Free online reading

## TABLE OF CONTENTS

ABSTRACT

TABLE OF CONTENTS

LIST OF TABLES

ACKNOWLEDGEMENT

DEDICATION

CHAPTER ONE

GENERAL INTRODUCTION

1.1 BACKGROUND OF THE STUDY

1.2 STATEMENT OF THE PROBLEM

1.3 RESEARCH QUESTIONS

1.4 RESEARCH AIM AND OBJECTIVES

1.4.1 Aim

1.4.2 Objectives

1.5 SCOPE OF THE STUDY

1.6 RESEARCH METHODOLOGY

1.7 SIGNIFICANCE OF THE RESEARCH

1.7.1 Significance to practice

1.7.2 Significance for future research

1.8 GUIDE TO THE THESIS

CHAPTER TWO

LITERATURE REVIEW

2.1 INTRODUCTION

2.2 COST ESTIMATION

2.2.1 Early cost estimation

2.2.2 Cost estimation in the Ghanaian Construction industry

2.3 PREVIOUS STUDIES ON COST ESTIMATION MODELLING

2.3.1 Definition of Cost Models

2.3.2 Models Categorization

2.4 COST MODELLING

2.5 BACKGROUND TO BUILDING COST ESTIMATION MODELS

2.6 CLASSIFICATION OF NON-TRADITIONAL COST ESTIMATING MODELS

2.6.1 Element-Based Floor-Area Model

2.6.2 Regression Model

2.6.3 Probabilistic Model

2.7 TYPES OF COST ESTIMATION MODELS

2.7.1 Traditional Models

2.7.2 Non –Traditional Models

2.7.3 New Wave Models

2.8 GLOBAL UTILISATION OF COST ESTIMATION MODELS

2.8.1 The Practice of Modern Estimation

2.8.2 The Need for Non-traditional Cost Estimation Models

2.9 BENEFITS OF COST ESTIMATION MODELS

2.10 THE BARRIERS TO COST ESTIMATING MODEL UTILIZATION

2.10.1 Lack of understanding

2.10.2 Lack of data management

2.10.3 Time constraints

2.10.4 Inaccurate outcome of models

2.11 STRATEGIES FOR UTILIZING COST ESTIMATING MODELS

2.11.1 Establishing Implementation Teams

2.11.2 Government and Constructors

2.11.3 Organisation of Workshops

2.11.4 Improvement of data management in firms

2.11.5 Introduction of model development in higher institutions

2.11.6 Enhancement of publicity on cost estimation modelling techniques

2.12 DRIVERS OF COST ESTIMATION MODEL UTILIZATION

2.12.1 Cost Advice

2.12.2 Cost information

2.12.3 Reliable estimated budget

2.12.4 Improve firm’s image

2.12.5 Reduce cost overrun

2.13 THEORETICAL FRAMEWORK

2.13.1 Activity Theory

2.14 CONCEPTUAL FRAMEWORK OF THE STUDY

2.15 DISCUSSION ON VARIABLES OF CONCEPTUAL FRAMEWORK

2.15.1 Background to building cost estimation models; Traditional and Non-traditional

2.15.2 Selection of a Model

2.15.3 Knowledge on Cost Models

2.15.4 The Barriers of Cost Estimating Model Utilization

2.15.5 Drivers for the utilization of Cost Estimation Models

2.16 CHAPTER SUMMARY

CHAPTER THREE

RESEARCH METHODOLOGY

3.1 INTRODUCTION

3.2 PHILOSOPHICAL ORIENTATION OF THE RESEARCH

3.2.1 Ontological and Epistemological considerations

3.2.2 Deductive and Inductive reasoning

3.3 RESEARCH DESIGN

3.4 RESEARCH METHODS

3.4.1 Qualitative Research

3.4.2 Quantitative research

3.4.3 Mixed method

3.5 DATA COLLECTION METHOD

3.6 RESEARCH POPULATION AND SAMPLING TECHNIQUE

3.6.1 Study Population

3.6.2 Sampling Technique and Sample Size Determination

3.7 DATA COLLECTION INSTRUMENTthe

3.7.1 Questionnaire Development

3.7.2 Selection of measurement format

3.7.3 Pilot Test

3.7.4 Structure of questionnaire

3.8 METHODS OF DATA ANALYSIS

3.8.1 Mean Score Ranking

3.8.2 Relative Importance Index

3.8.3 Factor Analysis

3.8.4 Ethical Issues

3.9 CHAPTER SUMMARY

CHAPTER FOUR

DATA ANALYSIS AND DISCUSSION

4.1 INTRODUCTION

4.2 RESPONDENTS CHARACTERISTICS

4.3 FREQUENCY OF PREPARING INITIAL COST ESTIMATE

4.4 KNOWLEDGE ON COST MODELS IN-USE FOR COST PREDICTION

4.4.1 Reliability of data analysis

4.4.2 Data analysis

4.4.3 Data analysis

4.4.4 Discussion of Cost Estimation Models in-use for Cost Prediction

4.5 BARRIERS TO THE UTILIZATION OF NON-TRADITIONAL COST ESTIMATION MODELS

4.5.1 Data Analysis

4.5.2 Discussion of Barriers to the Utilization of Non-Traditional Cost Estimation Models

4.5.2.1 Component 1 – Inefficient Techniques

4.5.1.2 Component 2 – Perception of model techniques

4.5.1.3 Component 3 – Unavailability of cost Data

4.5.1.4 Component 4 – Lack of Understanding

4.6 DRIVERS FOR UTILIZING NON-TRADITIONAL COST ESTIMATION MODELS

4.6.1 Analysis and discussion of Extracted Components

4.6.1.1 Component 1 – Efficient Cost Estimation

4.6.1.3 Component 3 – Risk Management

4.6.1.4 Component 4 – Improved Estimation Process

4.7 STRATEGIES FOR UTILIZING NON-TRADITIONAL COST ESTIMATION MODELS

4.7.1 Discussion on strategies for the Utilization of Non-Traditional Cost Estimation Models

4.8 CHAPTER SUMMARY

CHAPTER FIVE

CONCLUSIONS AND RECOMMENDATIONS

5.1 INTRODUCTION

5.2 RESEARCH QUESTIONS

5.3 ACHIEVEMENT OF RESEARCH OBJECTIVES

5.3.1 The first objective; To determine the level on awareness on types of cost estimation models in the construction industry;

5.3.2 The second objective; To identify barriers of cost model utilisation in the Ghanaian construction industry;

5.3.3 The third objective; To identify the drivers of cost model utilisation for cost estimation with regards to the construction industry in Ghana;

5.3.4 The fourth objective; To propose strategies for the utilisation of cost estimation models for cost estimation in the Ghanaian construction industry;

5.4 CONTRIBUTION TO KNOWLEDGE AND INDUSTRY

5.4.1 Contribution to knowledge

5.4.2. Contribution to industry

5.5 RECOMMENDATIONS

5.6 LIMITATIONS

5.7 FUTURE RESEARCH

5.8 CONCLUSION

5.9 CHAPTER SUMMARY

REFERENCES

## ABSTRACT

Construction cost estimation model deals with arrangement of data, assumptions and equations that permits translation of physical resources or characteristics into cost. These models serve as framework for forecasting the probable cost of proposed construction projects. It can be classified as either traditional or non-traditional depending on the cost variable formulation. However, in the Ghanaian construction industry, quantity surveyors traditionally estimate the initial cost of building projects using the traditional models, which have been criticized overtime for its inaccuracies. The purpose of the study was to provide account on cost estimation modeling as the basis for proposing strategies for the utilization of non-traditional cost estimation models. Specifically, it intended to determine the awareness level of quantity surveying professionals on cost estimation models; to identify the barriers of cost estimation models utilization; identify drivers of cost estimation model’s utilization and finally propose strategies for the utilization of non-traditional cost estimation models. A quantitative survey research design was adopted and utilized. Data collected were analyzed using, mean score ranking, relative important index (RII) and factor analysis (FA). The results revealed the existence of poor knowledge base among the quantity surveying professionals on non-traditional cost estimation modeling techniques. Furthermore, it discovered four critical barriers to the utilization of non-traditional cost models namely; inefficient model techniques, perceptions on cost models, unavailability of cost data and lack of understanding. Moreover, it was discovered that the utilization of cost estimation models for cost prediction is driven by; efficient cost estimation; cost advice; risk management and improved estimation process. The study proposed strategies for the utilization of cost estimation models which includes improvement of data management in firms; introduction of cost modeling development in tertiary institutions; enhancement of publicity about non-traditional cost estimation models and organization of cost modeling development workshops for quantity surveying professionals. The research recommends that, quantity surveying professionals should be educated on the benefits of non-traditional cost estimation models to improve their understanding and appreciate its utilization to help provide accuracy in cost estimation and planning of construction projects. It further recommended a future research which should focus on developing a framework for the utilization of non-traditional modeling techniques. The study however demonstrates evidence of barriers, drivers and strategies to the utilization of non-traditional cost estimating models based on the perspective of Ghanaian quantity surveyors.

**Keywords:** Cost Model, Cost Estimation, Barriers, Drivers, Strategies, Construction, Ghana.

## LIST OF TABLES

Table 2.1: Summary of cost estimation models

Table 4.1: Respondents characteristics

Table 4.2 Reliability Statistics

Table 4.3 Cost Estimation Models In-Use for Cost Prediction

Table 4.4 One-Sample Test

Table 4.5 Reliability Statistics

Table 4.6 KMO and Bartlett's Test

Table 4.7 Communalities

Table 4.8 Total Variance Explained

Table 4.9 Rotated Component Matrixa

Table 4.10 Reliability Statistics

Table 4.11 KMO and Bartlett's Test

Table 4.12 Communalities

Table 4.13 Total Variance Explained

Table 4.14 Rotated Component Matrixa

Table 4.15 : Strategies For Utilizing Non Traditional Cost Estimation Models

Table 4.16 One-Sample Test

## ACKNOWLEDGEMENT

Glory is to the Lord God Almighty who in His infinite wisdom granted me the grace and will power to embark on the research journey.

My profound and warmest gratitude goes to my Supervisor Nana Prof. Edward Badu for his encouragement, support, research guidance, constructive criticism and suggestions throughout the research process.

Furthermore, my gratitude to Surveyor Dr. Ernest Kissi, Daniel Oteng and Daniel Agyeman for their emended help, support and guidance in coming up with this research. Your time and support is greatly appreciated.

To the head of Building Technology Department, Lecturers as well as research Assistants I say God bless you all. Your selfless and patience in imparting knowledge to me has turned me out into a better person of which I am much grateful.

To all my friends who in diverse ways have been of help or motivation, I say God richly increase you.

## DEDICATION

This thesis is dedicated to my parents Mr and Mrs Agyekum and my wife Evelyn Adjei-Yeboah.

## CHAPTER ONE

## GENERAL INTRODUCTION

### 1.1 BACKGROUND OF THE STUDY

Construction projects require more capital and it demands that a cost plan is undertaken before its execution. The cost of construction projects is mostly needed by prospective clients to know the level of funds needed to handle cash flow (Kelly et al., 2002). Gunayin and Dogan (2004) argued that cost estimate is a major significant criterion in making decisions at the early phases of a building construction process which involves designing, bidding traditionally and constructing. It is vital to prepare cost estimates with high level of accuracy at every stage of the construction process especially initial stages. The exact cost estimates are necessary to the successful execution of every project as stated by Leung et al., (2005). According to Idoro (2009) time, cost and quality play a vital role in the success of every project, hence ideal to perform them at the minimum time, cost and quality.

An initial estimate often represents the cost of project and the budget for the client upon which major decisions are made as well as the first stage of appropriation and its economic feasibility (Henli, et al., 2005). The passage of time has evolved in many ways one can prepare cost estimates pertaining to construction projects such as single pricing and multiple pricing (Harris and McCaffer, 2013; Brook, 2004). According to Skitmore and Smith (1991), the process chosen for preparing cost estimate greatly relies on the level of information readily available and the duration. However, Gould and Joyce (2000) agreed that cost estimating accuracy largely depends on the information readily available and the time allocated for its preparations. Cost estimates establishes the cost of projects during the initial stage of construction projects and forms some important criteria to make decisions of whether the client must partake in the project or not. Hence, a great deal of importance to prepare these estimates adequately so as to avoid incorrect estimates which could probably lead to a poor project start up (Kwakye 1997). However, an observation by Leung et al., (2005) indicated that cost estimates are often produced within a rigid and hasty timeframe and most often inadequate data upon which major decisions are based on. Akintoye and Fitzgerald (1999); Leung et al., (2005) commented on short timeframe for cost estimating has contributed to inaccuracies in the preparation of cost estimates. Harris and McCafffer (2013) therefore suggested a considerable time say four to six weeks as the minimum time frame to come up with a cost estimate devoid of any problems thereof. It is worth noting that not much growth has been seen in the accuracy of cost estimates pertaining to the construction industry. Example, the level of forecast error established in single price estimating methods falls within and , (Lock 2003, Gould and Joyce 2000). Akintoye and Skitmore (1990) and Babalola and Adesanya (2013) reported accuracy of estimates between and for multiple price estimating methods. However, cost estimating models which involves the use of mathematical algorithms and parametric equations to estimate cost of projects with significant level of accuracy and fastness is not employed in Ghanaian building industry.

### 1.2 STATEMENT OF THE PROBLEM

The preliminary estimate of cost pertaining to construction projects is seen as a cost limitation for a project (McCaffer et al., 1984). The establishment of initial cost by estimation professionals is deemed necessary as it provides the basis for a prospective client to decide on whether to commence a project or not. Any inaccurate estimates caused by errors in the initial estimation process will bring about frustrations and dissatisfaction to clients (Akintoye and Skitmore, 1990).

In an effort to prepare cost estimates within an unfavourable time frame, some cost prediction methods are used. However, these methods have some problems which have effects on the accuracy of the cost estimates. Annan (2006) indicated that construction cost estimate experiences an overrun of 60% to 80% through a survey conducted on buildings for offices. In addition, Laryea (2010) revealed that consultants cost estimates in Ghana experience an overrun of 40%. Traditional building cost estimation strategy however has ended up ineffective due to computerisation and complexity of current construction projects (Ashworth 2010). Boussabaine and Cheernahm (1997); Mosaku and Kuroshi, (2008); Flood and Kartam (1994) submitted that cost overrun and construction cost escalation could occur at pre-contract and feasibility study stages of construction works. However, by feeding the relevant basic and easily accessible data into a typical cost model, a stable output in the form of cost estimate could be generated (Odusami et al., 2000; Ogunsemiet al., 2006; Oyediran, 2001).

The problem of high cost of construction materials in the industry has called for the need to adopt various techniques to ensure accuracy in estimating project cost (Lamuti, 2005). Kim et al., (2005) suggested that there is the need to use computerized estimation models to establish cost estimations of projects because of the growing number, size and complexity of construction projects. Aloysius (2010) noted that in predicting initial construction cost in the Ghanaian construction industry, quantity surveyors and cost estimators rarely use cost estimation models as compared to other techniques. In addition, Alorgli (2015) posited that there is no utilisation of cost estimation models at the initial stages of construction projects. Hence the goal of this study is to propose strategies for the utilization of non-traditional cost estimation models in the Ghanaian construction industry.

### 1.3 RESEARCH QUESTIONS

Questions pertaining to the study are asked below;

1. What is the level of awareness of cost estimation models with regards to construction industry in Ghana?

2. What are the barriers to cost estimation model utilization with regards to construction industry in Ghana?

3. What are the drivers of cost estimation model utilization in the construction industry?

4. What are the strategies that will help utilize cost estimation models in construction industry in Ghana?

### 1.4 RESEARCH AIM AND OBJECTIVES

#### 1.4.1 Aim

The aim of the study was to propose strategies for the utilization of non-traditional cost estimation models in the Ghanaian construction industry.

#### 1.4.2 Objectives

The following objectives were adopted to achieve the stated aim of the research:

- To determine the level of awareness of types of cost models in the construction industry;

- To identify barriers of cost model utilization in the Ghanaian construction industry;

- To identify the drivers of cost model utilisation for cost estimation with regards to construction industry in Ghana; and

- To propose strategies that will help utilize cost estimation models in construction industry in Ghana

### 1.5 SCOPE OF THE STUDY

It focused on quantity surveying section of A.E.S.L and other private quantity surveying firms in good standing with Ghana Institution of surveyors in Ghana. Contextually, the study looked at cost estimation models utilized by construction industries in Ghana and proposed the most effective strategies for the utilization of non-traditional cost estimation models.

### 1.6 RESEARCH METHODOLOGY

In the bid to attain the aim and objectives of the research, positivism in the ontological paradigm was adopted employing the use of quantitative approach in the data collection needed for the research after which descriptive analysis was conducted on the data. The research was carried out through field survey after purposive sampling technique to select the respondent population. Both opened ended and closed ended questions were adopted for this research. This study was descriptive research based on quantitative approach.

Analysis of data was performed by the use of the IBM SPSS software package (version 24.0). Descriptive statistics such as Relative Importance Index (RII), Mean Score Ranking and Factor Analysis were used to analyse the data.

### 1.7 SIGNIFICANCE OF THE RESEARCH

The study results are expected to be useful for various parties, among others: Quantity surveyors and cost estimators, institutions and the academia.

#### 1.7.1 Significance to practice

The research is expected to influence the construction industry professionals to appreciate cost modeling in estimation utilizing non-traditional techniques since the study will detail cost estimation modeling techniques which will improve on their level of understanding.

#### 1.7.2 Significance for future research

The gaps established through this research are expected to influence further empirical studies on cost estimation model utilization in the Ghanaian construction industry academically.

### 1.8 GUIDE TO THE THESIS

This research employed five different chapters; **Chapter one** dealt with general introduction: background, statement of problem, aim, objectives, scope, methodology and organization of the study. **Chapter two** concentrated on various literatures relating to cost models, types of cost models and barriers to cost models, empirical studies on cost models. **Chapter three** was research methodology; research philosophy, sampling frame (target group), instrumentation design (questionnaire), administration and tools for statistical Analysis. **Chapter four** covered the analysis of data and discussions of outcome. The strategies for cost estimation model utilization are thus developed from the data analysis and the discussion **. Chapter five** concluded the overall research and suggested recommendations.

## CHAPTER TWO

## LITERATURE REVIEW

### 2.1 INTRODUCTION

This section of the report gives an in-depth knowledge on the review of relevant literature related to the subject of cost estimation and cost estimation models in the construction industry. The chapter began by giving the overview and significance of cost estimation, defining cost modelling in the context of the study, barriers and drivers of cost estimation model utilization and strategies for the utilization of cost estimation models.

### 2.2 COST ESTIMATION

Gould (2005) established the definition of an appraisal estimation compared to the project cost subjected to the real construction. Butcher & Demmers (2003) defined cost estimation is a planned prediction of the estimated construction cost that will be incurred in certain specified buildings. It plays an important role from management to library planners at the design stages of a project in providing adequate information pertaining to the budget of the project and the facility. This implies that, cost estimating is the cost of physically constructing the project in the time required. Clough (1986) affirms that estimation of construction projects is gathering and examining the numerous components affecting the project pertaining to the total cost. Stewart (1991) employed a definition of cost estimation as emanating from the society of cost estimating and analysis (SCEA), as "the mechanism of having an estimate to the project subjected to the information readily available ". Ritz (1994) affirmed that the “cost of a project estimate is all cost involved in bringing the project into reality”. As defined by Ahuja et al., (1994) defined cost estimate as "the art of projecting the expected cost of projects”.

The Association of Advanced Cost Engineering defines the estimate of cost as the items of a project or efforts directed to the project within an agreed scope. Construction cost estimate is also defined as a clear representation of the total cost a project (Collier, 1987). Cost estimation is defined as the projected cost of a building considering the characteristics of other similar projects at the first stages of the project.

#### 2.2.1 Early cost estimation

Ahuja et al., (1994) stated that estimation is the core mandate of industries involved in construction; the inaccuracy of cost estimates can negatively influence the successful execution of building projects beginning from the early stages of a project through the tender estimates. They also stated that most projects turn to fail because of the inaccuracies of cost estimates. Before the execution of construction projects an estimate is provided, the result of the estimate devoid of ambiguity of such cost estimates is largely dependent on the attribute and characteristics of the one estimating. Raddon (1982) established skills to be the ‘appropriate use of tools pertaining to estimation, and judgement’. Nonetheless, estimation plays a vital role towards the achievement of a project. Ibrahim (2003) affirms that a cost estimates possesses three variables which includes; the quality, the quantity and the cost. These forms part of the factors pertaining to construction, which includes specifications and drawings. It states materials needed in terms of its quality as well as quantity needed and cost is subjected to two elements. An adjustment is made to suit the total cost of materials subjected to the quality and quantity of the materials to be used. The mandate of any cost estimator is to come up with an estimate of the intended cost of a project to carry into the future. The client has to have a foreknowledge of the total cost of the project before the commencement of the project. This establishes how the project can be managed during the design phase or early phase and the most cost-effective choice from a list of other design proposals. In most cases, pre-design cost estimation represents both budgetary and planning tools. Mostly, clients pay attention to how much it will cost them. Clients turn to shun away if the initial cost of the building project is too high, accounting to the loss of interest in such a project (Ashworth 1993).

The accuracy of a cost estimate at the initial stage is very vital because it represents a budget limit for the client as well as the basis for decisions pertaining planning and source of funding (Newton 1991). It serves as a cost management tool so as not to go beyond the budget when designing and to establish that the tender figure can be made known with a level of assurance to the team responsible for designs. However, it is important to note that the bid to estimate the cost within the shortest possible time is the cause of inadequate recorded past cost data, as is the norm, the correct preliminary cost establishing the processes is utilized (Skitmore et al 1999). The method to be used is greatly affected by the quantum of information needed, the duration needed in processing and the vast experience offered on the estimator’s side. In the provision of an initial estimate of cost accurately, nonetheless, the estimates are of higher importance to the organization providing the funding and the team carrying out the project. For the organization providing the funding, Oberlender and Trost (2001) affirmed the preliminary estimation of related cost is important in businesses to make informed decisions that include mechanisms for the creation of assets, probable screening of projects and the allocation of resources to subsidize future plans of planning and developing projects. On the part of the team carrying out the project, the success of projects is arrived by comparing the exact cost and the initial proposed estimated cost. The estimating method to be employed is greatly dependent on the data at hand (Akintoye 1992). As the project moves from one stage to the other, the estimates which have already been done changes and hence the probability used increases and is devoid of ambiguity. Holm et al., (2005) establishes why cost estimate should be produced.

The most widely used estimating techniques for projects relating to construction are functional unit method, the floor area unit method, the cubic method, the storey enclosure method and cost modelling (Ashworth, 1994). The mechanism adopted by the functional unit is often relied on when accurate information is available. This establishes the preparation of an estimated cost because of the similarity in projects when the only difference thereof is the size. Seeley (1996) affirmed the lack of precision as a major weakness, accounting the misappropriation of factors including the size, shape, construction form, material and finishing. The techniques pertain to public projects and to the initial phase of the establishment. The floor area unit is mostly employed in our building industry. It involves the measurement of the whole floor area pertaining to storeys. Cost estimation of a floor area of a proposed building can be established by the multiplication of the calculated square meter cost and the historical square meter cost.

The main setback encountered in the use of this method is the difficulty in the precise allocation of separate allowances pertaining to the project. Cost-per cubic-metre estimation cannot be relied on unless there is significant relationship between comparable buildings. The major set-back is the employment of deceptive simplicity. It appears to be a common phenomenon to mathematically deduce the building volume, but the problem rest on the inception of various factorial designs into the cubic unit-rate. It does not take into account allowances which also contribute to the total cost; In addition, cost variations coming from other foundation types are inconvenient to be added to the single unit (Seeley, 1996).

The storey enclosure method employs the measurement of the vertical planes and horizontal planes of the building. It is hardly used in industries because of the quantum of work required and its related cost. Cost models are employed in forecasting the required estimated cost (Ashworth, 1994). A cost model emanates from information at hand concerning performances. This information is examined which establishes the patterns which can be related to predictions made. The approximate estimate method is linked to the preparation of an initial estimate of the project based on data provided. There are similarities between the pricing method and some abbreviated bill of quantities, which establishes remarkable checks in the first phase of the project. The major set-back of this method is the amount of time spent in the process and only applicable in the subsequent phase of the plan pertaining to the design that has been established. The elemental cost analysis depends on projects previously performed. The cost depends on the external basis but the external unit cost is divided into elements and sub-elements, provides the environment to adjust cost to be incurred in the design new of projects in relation to the one which held previously. It is acceptable and understood by all those involved, nonetheless, it consumes time.

#### 2.2.2 Cost estimation in the Ghanaian Construction industry

The industry relating to construction is characterized by organizations involved in activities such as the construction of real estates, private and public infrastructure (Anaman et al., 2007). The building industry is generally concerned with building construction projects. The building industry is made of three participants; the owner (customer), the designer, and the contractor. The construction process involves the following; the owner requests the services of an engineering firm to come up with a design suitable and request for the cost, the project is then placed out of bid to contractors and the contractors perform the actual work (Loushine et al., 2007). The process of bidding however requires the preparation of estimates by either the quantity surveyor or the cost estimator. Cost estimation procedure is a method used by consulting firms or estimation department in clients‟ agencies to come up with the preparation of cost estimates based on readily available information (Assaf & Al-Hejji, 2006). In Ghana, Quantity Surveyors (QSs) graduate from both the universities and polytechnics (Fugar and Adinyira ,2009) who pursued programs in building technology and quantity surveying who are employed by quantity surveying firms to come up with the estimates of projects. A realistic and accurate cost estimate is of importance since clients of building construction projects rely on the initial estimates to make budget. It is also important because the fate of any building project relies on the accuracy of costs estimates. Cost estimators and quantity surveyors adopts various methods in preparing cost estimates, however a research by Maalinyur (2010) indicated the common cost estimation models used in Ghana is the cost per meter. There is therefore an indication that the Ghanaian construction industry continues to rely on the traditional model techniques in the preparation of cost estimates. Traditional building cost estimation strategy however has ended up ineffective due to computerization and complexity of construction projects in recent times as stated by (Ashworth, 2010).

The past years have yielded greater understanding and utilization of non-traditional cost models leading to the ability to decipher complex driven projects pertaining to technology so that the important tools influencing cost can be ascertained. The use of these estimates in budgeting, scheduling and control of projects comes with an impeccable use of resources.

### 2.3 PREVIOUS STUDIES ON COST ESTIMATION MODELLING

There have been many attempts in the past researching into cost estimation models. They were mainly part of more comprehensive models aimed at assisting contractors and organisations to estimate cost on projects (Kaka et al., 1996). Typically, these were achieved by the collection of data relating to similar projects and its overall characteristics.

Fortune and Lees (1996) investigated the technique used in cost modelling by firms in the north of England and Wales. Consulting quantity surveyors were 62% of the sample space in his study. The remaining was made up of multidisciplinary authorities. Fortune and Hinks (1998) also studied the utilization of models relating to cost in the provision of early cost advice or estimates by consultant quantity surveyors throughout England. In their study, it was revealed that the traditional method of estimation was in widespread use which included superficial, cubic, functional unit. A study on exploring the types of cost modelling techniques used in Malaysia by Bari (2012) indicated that the traditional cost estimation models were in widespread use whereas non-traditional models that were in widespread use included value management, resource-based model and life cycle models.

Stochastic modelling takes into consideration probability and uncertainty in the estimation (Bernie and Yates, 1991). Decision tree, utility theory and the Monte Carlo simulation techniques was largely employed in estimating cost. Brandon (1982) commented on the importance to conduct a research directly in developing the cost theory models. A few researches have been done in connection to the grouping of all modelling cost (Raftery 1984; Beeston 1987; Newton 1990; Skitmore and Patchell 1990). It was revealed that the cost modelling techniques that were mostly employed by South African quantity surveyors were the approximate quantities and the elemental analysis techniques, a study by Bowen and Edwards on traditional cost estimation methods. In addition to these, other methods that were included in the phases of the designing processes were the bills of quantities method as well as the superficial method. Gwang (2004) analysed various ways by which cost can be estimated. It was concluded that one can achieve the best by employing neural network techniques for cost estimation.

Newton (1990) establishes nine types of cost modelling studies; they are (1) data, (2) units, (3) usage, (4) approach, (5) application, (6) model, (7) technique, (8) assumptions and (9) uncertainty. With regard to the technique of cost models, ten types were classified; (1) Dynamic programming, (2) Expert system, (3) Functional dependency, (4) Linear programming, (5) Manual, (6) Monte Carlo simulation, (7) Networks, (8) Parametric modelling, (9) Probability analysis and (10) Regression analysis. Despite the evolving creation and development of newer techniques for constructing cost models, floor area and quantity based models which are manual techniques are widely used today. Results of the use of forecasting techniques have all been proven by Nigerian practitioners (Akintoye et al,. 1992), South Africa (Bowen and Edwards 1998), and United Kingdom (Fortune and Lees 1996; Fortune and Hinks 1998, Aloysius, 2010).

Yu, proposed the utilisation of PIREM (Principal Item Ratios Estimating Method) which integrates various existing methods including parametric estimating, ratios estimating, and cost significant model with advanced nonlinear mapping techniques, and develop a mechanism that distinguishes unit prices with the quantities of a cost item (Yu, 2006). Preliminary cost estimation based on case based reasoning and genetic algorithm, a research by Kim (2010) suggested that the estimation accuracy outweighs that of the conventional cost estimation models and therefore called for more development of such models and their implementations. Fellows and Liu (2000) a research on cost modelling, classified building cost models into two types. These included Clients oriented and customer oriented. The client-oriented determinist cost models are widely employed through the stages concerned with design to establish the related cost of the project. Most models are generated from stored historical data. Aloysius (2010) studied into initial building project cost and cost model utilisation. The study revealed that cost estimators and quantity surveyors seldomly use the non-traditional cost estimation models in the preparation of cost estimates. Additionally, Alorgli (2015) in a similar project, asserted to the fact that traditional cost modelling techniques are in widespread use among quantity surveyors in the Ghanaian building industry at the expense of non-traditional techniques which provides more accurate, reliable and faster cost estimates. Based on the literature reviewed, there is an indication that the application of cost models in early cost estimation of building projects is a promising area. However, there exists a gap of literature on strategies for the utilisation of non-traditional cost estimation models. This has therefore formed the basis of the research, thus a study to propose a strategy for the utilization of cost estimating models in the Ghanaian building industry.

#### 2.3.1 Definition of Cost Models

A model can be described as simplified representation of complex reality. Vermande et al., (2000) describe a model as system that use simple inexpensive object to represent complex or uncertain situations. The system that makes use of this is termed modeling. To this end therefore, modeling is the process of converting complex real-life problem situation to simple representation of the problem situation. Below is possible mode of derivation of empirical models using model of building process:

Formulate the Problem Collection of data Analysis of Data Model building Optimum Model Evaluation of Model Testing Application

(Source: Ashworth, 1994).

#### 2.3.2 Models Categorization

Models can be classified into two broad categories: Product-based cost model and Process-based cost model

** Product-based cost model: ** These types of models are considered best for finished products (Moore et al., 1996 and Ferry, 1999). These types of models do not consider details of the design and basis its judgments on the parameters pertaining to the building. Such parameters are as follow: i) the floor area of the proposed projects (gross or net). ii) The quantum of the proposed project iii) Some user ‘s parameters, such as number of pupil places for a school or number of bed for hospital.

** Process-based cost model: ** This is the type of model that deals with construction items process of formation. This is adjudged the most accurate of the models. It is often argued that it is a process that actually generates costs; however, the cost cannot be generated until the form of building has been conceptualized. With this, process approach could not be best approach to be adopted at early design stage, since little information would be available for analysis. This view was supported by Moore (1999) that attempting to model construction process at too early stage can result in over-riding of the design process in order to arrive at bricks-and-mortar solution before the user criteria have been properly worked out. Process -based cost models, can further be classified into sub-types, within the context of probability model and deterministic model or combination of the two.

** Probability Models: ** These models recognize the fact that specified variables can only be estimated using the probability theory. Majority of models falls into this category.

** Deterministic Model: ** This type of model affirms that variables can be accurately be estimated. Models can be classified based on source of extraction or abstraction. They can be classified based on their degree of abstraction (Adedayo, 2006). There are three groups under this: Abstraction or mental models (High abstraction); Physical models (Low abstraction); and Symbolic models (Moderate abstraction).

*Abstraction or Mental Models:*Mental models are models that are ill-structured representation of reality, that cannot be feel or touch physically. That is, they lack physical or symbolic configuration. These types of models involved high level of creative imagination; they are unclear image of complex objects that have not been finalized. Summarily, abstract or mental models require high level of abstraction or creative ability.

*Physical Models* *:* Physical models bear semblance with the real objects. They are usually a prototype of the real objects or possess characteristic that reflects the features or function of the real objects. There are two types of physical models: Iconic Models: Iconic Models are used to represent real features of an object. Scaling system is often used in the feature representation. The scale could be upward or downwards. They could also be presented in modular form (using system of dimension grids), three dimension forms, like model card to two dimensional models like sketches, photography ‘s and paintings. The aim of iconic model developments is to physically represent client needs and requirements, e.g. architectural building models, model airplane, model train and car and so on and so forth. The scale used in iconic model is to convey design ideas to clients for feedback; and Analogue models: Analogue models are like physical models but they may not look exactly like the reality. They aim at performing basic functions instead of emphasizing and communicating ideas about appearances. They may or may not look like the real thing. Examples are flow diagrams, maps, circuit diagrams, building plans organization plans (Adebayo, 2006; Ashworth, 1994).

** Symbolic Models: ** These types of models represent ideas using numbers, notations, mathematical formulas, musical notation. They have lower level of abstraction when compared with mental models. There are two types of symbolic models; namely: Mathematical models: these models are also symbolic in nature. These models use symbols to express or simplified relationship between variables of complex problem e.g. y = mx + c Where y stands for dependent variable, c stands for intercept on y-axis, x represents dependent variables, m is the slope of the graphical relationship. This is an example of mathematical model. These types of models find application in operation research, decision analysis, production management, complex computation problems, optics, material allocation, construction programming, electrical engineering, naval architecture and shipbuilding; and Verbal models: These models are referred to as written form of mental models of ideas. Examples are: poems, plays, stories, theories, television adverts of products. The models utilized tools of figurative expression, beautiful painting of scenario to appeal to people or customers sense of reasoning.

### 2.4 COST MODELLING

Cost modeling is the clear depiction of all processes that establishes an acceptable series of data input (Bari, 2010). The process is widely subjected to thorough examination of all accumulated data influencing cost. The importance is to produce accurate information that can be dependent on for accurate decision making. Ji et al., (2010) affirmed it as a nominal representation of a system that contributes high on cost. It plays a higher importance through helping consultants of projects and contractors to provide a more accurate, cost efficient and reliable advice on cost to their respective clients.

Willis and Ashworth (1987) described cost modeling as a new technique employed to establish the estimated cost of an intended project. Ferry and Brandon (1991) however defined it as a depiction of a process showing all factors contributing to cost. Koo et al., (2011) establishes the cost models which forecast the cost of a building project with the aim of making an informed decision. Cost models serves as tools through which management make informed decisions pertaining to the area of the design pertaining to the building (Skitmore and Marston, 1999). Its purpose is the ability to produce accurate cost estimates (Elhag and Boussabaine, 1998) for both contractor and client (Ashworth, 1999). Cost model in this document shall be referring to as a nominal expression that predicts the cost of a proposed project taking into consideration the factors which influence its cost. The historical view of cost models categorizes it into three prominent sets below: (i) the first-generation models. This was used widely for cost planning approach prior to 1950’s. It involved using the functional elements of a building (ii) the second-generation models. They are based on regression principles. They have been utilized from mid-1970’s to date. (iii) The third-generation models which are based on Monte Carlo simulation principles (McCaffer 1975; Yaman, 2007). Grouped into probabilistic and deterministic models, the deterministic models, all the variables can be evaluated precisely. However, in the case of the probabilistic models, only few variables are known with certain degree of certainty. Yaman (2007) also grouped the cost models according to their features. The main cost estimation models known traditionally is grounded on quantities including, mono-evaluated expense estimation models utilized as a part of the schematic design stage, (for example, unit, square, 3D shape and building construction envelope). Again, various cost estimates based on personal intuition and judgments through the vast experiences and data consisting of cost from previous works were established by Arafa and Alqedra in 2011. In turn, sense of estimate is an attempt to assess or estimate a value through analysis and calculation based on experience (Dipohusodo, 1996).

Plan of budgeting pertaining to construction only seeks to establish the estimated cost to be incurred in the building process. Thiry (1997) stated one of the ways used to compile estimates of cost into the offering price is the cost model. According to Poh and Horner (1995), mechanism used to establish cost of a project pertaining to building is the parameter technique, where prices are estimated from the price of unity building floor area. According to Ashworth (1994) cost model is employed to estimate cost of project. A model which best explains the collected data in relation to price or the project. Multiple linear regression analysis is also employed to estimate cost or ascertain the price of cost. Thiry (1997) presented the principal emphasis was on efficiency, then in making the model should refer to the practitioners who are experienced in determining the price. Cost models can provide a variety of uses for such estimates of the cost components of planning and control system in the bid to alter the value forecasted before and after the whole contract.

### 2.5 BACKGROUND TO BUILDING COST ESTIMATION MODELS

Dating back to 1950’s, an immense research with the aim of understanding the relationship between the parameters relating to the design and the cost of the building and to come up with models in other to be able to forecast and predict the cost of building. Cost modelling is sometimes referred to as formation of system which components have some level of influence on the cost of a project (Holm, 2005). Based on history three phases of cost modelling have been identified. The first to consider is the first-generation model which was adopted from the “basic element of oriented cost of planning approach”. This model was introduced in in the late 1950’s and used extensively till the late 1960’s. The second model introduced was the second-generation model. This model was based on regression analysis which started around the mid 1970’s (McCaffer, 1975). The last model was the third-generation model which had its bases on Monte Carlo Simulation technique techniques which started in early 1980’s (Touran, 1992). Furthermore, cost estimation models can be grouped based on the characteristics they exhibit. With regards to this, two types of cost estimation models are identified and they include; traditional cost estimation model and the non-traditional cost estimation model. The traditional cost estimation model was generally based on quantities that is, models based on the resources used in the construction phase, models based on the building operational units and functional elements of the building (Akintoye &Fitzgerald, 2000; Ashworth, 1988; Seeley, 1976; Bledsoe, 1992; Flanagan &Tate, 1997; Mann, 1992; Mc Caffer et.al, 1984; Newton, 1991; O’Brien, 1994).

According to Boehm et al., (2000) and Singh (2014) most cost model emerged around 1970’s, which consisted of Software lifecycle management (Putnam and Myers, 1992). Checkpoint (Jones, 1997), Prices (Park, 1988), Surveillance, Epidemiology and end results (Jensen, 1998) and Constructive cost model (Boehm, 1981). Although most analysts began extensive engagements at the advancement of cost estimation models in the same time, they all went up against similar challenges as programming developed in size and significance. The field of programming estimation maintained the enthusiasm specialists who made it possible to establish the programming design expense model in every other field of work.

### 2.6 CLASSIFICATION OF NON-TRADITIONAL COST ESTIMATING MODELS

Element-based floor-area models, regression models, and probabilistic models are the classifications developed by Raftery 1987 in his bid for classification of construction cost models.

#### 2.6.1 Element-Based Floor-Area Model

The model being propounded in 1962 emanating from the United Kingdom employed the use of tenders and divided it into various categories. The cost planning technique used was researched by Ferry and Brandon (1980). The set-back (cost) not only affected floor area alone but other factors (Heng Le et al., 2005).

#### 2.6.2 Regression Model

The model mostly employed is the multiple regression analysis (Ashworth, 1994). The technique finds a model which best explains what is at hand (Norusis, 2004). Al Momani, (1996) argued that this technique is often utilized when the relationship used amongst variables is not unique, there is the ascertainment of similarities amongst variables. Another model whose purpose was to provide cost estimation using regression analysis for public school buildings (Skitmore and Ng (2003). According to Heng Li et al., (2005) a new model was again introduced and this model was used to calculate the possible cost for all office buildings in Hong Kong quoted Kouskoulas and Koehn (1974) designed an early design cost estimation model based on regression analysis.

#### 2.6.3 Probabilistic Model

According to Jaggar et al., (2002) the method which are mostly used is the Monte Carlo Simulation where activities were simulated over a period based on the life history of the system to be studied. These model techniques are not acceptable compared to the traditional models. The term model represents a misnomer, Monte Carlo evaluation of a construction cost is employed by estimator’s subjective perceptions of probability distributions for each of a set of reasonably independent sub-systems (Raftery, 1991).

### 2.7 TYPES OF COST ESTIMATION MODELS

Since 1950’s, there has been immense research with the aim of understanding the relationship between the parameters relating to the design and the cost of the building and to come up with models in to be able to forecast the cost of buildings. Cost modelling can also be termed the formation of a system defined with all those variables having an influence on cost (Holm, 2005). On the basis of the historical development, cost models are grouped into three. The first-generation models started from functional elements of building-oriented cost planning approach in England at the end of 1950’s and was extensively used until the end of 1960’s. The second-generation models started from the regression analysis since mid-1970’s (McCaffer, 1975). Touran (1992) argued that the third-generation models that were used in the 1980’s were generally based on Monte Carlo simulation technique. Studies have revealed that the model pertaining to cost can be categorised into two namely deterministic and probabilistic models. For deterministic models, it is presumed that the values can be allocated with any form of variables and can be accurately estimated. On the other hand, in probabilistic models comes with the affirmation that even though some variables are uncertain they can still be calculated. However, models can also be grouped based on their characteristics. In a broader perspective, the traditional cost estimation models are calculated based on quantities. The second model to be considered is the non-traditional model, these models are composed of new techniques and practices (Akintoye & Fitzgerald, 2000; Ashworth, 1988; Seeley, 1976; Bledsoe, 1992; Flanagan &Tate, 1997; Mann, 1992; Mc Caffer et.al, 1984; Newton, 1991; O’Brien, 1994). The level of knowledge on the types of cost models facilitates the adoption of a kind of cost model technique for cost estimation (O’Brien, 1994).

However, Naaret al (2012), Ashworth (1994), defined cost model as “the mechanisms used for calculating the estimated cost on an intended building project”. Therefore, all methods techniques or procedures used by quantity surveyors for cost estimation or cost forecasting may be termed as cost models. In his classification, the types of cost estimation models included: Traditional models: Conference, Financial method, Functional unit, Superficial-perimeter, Cube, Storey enclosure, Approximate quantity, Bill of quantities; Non-traditional models: Statistical or Econometric model (Regression analysis and Causal model),Risk and Simulation model (Monte Carlo Simulation and Value Management) Knowledge - based model; Resource- based model; Life cycle model; Artificial Intelligent System (Intelligent system (Neural network and Fuzzy logic) and other models (Environmental and Sustainable development).

#### 2.7.1 Traditional Models

** The functional unit method:** According to Seeley (1996) this technique is utilized when there is adequate information readily available to the estimator as a matter of fact on a specific sort of venture to match some final item to development costs. It permits an appraisal to be arranged for a comparable task when the main significant distinction amongst ventures is their size, the utilization of this method is restricted to open ventures and or ahead of schedule phases of task definition where next to no outline has been embraced.

** The cubic method:** (Seeley, 1996) argued that the cubic technique for assessment establishes the expense of a building to its volume. Taken a toll for each cubic-meter appraisals are somewhat temperamental unless basically indistinguishable structures are thought about, as there is no link between the expense and the volume of the building. The essential shortcoming of this system is its misleading straightforwardness. This technique neglects remittance pertaining to shapes, story tallness and number of stories, and for section separating, all of which impact cost; also, cost varieties emerging from contrasts, for example, elective establishment sorts are hard to fuse in single unit rate. This method has mostly been employed since ages, but has not been used due to its set-backs. Most countries in Europe have its architects and engineers being familiar with the related cost of building. Architects keep a document called a cube book. Ascertaining the cost of projects that are new is by calculating its volume and selecting an appropriate rate from the cube book

** The storey enclosure method:** The storey enclosed system is in light, the zone and vertical planes of buildings and aims to establish an evaluating framework. The system has encountered a very low turnout in its use in the industry, in view of the volume of work included and the deficiency of distributed expense information for its use.

** Conference Estimate: ** This method is relied on whenever there is a need to prepare an early estimate of price. It is subjected to the views of people and fails to follow in any specific order. It concerns itself with appropriate experience of similar cost estimation pertaining to similar project (McNeil, 1981).

** Superficial Floor Area Method:** The system includes measuring the aggregate floor region of all stories between outside dividers without conclusions for things, for example, inside dividers. By increasing the verifiable square-meter expense by the computed square meter of floor range for the project, a pre-construction preparatory expense gauge can be resolved. The significant shortcoming of this system lies in the failure to consider factors such as storey height and number of floors (Ashworth, 1991). The area is subjected to internal measurement of dimension not considering the deductions pertaining to internal stairs, walls and lift zones (Berthouex, 1972). Brook (1998) emphasized that estimation could be made pertaining to inflation and location but when concentration is geared towards specificity. Ashworth (1994) also emphasized that attention should be directed towards all processes in conjunction with the projects of construction anytime the rate is employed. This comes with its related problems but mostly finds its way in the Ghanaian Industry. In the UK, the source of data is acquired from the previous bills of quantity. The standard form of cost analysis can also help in establishing rates for designs provided early. Tender predictions are arrived at, from the analyses of quantity bills derived from past contracts which are related as well as the information at hand.

** Superficial Perimeter Method:** Southwell (1971) stated that the floor area was the greatest single variable –correlated price produced a formula that concerns in the engagement of floor area space with the length of the building perimeter. The ratio of the floor area is very paramount in designing of buildings and comes out with an unambiguous design. The method has since been unemployed due to difficulty in adjusting to change by the quantity surveyors.

** Unit Method **: The method comes with the selection of a standard unit of accommodation and multiplying by an appropriate cost per unit, the unit may represent (McNeil and Hendrickson, 1982). The method finds itself in most construction industries being utilized during the start of the construction phase (Brook, 1998). This is solely dependent on the relationship existing between the cost of construction and number of functional units it holds. Nonetheless various sites, specification changes, and inflation can always be adjusted. The set-back for this method is its lack of precision.

** Story-Enclosure Method:** James (1954) propounded a new way by which calculations can be done. Twice the area of the lowest floor, the area of the roof measured on plan, twice the area of the upper floor plus an addition of 15% for the first floor, 30% for the second floor, 45% for the third the area of the external walls. It however established itself to be the best compared to other methods like the single-price method but since it is not mostly employed, it is difficult in verifying. The weighting employed is subjective and dependent on every building. The disadvantage of the single rate approach was mostly experienced around1954, however it may be employed.

** Estimation model based on Bills of quantities:** BOQ is the traditional system which is usually utilized by development organizations to figure the expense of a venture in the point by point outline stage and all through the development. BOQ frames a standout amongst the most essential piece of the delicate reports together with specialized drawings, details, states of agreement, and so on when the information of a task is given in details, each item of work expected to finish the undertaking is recorded and evaluated by estimators (Akintoye and Fitzgerald, 2000). Bills of quantity to determine the cost of project, a detail drawing from which the Bill of quantities (BOQ) are ascertained. The set-back of this model is the inability to produce an estimated amount subject to the commencement of the project.

** Estimate Based on Engineer's List of Quantities:** Estimation is done based on the components available and the related quantities from which quantities of the project cost is ascertained. Payment is based on the level of completed work. This estimation establishes the pedestal of details which can be used as a yardstick to the projects progress (Skidmore and Ng, 2003).

#### 2.7.2 Non –Traditional Models

** Statistical model:** The model establishes the approximation of realities and the predictions that emanates from the approximation. The equation being utilized is the statistical model. Statistical models include multiple regression models and causal models.

The steps illustrated in figure 2.2 depict the stages in the statistical modelling process;

Abbildung in dieser Leseprobe nicht enthalten

Figure 2.2: Statistical Cost Estimating Procedure. (Adapted from Black 1984)

The best approach is to identify the problem. It deals with determining the aims and areas of the study (Black 1984) followed next by the data collection, historical data are very important in parametric estimation (Rose 1982). The historical database produces the parameters for engineers (Meyer et al,. 1999). Data collection is a very important stage. Without sufficient data, estimation will be inaccurate. The third step is the normalization of the data. Mostly, in building construction project, the cost data for each project must be subjected to time and location (Orczyk 1990). The fourth step is the parametric cost estimating model development. This technique estimates relationship in the cost estimation models by using mathematical models or graphs. CERs are statistical models that characterize the cost of a project as a function of one or more independent variables or parameters. Rules of thumb and the unit-quantity method are not recognized as CERs (Ostwald, 2001). Hegazy et al., (1998) argued that traditionally cost estimating relationships are developed by applying regression analysis to historical project information (regression analysis establishes the relationship existing between the parameters and cost, and the mathematical form of the model (Black, 1984; Orczyk, 1990).

In recent times, countless experiments have been conducted pertaining to neural network method as an alternate method to statistical method of coming up with parametric cost estimating model (Hegazy et al., 1998). Other researchers had also affirmed the utilization of that method in projects of construction (Moselhi et al., 1992). Hegazy et al., (1994, 1998) explained that the neural network is regarded as the black box approach since the neural processing is difficult to be traced and explained which however dwells on trial and error in setting the connection weights during the model development. Both concludes that the neural network research to date is quite experimental because neural network theory does not yet provide applicable rules, optimal setting of control variables and topologies (Bode, 1999). Krieg states that the success of the parametric model depends on capturing the repeatable characteristics of the building projects to be estimated (Krieg, 1979). The fifth step of the parametric estimating procedure is the establishment of model limitations. Since the model is usually developed from a limited data set, its validity is dependent on the ranges of variables used in the model (Orczyk, 1990). Rose (1982) affirmed that the interpolation and extrapolation must accurately be done and the range used to estimate the validity of cost must also be known. For that matter Orczyk (1990) concord that the extrapolation must be done beyond the range which will produce undoubted results. The final step which is the sixth method is the documentation of the model development process which supports the proper implementation process (Black, 1984).

** Multiple Regression Model:** Kim et al., (2004) were able to establish some draw backs of regression estimation model. The model failed to define the mechanism aiding estimators to choose the appropriate cost model; a sure sort of different mathematical equation and its information are thought to be suitable for the relapse comparisons; there must be a change on the variables employed. On the other hand, relapse cost estimation models have been utilized for evaluating expense subsequent to the 1970s in light of the fact that they have the benefit of a best unmistakable numerical premise. Multiple regressions generally come in this format C = X1+ X2+ …Xn, the C aggregates estimated project cost (the dependent variable) and X1; X2. Xn are measures parameters which help in assessing C. For instance, X1 can be the gross floor zone, the number of stories... etc and d1; d2… …. dn are the coefficients from the regression analysis.

** Causal model:** Estimation is based on the presumption that future values pertaining to a variable of mathematical function of the values of other variables. It is mostly employed when there is much data derived from historic data.

** Risk or Simulation model: ** The most commonly used technique is the Monte Carlo simulation which relies on activities simulated over a period and has a history to be researched on. The techniques are not easy as traditional deterministic approaches which tends to be more readily accepted rather than ranges to which confidence limits are attached (Jaggar et al., 2002). Here the name “model” may well be a misnomer, the approach is normally a simple Monte Carlo evaluation of a construction cost compiled using the estimators’ subjective perceptions of probability distributions for each of a set of reasonably independent sub-systems (Raftery, 1991).

** Resource based model: ** The currently dominant view of corporate strategy – resource-based theory or resource-based view (RBV) of firms – basis most concept on economic rent and the notion of companies having the ability to collect capabilities. Traditional strategy models such as Michael Porter's five forces model focus on the company's external competitive environment. In contrast to the Input and Output Model (I/O model), the resource-based view is grounded in the perspective that a firm's internal environment. "Instead of focusing on the accumulation of resources necessary to implement the strategy dictated by conditions and constraints in the external environment (I/O model), the resource-based view suggests that a firm's unique resources and capabilities provide the basis for a strategy. The business strategy chosen should allow the firm to best exploit its core competencies relative to opportunities in the external environment."

** Case- based reasoning (CBR) model: ** This involves the use of old solutions to problems to solve new problems (Singh, 2014). CBR frameworks have been produced of late for all areas of construction, namely architectural and or basic outline, term and expense estimation, development process, security arranging, offer choice making, choice of strategy and administration. The CBR methodology is like the master judgments that rely on the utilization of experience to take care of issues. According to Kim et al., (2004) the experts solve a problem using the procedure below: Observe the key qualities depicting an issue; Identify these qualities in past comparative issues they would say; and Estimating the heading of the new issue on the premise of the comparative experience issues with mental conformity.

A CBR framework, roused by the recalling of similitude’s in specialists' thinking, comprises of four sub-forms: Old cases which speak to encounter the framework obtained are put away for a situation base; when another case is displayed to the framework, the CBR framework recovers one or more put away cases like the new case as per the rate comparability; Users endeavour to illuminate the new issue by adjusting the recovered issue(s), and the adjustment is taking into account the contrasts between the put away issues and the new issue, unless the recovered old issue(s) is a nearby match and this recovered case most likely must be changed; and the new arrangement is held as a piece of the put away cases all through the test.

** Knowledge based model: ** Knowledge modelling is concerned with the standard specifications attributed to a specified product. Knowledge-based engineering is a process of computer-aided usage of such knowledge models for the design of products, facilities or processes. It employs models pertaining to knowledge based that has to be modelled. If the knowledge representation language enables to express both, then the knowledge model and the information model can be expressed in the same language (or data structure).

** Cost estimation model based on cost significant items: ** Estimating depending on cost significant items has its basis on historical data by employing historical data of completed projects. The cost of each element is subjected to the cost of projects similarly proposed. It can be said the basic strides in the advancement of a building estimation cost model in light of cost sensitive work packages are: Bills of quantity (BOQ); identifying similar projects and Computing the cost significant value factor (CSVF). According to Horner and Zakieh (1993) Cost Significant estimation usually does not pay attention to small works. Vilfredo Pareto’s principles explains that 80% of the effect of problems is instigated by 20 % of the causes. This standard principle is known as 80:20. Relating it to the construction sector, it can be said that 20 % of the work items can account for 80 per cent of the cost of the project.

#### 2.7.3 New Wave Models

The new wave models comprise artificial intelligent system (Neural network and Fuzzy logic) and other models (Environmentally and Sustainable development).

** Neural network Model: ** Numerous studies have made use of the neural network model for prediction and optimization. The research by Shtub et al., (2002) cited in Gunayin and Dogan (2004) postulated that the manufacturing sector is made up of ANN models. The neural network model is a computerised system technique that restructures the learning procedures of the mind of human. Neural networks are widely associated to numerous industries not including the construction sector. The operation of NNs to development has been broadly mulled over. Also, scientists have found the utilization of NNs to enhance the precision of expense assessing past that of the relapse models (Gunayin&Dogan, 2004).

Abbildung in dieser Leseprobe nicht enthalten

Fig 2.3: The architecture of the neural network model (Arafa and Alqedra, 2011)

Results from several analysts indicates that in comparing the relapse model to neural network, the later gives more accuracy than the former. One leeway of the neural network is the amount of yields and inputs which cannot be confined and the number of shrouded layers and the quantity of concealed neurons that are characterized. On the side of its demerits, a significant fact is the time spent in deciding the quantity of concealed neurons. Hegazy et al (n.d) recommended that one shrouded layer is adequate to produce a discretionary mapping in the middle of inputs and yields (Arafa and Alqedra 2011).

** Fuzzy logic: ** Functional systems must provide decision making abilities. Some also utilize classical convention logic, which will always provide affirmative or non-affirmative answers, meaning “white” or “black”, “no” or “yes”, “small” or “large”, etc. The answers employ accurate values. The basic logical operations employed are the AND, OR and NOT operators. The function of a classical set of values is characterised by either conforming to the set or not. Classical set theory operations are realized using classical logical operators among others Korol and Korodi (2010) indicated that the operation of classical theories can be achieved by using the classical logical operators based on this it could determine whether the values of the classical set will conform to the main data set or not. In many instances the set values do not present enough relevance needed, this is because correct answers cannot be given only through two fix values. It is often said that “vagueness and impreciseness are components of everyday life”. The idea behind the fuzzy logic theory is to replace the set of truth values with the entire interval practically to take a much more complex decision. A membership function depicts the extent of elements of membership between the intervals. In addition, the membership function engulfs itself in fuzzy logic which is represented in a general form.

The table (2.1) summarises cost estimation models identified in literature with their references indicated.

Table 2.1: Summary of cost estimation models

Abbildung in dieser Leseprobe nicht enthalten

### 2.8 GLOBAL UTILISATION OF COST ESTIMATION MODELS

Industry and Government practitioners in the US commonly used parametric or statistical cost model to engage a various form of analyses (Department of defence, 1999). Practitioners (users) lay emphasis on the preparation of proposals, evaluation and the cost of negotiations incurred. As well as the cycle time minimizes the adoption of statistical cost estimation. There was the need for organisations to collaborate with government to demonstrate the technique used in estimation to accomplish the task using parametric techniques. “In December 1995, the Commander of the Defence Contract Management Agency (DCMA) and the Director of the Defence Contract Audit Agency (DCAA) funded the Parametric Estimating Reinvention Laboratory (PERL) under the auspices of the Parametric Cost Estimating Initiative. Its purpose was to subject it to test by using the auspices of the parametric cost estimating initiative”. The aim was to examine parametric estimating techniques on proposals and provide the mechanisms through which others can have access to the techniques. The core mandate of the PERL was to test the technique on the actual proposals submitted this was done by developing a case study using the lessons learnt and the best practices identified as the bases, another mandate is to also identify the suitable technique which is more appropriate to be used for the estimate, moreover they established the formal guidance that need to be followed for future implementation, they also negotiate the type of estimating system to be used in the proposal. With regards to the study, 13 PERL were tested and implemented. The estimate encompasses all the major uses of specific element of cost. It was identified by the team that the use of parametric techniques provided an opportunity for an in-depth estimation and negotiation. In addition, the teams reported proposal preparation, evaluation and negotiation cost savings of up to 80 percent and reduced cycle time of up to 80 percent. The contractors with feedback from their Government team members, changed their estimating system policies and procedures to ensure data received are consistent. The Parametric Estimating Initiative closure report contained all remarkable practices for the implementation of these techniques.

During the early and mid-1990’s, both Industry and Government through the Parametric Cost Estimating Initiative (PCEI), accessed the ability to evaluate the ability of parametric estimating techniques and tools to support Government proposal cost estimating required to increase the utilization of historical data in the estimating process, increase estimate realism and reduce the costs associated with proposal preparation, evaluation and negotiation. As a result, contractors today generally use parametric techniques to improve their Government contracting practices, as well as the quality of their estimating. In the early 1960s, the U.S. Air Force, was dissatisfied with the bottom-up estimating technique used, a request was therefore sent to RAND Corporation to look for a better way of making program cost estimates. Based on the request, RAND developed the idea of a statistical estimating model where estimating relationships were found between projects or program costs and one or more technical parameters of the project or program. Cost estimating relationships (CERs) technique was the concept presented to the Air Force and concluded with a suggestion that one or more technical parameters might be useful for predicting project or program costs (Black 1984, Krieg 1979, Orczyk 1990). In Malaysia, most quantity surveyors and cost estimator’s estimates cost of projects using either expert judgement or algorithmic cost models, the mixed techniques produced more accurate results (Holm et al., 2005). Furthermore, utilisation and development of cost estimation models is in no exception in Nigeria and other developed countries because of the contributing factors to construction cost estimation (Ahuja et al., 1994).

#### 2.8.1 The Practice of Modern Estimation

Ashworth and Skitmore (1982) emphasized on the need of attaching great importance to any estimation method employed when predicting cost estimate. Ogunsemi (2006) affirmed that when 10 percent depicts the difference between the cost of the project and the actual cost incurred is a plus to estimation of cost and a sign of excellent estimation. It is quite unfortunate we cannot base realism on this fact proposed with the fact that, studies have been conducted into this issue with relevance to developing countries. Estimating the cost of projects largely dependent on various factors (Chua et al., 1999; Akintoye 2000; Morrison 2004; and Enhassi et al., 2007). A study by Wu and Cheng (2005; Lowe et al., 2006; Kim et al., 2005) and Marzok et al., (2008) on traditional methods to estimation explained that the importance in estimation model is to estimate cost of projects depicting a substitute to the current approach and reduce uncertainty to the best minimum level. At the design phase of a project, projects cost can be ascertained using cost models, however it turns out to be more difficult in ascertaining the objective in association with the building cost estimation (Ashworth, 2010). Ashworth (2010) emphasized on traditional building cost estimation strategy ending up ineffective in time, since undertakings have gotten to be bigger and refined. The traditional system of cost estimating has been substituted with computer-based systems that are simpler, saves time and more accurate. Computer based estimating models depend extensively on database from past project information for its development. The technique has significantly improved cost estimation process in the construction industry.

#### 2.8.2 The Need for Non-traditional Cost Estimation Models

The importance of using computerized estimation models cannot be underestimated (Kim, et al., 2005). At the stage pertaining to design, the cost of the project can be estimated using cost models. But it is turning out to be all the more difficult to accomplish the set goals in association with cost (Ashworth, 2010). Ashworth (2010) also emphasizes stated on traditional building cost estimation strategy has ended up ineffective in time, since undertakings have gotten to be bigger and refined. The traditional system of cost estimating has been substituted with computer-based systems that are simpler, saves time and more accurate (Ashworth, 1994). Computer based estimating models depend extensively on database from past project information for its development (Ashworth, 2010, Kim et al., 2005). The technique has significantly improved cost estimation process in the industry. In request to get a more precise building estimate, cost datasheet must be accurate and reliable (Singh, 2014).

### 2.9 BENEFITS OF COST ESTIMATION MODELS

A model represents a physical object usually smaller; models provide assistance to those mandated to provide the building cost in the bid of improving the performance. Raftrey et al., (1993) and Ashworth (1998) came up with cost model development with its common use to projects relating to construction prices has proved its usefulness in the advantages below; it provides accurate information through which meaningful decision can be made, Information accuracy of decisions can be ascertained hence bringing about confidence in the decisions made, information on cost can easily be ascertained and it contributes significantly to cost advice (Ashworth, 1988). In addition, other research also revealed that there exists a high-quality link between the technical and cost proposals. The data is well understood through the calibration and validation activities. Ashworth (1991) is of the view that early costing cannot be done in any other way except the use of cost models which in addition is much easier to handle scope, technical and performance changes.

### 2.10 THE BARRIERS TO COST ESTIMATING MODEL UTILIZATION

Although there are benefits to the utilization of cost estimation models at the initial stages of construction projects, there are still some barriers in implementing cost modeling in construction cost estimation. However available literature has identified some barriers that impede the utilization of cost modeling in cost estimation. Skitmore and Marston (1999) presented set-backs in estimation of cost modeling techniques at initial phase of the process subjected to the additional requirement needed to change the information pertaining to the design into process information. Formosa (cited in Bowen, 1993) affirms the minimal usage of cost modeling techniques in the initial stage is subjected to low level of understanding by the team responsible for designing of the process pertaining to construction. Ogunlana (1989) considers the estimators of design not to have the resolute skills to provide construction cost estimate as compared to construction contractors. Skitmore and Patchell (1990) and Bowen (1993) also confirmed the low level of understanding of “design cost estimators. Ashworth (1999) affirms that the cost estimators are not given adequate data whilst Ogunlana (1989) contended that, should the data be made readily available, unfamiliarity would act as a set-back to its utilization. These are affirmed by considering an earlier study by Fortune and Lees (1996). Additionally, misunderstanding of cost modeling techniques had been a clear contributor to the inability of quantity surveyors employing the said techniques. Ogunlana (1989) noted that” constraint of time is problematic and a set-back in relation to other modeling techniques pertaining to cost.” Bowen and Edwards (1998) ‘affirms that attention is directed towards construction resource effects on building design.”

#### 2.10.1 Lack of understanding

Briand et al., (1998) established various factors for the low implementation of model based-approaches of cost implementation. Formosa (1993) contended that the main objective behind the low level of utilization is the unavailability of adequate information. Cited in Bowen (1993), Ogunlana (1989) made mention of design cost estimators not having the needed skills to examine and provide the cost of materials. Skitmore and Patchell (1990) and Bowen (1993) made mention of the low level of understanding pertaining to construction professionals as the main cause of their inability to use other forms of techniques for cost estimation.

#### 2.10.2 Lack of data management

Ashworth (1999) affirmed that cost estimators should have the information needed to estimate cost, whilst Ogunlana (1989) contended that the” time constraints of cost modeling techniques; establishing them to consume more time in relation to other techniques Fortune and Lees (1996) also established some arguments in their initial studies trying to establish the link that existed pertaining to firm which is been contracted and the organization.

#### 2.10.3 Time constraints

Ogunlana (1989) affirms the set-backs of the modeling techniques: classifying them to consume more time. It is believed that the rigorous process quantity surveyors have to go through before arriving at the estimated cost at the early stages using these techniques is not time saving. Bowen and Edwards (1998) affirms that “time constraints’ have a great effect on the cost estimation at the early stages of the project when cost modeling approaches are used.

#### 2.10.4 Inaccurate outcome of models

Models when not properly calibrated would produce unreliable estimates. It is very important to secure a sound data so as to ensure confidence in the model developed. Due to the inability to secure sound data, Briand (1998) opined that this normally leads to errors in the model developed.

### 2.11 STRATEGIES FOR UTILIZING COST ESTIMATING MODELS

Strategies involves a plan of action designed to achieve an aim or provision of a solution to a problem. In as much as barriers exist to the utilization of non-traditional cost estimation models, there are however strategies which can be adopted to aid the use of cost models. Establishing implementation teams, government and constructors, organization of workshops, improvement of data management in firms, introduction of model development in higher institutions and enhancement of publicity on cost modelling are the existing strategies for implementing cost estimation modelling.

#### 2.11.1 Establishing Implementation Teams

Implementation project teams (IPTs) includes all players and facilitators pertaining to the evaluation, negotiation and implementation of a parametric model techniques (PCA, 1960). The IPT process ensures the understanding of sophisticated cost estimation techniques. A well-structured management process was essential in the implementation of the process (Department of defense, 1999).

#### 2.11.2 Government and Constructors

The endorsement of government and the contractor is key in the utilization of newer techniques in advanced countries, having researched to establish the benefits of these techniques, they opted for the techniques to aid in their estimation process (Department of defense, 1990). The collaboration between both parties’ breaded support from government to the successful execution of the project when the needed and relevant structures were put into place resulting in the utilization of new techniques for effective and reliable cost estimation.

#### 2.11.3 Organisation of Workshops

The estimation of cost and its relationships is not dependent on the jurisdiction of an expert in statistics; it requires the knowledge of the application of probability and measurement of statistics (Oblender 2001). A deeper understanding of the modeling technique is needed to facilitate its use for cost estimation (Ashworth 1999). A joint training is needed to provide detail understanding to users and the implementation team to breed understanding. It is therefore of importance for stakeholders to organize workshops for members to gain in-depth knowledge on available techniques for cost estimation modeling.

#### 2.11.4 Improvement of data management in firms

Skitmore (2001) opined that the availability of accurate and relevant data for the development of cost modelling improves its utilization. Oblender (2003) also suggested that such data aids in achieving value for money through the optimisation of cost.

Boussabaine et al., (1999) in their study emphasised that client’s information and data Kissi (2016) the Ghanaian construction industry lacks good historic data management to be used for early cost estimation and minimization of project risks. Unfortunately, the lack of available data inhibits the interest in developing cost models (Shehatto 2013), however the improvement of data management among construction firms could give a different environment where there would be enough data for cost model development.

#### 2.11.5 Introduction of model development in higher institutions

Modelling developing introduced in higher institutions has seen to gain more grounds which un-doubtfully have aided its understanding by breeding modelling professionals. It is believed that resolving its understanding at that level helped professionals with the understanding and utilization. Incorporating these modules into the higher institutions curricular has helped the recognitions and utilization of modelling techniques in developed countries (Department of defence, 1999). A research by Parametric Cost Estimation Association (Department of defence, 1999) suggested that models development introduced in the curricular of higher institutions has helped in its development and usage. Un-doubtfully, this is revealed in the utilisation of cost modelling in developed countries including USA, UK, Malaysia and Singapore.

#### 2.11.6 Enhancement of publicity on cost estimation modelling techniques

Persistent promotion of new product to the public is of essence it requires constant brief, education and practice for a particular product or technique to be understood and patronised. It is for this reason why non-traditional estimates should be promoted so as to obtain more attention by professionals. Earlier discussion in this study revealed the benefits of utilizing cost modelling techniques and hence the need to give it the necessary recognition. Bledsoe (1992) opined that the positive benefits of cost modelling demands a better way to communicate to the industry. The study by Smith et al., (2000) suggested that, if the positive effect of utilizing cost models is sold to the industry, it will maximize its appraisal.

### 2.12 DRIVERS OF COST ESTIMATION MODEL UTILIZATION

Drivers of cost estimation modeling are forces which create positive pressure to utilize cost estimation techniques by providing opportunities or threats which must be tackled in the estimation process. The driving forces of cost estimation models are: the need for readily accessible information at the design stage (Camargo, 2003; Scanlan et al., 2002), cost advice must be provided (Castagne, 2008), potential elimination of cost estimation errors (Castagne, 2008), increased pressure to provide reliable cost estimate (Ashworth 1994), growing demand for speed in estimation process(Ashworth 1994), difficulties in the identification of cost drivers in estimation (Skitmore, 1987 ), reduce the cost of preparing project proposals (Lamboglia et al., 2008; Asiedu & Gu, 1998). Provide leaders of projects the ability to access various options (Duverlie & Castelain, 1999; Watson & Kwakye, 2004). Crow (2002) argued the likelihood of serious overruns of budget and schedule, addressing the risk and uncertainties associated with project cost and schedule, improvement on organizations image are also driving factors for cost estimation modeling and essential tool because of the complexity in projects.

In order to understand the choice of certain strategies, it is prudent to critically consider the drivers that create the desire to implement these strategies. Ashworth and Briand (1998) suggested a number of motives to utilize cost models in estimation for construction projects. The drivers often demonstrate the forces that push firms and individuals towards implementing cost modeling into the estimation process. However meaningful and well-designed motives make it easier for firms to achieve the desired benefits. Below are the drivers for cost estimation model utilization;

#### 2.12.1 Cost Advice

Available literature suggests that having the knowledge of the probable cost of a project helps project participants to decide on other alternatives for the project at stake. Castagne (2008) opined that cost models are decision making tools used by cost advisors when forecasting building cost at the pre-tender stage. Cost modeling out-rightly provides cost advice by considering varous factors. Cost advice is very important at the early stages of a construction project because, clients rely on advice from professional regarding their cost commitment in order for them to make an informed decision (Ashworth, 2010).

#### 2.12.2 Cost information

Projects cost information at the initial stage of a project is seen as vital for construction projects (Camergo, 2003). It is seen as the document that guides the client or the stakeholder in knowing his or her financial commitment on the proposed project. In using Cost models for cost estimation, it is able to provide the details of probable cost of the proposed project which in turn aid in decision making (Newton et al., 1986). Cost Information provided prior to the beginning of construction projects is very vital for both the constructor and the client. Cost models serves as tools through which management make informed decisions pertaining to the area of the design pertaining to the building (Skitmore and Marston, 1999). Constructors need initial cost estimate to be able to budget and prepare good tender document even if drawings are not available for an average project. Early cost information is required by clients to assist in their decision-making process of embarking on a project.

#### 2.12.3 Reliable estimated budget

Skitmore (1999) identified that the need to provide a reliable cost estimate is one of the primary drivers for cost model utilization among quantity surveying firms and individuals. The USA parametric reinvention lab pointed out that statistical and parametric techniques and tools if properly utilized could result in reduced cost overruns in cost estimation and provide reliable estimated budget for the construction and other fields of science.

#### 2.12.4 Improve firm’s image

Saving the image of organization and individuals force them to act in a more responsible way. Skitmore (1996) opined that the growing number of stakeholders concerned about cost estimation issues as well as market competition puts pressure on firms to adopt tools and techniques in other to improve their services. The USA Air force (1960) been dissatisfied with the bottom-up method of estimation requested for new techniques to improve on their method of estimation for their construction and other related projects. This request brought on board RAND to develop the idea of statistical estimation model which caused a change in their estimation process (PCA 1960). Contractors today use statistical techniques to improve their cost estimation practices owing to the fact that cost modelling is seen as significant to construction cost estimation (Ahuja et al., 1994).

#### 2.12.5 Reduce cost overrun

Available literature suggests that the consideration of cost for a proposed project at the initial stages is more cost effective; cost models are however able to provide a considerable financial implication on a proposed project (Bousabaine, 1999). This helps clients to stay within budget. Previous research also shows that it is possible to stay within budget to avoid cost overrun and projects abandoning with cost modeling as tool used in forecasting project cost (Akintoye, 2000).

### 2.13 THEORETICAL FRAMEWORK

The underpinning theory of the study is the activity theory. Activity theory postulates that activities of human beings are affected by economic, social and cultural factors from internal and external sources. The aim for selecting activity theory was based on the fact that it highlights potential issues that affect the success or otherwise of a task or project such as cost estimation.

#### 2.13.1 Activity Theory

The activity theory is a cross investigative hypothesis for contemplating man as a performer in a socio- social, chronicled context. It is taking into account the thought of the double procedure of man and ancient rarities forming and being molded by societal and corporeal setting (Cole, 1996). This hypothesis is ancient rarity intervened and object-oriented activity (Vigotski, 1997). The connection between human operators and objects of environment is mediated by social means, tools and symbols. Activity theory may offer a conceptual framework to portray the structure, advancement and setting of tasks such as early cost estimation. As opposed to subjective science, which concentrates on the examination of the individual as a different substance, the unit of examination of activity hypothesis is human action. This activity can be portrayed as an action composed as an item that moves the action. The article orienteer’s states that people live in a reality which is focus in an extensive sense; the thing which constitute this reality have not quite recently the properties which are considered as item as showed by characteristic sciences however social and socially portrayed properties too. The hypothesis of activity places the underscore on the significance considering the setting, just as it is expected that the activity is essentially arranged in communal and corporeal setting. As per Nardi (1996), activity theory does not see cognizance as an arrangement of incorporeal subjective doings, and do not restrict it in mind; however, they see awareness as situated in regular exercise. In this way, human co-operations are considered as social, progressively composed, taking into account mediums. The theory of activity concentrates on the foundation of communications: the procedure of internalizing the apparatuses. The procedure comprises in an aggregate allocation of the instruments and the compelling utilization. The key components in internalization process inside of a group are: (i) specialists' discernment; (ii) social communications; (iii) nature of exercises. This procedure of internalization is connected to the design stage cost planning which is a key principle upon which the study is underpinned.

Vigotski (1997) figured how first thoughts regarding intervention of awareness by appropriating. Marxist thoughts regarding how intervene the work activity, the research expanded those thoughts to incorporate the use of psychological instruments to facilitate thought. The activity is at that point made out of a subject and an item interceded by an apparatus. The participant could be a man or a team occupied with the activity. The subject holds an item which gives it a specific bearing. The intervention could happen by the utilization of a wide range of sorts of apparatuses.

The activity theory accentuates on social elements and on co-operations in the middle of operators and their surroundings clarifies why the rule of device intercession plays a central part inside of the approach. As a matter of first importance, equipment defines the way human beings interface with the world. What's more, concurring to the rule of internalization or externalization, forming outer exercises eventually brings about molding inner one in forming inner ones. Secondly, instruments for the most part mirrors the experience of individuals who have attempted to take care of comparable issues at a prior time what's more, imagined or altered instruments to make it more productive. This encounter is collected in the auxiliary fitting ties of instruments (shape, material) and also concerning the learning in the usage of equipment. Tools are made and changed amid the improvement of the activity itself and convey with them a specific society and chronicled remainders from that development. In this way, the utilization of tools is a method for the aggregation and transmission of social learning. It impacts the nature, not just of outside conduct, additionally on eventual rational working of people (Bannon, 1997). Human activity is interceded by objects-inner and external. These apparatuses may be cues, dialects or machineries. These are made by individuals on impact control, above conduct. Obsolescent have a related society and history and perpetual quality that exist crosswise over time and space.

### 2.14 CONCEPTUAL FRAMEWORK OF THE STUDY

Any study is developed from related concepts or theories and provides reasons why the problem understudied exist by providing the correlation between variables (Jabareen 2009). It guides the researcher to decide which direction to follow in the study. Conceptual framework is an important element in research study which describes ideas to the study, together with the key variables (Miles and Huberman 1994). It provides the structure for the whole work based on literature. Jabareen (2009) further described conceptual framework as a network of linked concepts. It is crucial to establish a conceptual framework for the study because it aids in articulating the theoretical perceptions of the research. It also helps the researcher to explore ideas on how the research problem would be tackled.

The study is however based on a framework proposed by Mobley et al., (2006), which highlights the relationship between knowledge, drivers, and barriers and capturing a strategy for the implementation of cost estimation modeling.

It provided the foundation for cost estimation modeling which involves the adoption of cost modeling techniques in estimating the cost of building projects. This study emphasizes on the strategies for the utilization of cost models in estimating building projects which are significant techniques to improve the reliability and accuracy of estimation (Ashworth, 2004). The framework however proposes that, knowledge on cost models in-use, barriers of their utilization as well as the drivers creates the desire to implement the strategies that would enhance the utilization of cost models in estimation.

### 2.15 DISCUSSION ON VARIABLES OF CONCEPTUAL FRAMEWORK

The study is however based on a framework proposed by Mobley et al., (2006), which highlights the relationship between knowledge, drivers, and barriers and capturing a strategy for the implementation of cost estimation modeling. The framework however proposes that, knowledge on cost models in-use, barriers of their utilization as well as the drivers creates the desire to implement the strategies that would enhance the utilization of cost models in estimation.

#### 2.15.1 Background to building cost estimation models; Traditional and Non-traditional

Over the years, efforts have been made in identifying the cause-effect relationship that existed between the cost for building and the actual estimated cost that is used as parameters in estimating the cost of a building as far back as in the early 1950’s. Cost modelling is sometimes referred to as formation of system which components have some level of influence on the cost of a project (Holm, 2005). Based on history three phases of cost modelling have been identified. The first to consider is the first-generation model which was adopted from the “basic element of oriented cost of planning approach”. This model was introduced in in the late 1950’s and used extensively till the late 1960’s. The second model to be introduced was the second-generation model. This model was based on regression analysis which started around the mid 1970’s (McCaffer, 1975). The last model to be introduced was the third-generation model which had its bases on Monte Carlo Simulation technique techniques which started in early 1980’s (Touran, 1992). The models were further grouped into probabilistic and deterministic models, with the deterministic models; all the variables can be evaluated precisely. The other aspect which is termed probabilistic is also based on the assumption that even though some values are not certain calculations can still be made. Furthermore, cost estimation models can be grouped based on the characteristics they exhibit. With regards to this, two types of cost estimation models are identified and they include; traditional cost estimation model and the non-traditional cost estimation model. The traditional cost estimation model was generally based on quantities that is, models based on the resources used in the construction phase, models based on the building operational units and functional elements of the building operations. The second which is the non-traditional models that is to say, models comprising new techniques and practices; e.g. the experimental models, regression models and simulation models (Akintoye& Fitzgerald, 2000; Ashworth, 1988; Seeley, 1976) (Bledsoe, 1992; Flanagan &Tate, 1997; Mann, 1992; McCaffer et al., 1984; Newton, 1991; O’Brien, 1994).

#### 2.15.2 Selection of a Model

Generally, every model has an underlying assumption which can either be implicit or explicit. For accuracy, there is the need to clearly understand the underlying assumptions of the model (Boehm et al., 2000). On instances where the assumptions form part of the model itself, it becomes difficult to assess and evaluate the model however it is worth to note that cost is always uncertain. Therefore, efforts should be made to different between stochastic method that is a measure of uncertainty and the deterministic method; this method does not have a formal measure of uncertainty. Generally, selection of models could be based on the following the characteristics of data to be employed, quantitative, qualitative, large, small, general knowledge of the problem to be modelled, general knowledge about the boundary conditions of the model, errors that the model can generate, input and output targets and possible consequences, understanding of accuracy, reliability, validity, confidence and sensitivity of the model to be selected (Naar et al., 2012). Non-traditional model techniques are however selected based the aforementioned attributes thus enabling the quantity surveyor to be more knowledgeable before using the techniques. Upon adhering to these guidelines would create the enabling environment to choose right a modelling technique for accurate cost estimation.

#### 2.15.3 Knowledge on Cost Models

The knowledge on cost modelling techniques guarantees its utilization. Ashworth (1996) posits that the level of knowledge on a particular model technique increases its usefulness. The most crucial action to promote the implementation and utilization of cost estimation modelling is the development of the awareness of the building industry professionals about the benefits of the concept (Bowen, 1993). With the issue of cost estimation modelling for the construction industry being treated as a matter of urgency, professionals ought to take proactive actions to promote cost modelling within their various scope of practice and be responsive to the need for a better cost estimation. This is important because the concerns about the impact of preliminary cost estimation of building projects has led to the heightened awareness of the need for cost estimation modelling (Skitmore, 1990). The application of cost modelling techniques for estimation practices is however dependent on the awareness, knowledge and understanding of the consequences of its application. Adequate application of cost estimation modelling can be realized if designers are aware of the fact that the concept of cost estimation modelling is very crucial at the preliminary stage. With the demand for preliminary cost estimates, the most crucial decisions with regards to project budget are made during the design and preconstruction stages (Skitmore and Martinson 1999). In that way, modelling techniques will be used to help minimize errors in construction estimates. But the question that remains is “what the level of awareness of cost estimation modelling is?” This can be answered by considering the historical perspectives, the knowledge and the implementation of cost estimation modelling.

#### 2.15.4 The Barriers of Cost Estimating Model Utilization

The additional assumption presented in the cost estimation modelling techniques makes its utilization very difficult especially at the early stages as a result of the conversion of design information into process information (Skitmore and Marston, 1999). The low level of the usage of cost modelling technique at the early stage of projects is as a result of lack of understanding by the project team members (Formosa as cited in Bowen, 1993) the project design team members who are supposed to estimate cost for the project sometime lack the needed skill to do so (Ogunlana 1989).

According to Skitmore and Patchell (1990) and Bowen (1993) it is difficult to use “design cost estimators” due to lack of understanding on the part of contractors to accurately use model techniques. Moreover, Ashworth (1999) revealed that the data necessary for cost modelling techniques is mostly not available to the professionals with Ogunlana (1989) affirming that instances where data is available, it somehow becomes difficult to use because most estimators are not familiar with the model techniques. However, Fortune and Lees (1996) supported the argument in the study. Their analyses were based on the linkages that existed between the type of organisation and the type of cost modelling technique used. The writer further explained that the low level of the usage of the cost modelling technique is due to lack of understanding by quantity surveyors. According to Ogunlana (1989) time factor is also another constrain to the cost modelling technique. It is further explained that time constrain do not allow more attention to be paid to details with regards to the construction resources as well as the construction design. It does not give a “cove” between the existing techniques (consultant quantity surveyors) and the construction process they are attempting to model (Bowen and Edwards, 1998).

#### 2.15.5 Drivers for the utilization of Cost Estimation Models

Drivers of cost estimation modelling are forces which create positive pressure to utilize cost estimation techniques. It is considered as an opportunity or threat which must be addresses. Drivers must be able to provide a favourable condition for the implementation and acceptance of the model. It is important to put more focus on the drivers because the driving forces in cost estimation model are to be incorporated in the estimation as early as possible (Camargo, 2003; Scanlan et al., 2002) in other to provide cost advice in a more informed manner (Castagne, 2008) and also help to eliminate potential errors in cost estimation (Castagne, 2008), increased pressure to provide reliable cost estimate (Ashworth 1994), growing demand for speed in estimation process (Ashworth 1994), difficulties in the identification of cost drivers in estimation (Skitmore),reduce the cost of preparing project proposals (Lamboglia et al.,2008; Asiedu & Gu, 1998). Moreover, identifying the drivers provide more options for project stakeholder to consider make the best decision that will yield the maximum result (Duverlie & Castelain, 1999; Watson & Kwakye, 2004).

Crow (2002) argued that the likelihood of serious overruns of budget and schedule, addressing the risk and uncertainties associated with project cost and schedule, improvement on organizations image are also driving factors for cost estimation modelling. These are important function, because modern projects are often enormously complex.

#### 2.15.6 Strategies for Utilizing Cost Estimating Models

Strategies involves a plan of action designed to achieve an aim or provision of a solution to a problem. In as much as barriers exist to the utilization of non-traditional cost estimation models, there are however strategies which can be adopted to aid the use of cost models. Establishing implementation teams, government and constructors, organization of workshops, improvement of data management in firms, introduction of model development in higher institutions and enhancement of publicity on cost modelling are the existing strategies for implementing cost estimation modelling. A deeper understanding of the modeling technique is needed to facilitate its use for cost estimation (Ashworth 1999). It is therefore of importance for stakeholders to organize workshops for members to gain in-depth knowledge on available techniques for cost estimation modeling. Boussabaine et al., (1999) in their study emphasised that client’s information and data Kissi (2016) the Ghanaian construction industry lacks good historic data management to be used for early cost estimation and minimization of project risks. Unfortunately, the lack of available data inhibits the interest in developing cost models (Shehatto 2013), however the improvement of data management among construction firms could give a different environment where there would be enough data for cost model development.

### 2.16 CHAPTER SUMMARY

The chapter discussed a brief overview of cost estimation, cost estimation models and their impact on initial cost estimate of construction projects in the construction industry. It looked at empirical studies on cost estimation models for construction cost estimation. The various techniques used in predicting the initial cost of construction projects were also reviewed. These techniques were categorised into two; namely traditional and non-traditional estimating techniques. Furthermore, the concept of non-traditional cost estimating models, barriers, drivers and strategies of utilizing non-traditional cost estimation models were considered under the review of the study.

Finally, the conceptual framework for the study was developed.

## CHAPTER THREE

## RESEARCH METHODOLOGY

### 3.1 INTRODUCTION

The chapter states distinctively the design and method of the study, the sampling method or procedure, the demographic status under study and the method used in the collection of data apt to the research objectives. It establishes all the methods and strategies utilized in the study involving the design and the creation of questionnaires.

### 3.2 PHILOSOPHICAL ORIENTATION OF THE RESEARCH

This study has been greatly influenced by knowledge, values and philosophical queries of existence. Epistemology, ontology and methodology pertaining to philosophical matters should be presented clearly because any research instrument is formed by these choices (Christou et al., 2008). Jacob (1987) explained that epistemology is the knowledge we possess about the world we live in; ontology is defining nature’s reality; how nature is real; and methodology is describing the process by which we develop knowledge about the world. A researcher’s perspective on epistemology and ontology underlines and governs the theoretical perspective and the entire research process whether positivism or interpretivism (Marsh and Stoker, 2002). The perspective theory will be imbedded in the objectives of the study and prescribe the researcher’s decision on the methodology to be used, which will form part of the process used (e.g. questionnaire or interview). According to Marsh and Furlong (2002) the decision a researcher takes on this is very key, since “they redefine the mechanism used as all as its process”

#### 3.2.1 Ontological and Epistemological considerations

Ontology is defining nature’s reality; how nature is real. It can also be described as “the product of one’s mind” (Burrell et al., 1979). Thurairajah, et al., (2006) emphasized that ontology contributes significantly to decision making, establishing the realities, certainty and socially accepted ways which can only define clearly subjecting the human actors to examination. The view of reality based on the researcher’s assumption plays a vital role in all other assumptions and serves as a basis of various assumes, what is assumed here predicts the researcher’s other assumption (Hay, 2002). Two basic distinctions can be made here: firstly, the world isn’t real but rather socially and discursively made and hence rely on a particular time or culture-hence the expression anti foundationalism. Secondly, there is an independent real world upon which foundations life is built-hence the name foundationalism, which is in tandem with this research (Marsh et al., 2002). The researcher adopts objectivism at the ontological level (to ascertain the existence of a real world which is sovereign to our knowledge; it represents the living being theory) which explains the problems confronting the funding of infrastructural development which represents facts which are externally beyond the influence of the researcher.

Epistemology spells out the nature, validity and limits of inquiry (Rosenau, 1992) which Hughes et al., (1997) buttresses with his statement “How probable it is without difficulty to have access to the knowledge of this world?”, Sarantakos (2005) describe epistemology as a philosophical branch of research that controls the process of knowledge acquisition combined with its validation. Epistemology spells out the mechanism acquiring knowledge and its validation (Gall et al., 2003) which Babbie (1995) simply puts it as the science of knowing. Epistemology is a philosophical branch which concerns itself with what is right for an individual; positivism and interpretivism (Streubert & Carpenter, 1999). Approach to knowledge was largely dependent on the positivist approach. Scientifically, facts can be accumulated for the positivists (Osei, 2010). The study is of the view that identifies alternative funding methods needed for infrastructural development that must be conducted in a genuine way devoid of the researcher’s influence.

#### 3.2.2 Deductive and Inductive reasoning

A research approach can also be grouped into deductive or inductive. A testing theory also known as a deductive approach describes when a researcher strategically uses designs and theory being subjected to test a theory. Furthermore, Perry (2001) added; it is a study in which a conceptual framework is established, tested by observing empirically; particular instances are removed from the general influences. Empirical observation is mostly the mechanism used in Deductive research (Hussey et al., 1997).

The inductive approach also termed as theory building, includes the researcher gathering data in the bid of developing a theory (Saunders et al., 2003). Hussey et al., (1997) affirmed that it is concerned with the observation of tentative reality; laying emphasis that inferences are gathered based on specific conditions depicting the opposite nature of the deductive method, because it turns to move away from observing individuals to statement of general pattern or laws”. This research was administered by the usage of adequate available data which conforms with the use of quantitative methods for analyses (Travers, 2001) and draws on the deductive reasoning which emanates from certain views of conclusion (Burney, 2008).

### 3.3 RESEARCH DESIGN

The study design had its basis on Saunder et al., (2007) research onion idea as shown in fig 3.1. The research onion as retorted by Bryman (2012) is quite useful due to its adaptability to a wide range of research methodology. This research was therefore designed taking into consideration the philosophical view point, research approach, research strategy, time zone and data collection method all with the aim of achieving the research aim and objectives

Abbildung in dieser Leseprobe nicht enthalten

Figure 3.1: Research onion (Adopted from Saunder et al., 2007)

The research design provides a logical link between the research aim, framing of research questions, the data collected and then finally reporting on the data.

This stage defines how the research was carried out. The design has a number of different approaches including experiment, case study, action research, archival research, survey, and ethnography. Due to the fact that this research was done in an uncontrolled environment, experimental strategy was ruled out. Again, action research is characterized by the examination of trends, participation and intervention of the researcher, it was therefore not ideal for this work. The assessment of a single unit based on case study of a subject in the bid to set the main characteristics of the subject and to draw generalization would not be used for this work, since the attention is not on a single unit but on a sample representing the population. From the above, the survey strategy was best option adopted for this work. The survey strategy is in line with Cohen et al., (2005) observation, which stated that, positivism research uses the traditional strategies of surveys and questionnaires. Survey research strategy dwells on sampling a representative portion of the population. This strategy produces quantitative data which can be analysed empirically and it is consistent with the adopted research approach above. Ethnography which involves the observation of people, examining their cultural interactions and their meaning was also not suitable for this study.

### 3.4 RESEARCH METHODS

The research method used was a sole decision by the researcher. The researcher chooses the appropriate method to be used in the gathering and subjecting the data to analysis (Silverman, 2004).

#### 3.4.1 Qualitative Research

Qualitative research establishes the method to be used in establishing the meanings which are formed in accordance to the propinquity amongst the one investigating the topic (Denzin and Lincoln, 1998). Berg (2001) contributed that this type of research establishes purpose, impression, description, attributes, metaphors, representations and description of things.

#### 3.4.2 Quantitative research

Quantitative research is a research into problems of human being subjected to testing on a hypothetical basis comprising of various variables being subjected to measurement with numbers and examined with all the processes which are statistical in the bid to ascertain accuracy of the theory (Creswell, 1994). The study used the quantitative research method with the aim of gathering and analysing data in the bid to ascertain and confirm the research hypotheses

#### 3.4.3 Mixed method

When both qualitative and quantitative methods are used, it produces a rich understanding of the phenomena and the triangulation is explained whiles concentrating on the important research findings. The mixed method comprises of assumptions which are philosophical which directs and guides the gathering and analyses of data as well as the combination of both quantitative and qualitative utilization comes during the principal stage of the problem pertaining to the research ( Tashakkori &Teddlie, 2003).

### 3.5 DATA COLLECTION METHOD

To analyse the relationships of the variables emanating externally which have influences, they must be disengaged technically to be able to solve the research problem at hand from any external influences (Nenty, 2009). Primary and secondary sources were used during the research because of the credibility it would produce (Patton, 2002).

Both primary and secondary data were utilized. The primary data were accumulated by using questionnaires administered to respondents from the population of the study whiles the secondary data dwelt on literature from books, journals and academic papers.

### 3.6 RESEARCH POPULATION AND SAMPLING TECHNIQUE

#### 3.6.1 Study Population

The principle under which the researcher undertook the sampling was based on the fact that, data should be collected from every member of the population. This population is however divided into sub- techniques for the purpose of the study (Denscombe, 2010).

The population of this study was the quantity surveying Section division of the Ghana Institution of Surveyors (GhIS) Ghana. According to (Polit et al, 1998; Young, 2006) it is important for every study population to meet the needs of the study. However, the Ghana Institution of Surveyors serves as a professional body recognized under law to govern the activities of Quantity Surveyors. To be legally recognized as a practicing quantity surveyor, you must belong to the Institution of Quantity Surveying (Badu and Amoah, 2004). Geographically, the study was limited to Greater Accra and Ashanti Regions of Ghana. These locations were selected due to its high number of registered Quantity Surveyors, presence of consultancy firms and client organisations, where the target group (Quantity Surveyors) can be found and the proximity to the researcher.

#### 3.6.2 Sampling Technique and Sample Size Determination

Sampling is defined as “the use of accurate processes to select most suitable sample for the study” Kumekpor (2002). Though the universe is not subjected to change, the samples used can change on the objective of the research. Nonetheless, for a sample to come up a success there must be homogeneity in the population (Kumekpor, 2002). There are two main sampling techniques. These are the probability and non-probability sampling techniques. Kumekpor (2002) defines the probability sampling techniques as the one where randomness employs selection of respondents while the sampling method for non-probability is the one where randomness is difficult to achieve but rather respondents are selected based on convenience. On the other hand, Mustafa (2010) regards the non-probability sampling method to be based on personal judgment where the researcher “selects a targeted number of a unit of a whole sample either by purpose or otherwise largely dependent on the objects being enquired so that significant components can stand as the accurate attribute of the population added to the sample.

The purposive and snowball sampling techniques were adopted in respondent’s selection by the researcher. Judgmental selection of items of choice that is purposive sampling was used with the notion that they would provide the best perspective on the subject of interest. Thus, those specific perspectives are intentionally included in the study (Kothari, 2004). The snowball sampling technique helps one subject within the frame to provide the name of another subject which in turn gives the name of the third subject and follows in that order (Sauanderset al., 2007). Guided by the aim of the study, it was considered to be more effective if the sample for the study covered questions in public and private organizations. The questionnaires were sent to Quantity Surveyors within the geographical location of the study. The combined techniques of purposive and snowballing were used to reach the respondents.

According to Fisher (2010), students using the survey approach should use a much structured approach. Thus; the characteristics and properties of the subject matter must be well defined before the start of the research so that accuracy can be ascertained as well easy categorization. Defined by Kumekpor (2002) as the “sample the researcher is interested to understudy in his research”. The membership of the quantity surveying section of the Ghana Institution of Surveyors consist of 39 fellows; 357 professional members; 37 technician members and 43 firms; this sum up to 433 members and 43 firms (GhIS, 2016) taking into consideration office of Architectural and engineering services limited. Therefore, the sampling frame for this research were 44 Quantity Surveying operating firms. However, the size of the sample frame was subjected to the Kish (1965) as elaborated below

Abbildung in dieser Leseprobe nicht enthalten

**Calculating the sample size**

Abbildung in dieser Leseprobe nicht enthalten

However, an additional 10% of the total sample size (8 number) was added to cater for non-responsive respondents. This resulted to a total sample size of 89 respondents.

### 3.7 DATA COLLECTION INSTRUMENTthe

#### 3.7.1 Questionnaire Development

The primary information or data for this research were collected using questionnaire. Questionnaire is a set of queries prepared to request for information from respondents in a research. The data from the responses are normally coded into statistically data useful to the research topic (Roopa and Rani, 2012). The questionnaires for this research were sent out both in printed and electronic form to the target respondents. The questions had predetermined answers on ranked scales for respondents to answers the questions from. The design and administration of questionnaire is important since the responses can be used to make general statements about specific groups or people or entire populations (Roopa and Rani, 2012). The responses from the survey can become useless with inappropriate questions and bad scaling, as it may not accurately reflect the views and opinions of the participants. To check for the accuracy of the questionnaires sent out and to remove as much as possible all forms of ambiguity, it was pre-tested on ten (10) quantity surveyors. This was a useful check for the questionnaire to make sure issues were accurately captured in the questionnaires to elicit the needed information or answers from participants. The structured questions were designed to gather specific and appropriate data, which can be amended for statistical analysis. The questionnaires were administered using a face-to-face as well as mailing to few quantity surveyors who could not be reached easily for response. This admittedly gave a better response rate.

#### 3.7.2 Selection of measurement format

Most of the questions were constructed using different Likert scale depending on the objectives of the questions. The respondents were asked to indicate whether they agree to a set of given statements on cost estimates using the scale of 1-5 where: 1= Disagree; 2= Strongly Disagree; 3. Neutral; 4= Agree; 5= Strongly Agree.

#### 3.7.3 Pilot Test

Awanta and Asiedu-Addo (2008) stressed that designing a questionnaire which is reliable is possible because the answers are considered but may be inaccurate because it fails to reward the concept it intends to examine and fail to attain the objectives of the research. According to Cohen et al., (2003) citing Wilson and Maclean (1994) piloting is able to help in establishing the reliability, validity and practicability of the questionnaire because it serves among other things to ascertain the clarity of the questionnaire, provide feedback to validate text items and also make sure items in the data provides answers to the question. With the above concern in mind, the questionnaires were evaluated for content as well as face validity. The research instruments were pilot tested using some quantity surveyors in Greater Accra and Ashanti regions of Ghana. The rationale for the pilot testing was to establish the reliability and validity of the questionnaire. The process was conducted with care and attention given to comments generated from the respondents. The data collection instrument was then taken through a rigorous evaluation and necessary amendments made before the main data collection was conducted.

#### 3.7.4 Structure of questionnaire

The questionnaire was designed to achieve the research objectives. It consisted of an introduction which gave a brief description of the research, its purpose and objectives. The initial part of the questionnaire identified the demographics of the respondents, while the second and third part related to the main aspect of the work. In effect the second part covered the knowledge levels on various types of cost estimation models in use whiles the third part considered the barriers of cost model utilization as well as the drivers on cost models. The respondents were also requested to answer general information relating to their classification and experience in construction.

### 3.8 METHODS OF DATA ANALYSIS

Data analysis is the method extracting meanings out of any data set and drawing conclusions from them. The questionnaire data were analysed using descriptive analysis, one sample t- test, relative importance index, mean score ranking, factor analysis and reliability test from Statistical Package of Social Sciences (SPSS) Version 24.0.

Data obtained from administrative research instrument was taken through data cleaning and entry into a statistical program where questions and responses were coded. Data was collated into tables, graphs and inferential results where necessary. Collated data were then subjected to descriptive and inferential analysis to enrich the research conclusions. A test or examination is reliable if it generates parallel outcomes in recurrent administration when the characteristics under measurement doesn’t vary within the span between measurements, notwithstanding the test might be carried out by different persons and other methods of the test are employed (Field, 2009).

The Cronbach’s alpha measures basically squared correlation existing between the true and observed scores. The check for reliability, which is generally conveyed by Cronbach’s Alpha, is a famous method (Cronbach, 1951). Cronbach’s Alpha proposes that, respondents which are totally identical should have same score whiles very non-identical respondents should also get totally different scores from questionnaire evaluation (Field, 2009). Besides, “alpha is grounded in the ‘tau equivalent model “which has an assumption hence subjected to measurement by the similar latent traits on the same scale.” (Tavakol and Dennick, 2011). Reliability as indicated by Cronbach’s Alpha value increases from 0 to 1. Usually, a questionnaire having a value of about 0.8 (or anything above 0.7) is regarded to be reliable (Field, 2009). The interpretation is also done bearing in mind that normally distributed test scores exhibit high level of consistency (value) than those that are skewed either wards.

#### 3.8.1 Mean Score Ranking

There are different types of mean such as the geometric, harmonic weighted means etc…, and as such different formulas for calculating the means to be ranked. After the means have been computed for each set of variables, they are ranked in ascending or descending order for comparison. The following formula was used to deduce the mean score;

Abbildung in dieser Leseprobe nicht enthalten

#### 3.8.2 Relative Importance Index

The relative importance index was used to analyse some of the data by computing to deduce their rankings as below. Data was also analysed by ranking for example whether the respondents agreed or disagreed with the statement. The ratings of the statements by the respondents were placed against a five-point scale were converted and combined to deduce the Relative importance indices as follows:

Abbildung in dieser Leseprobe nicht enthalten

Where RII = Relative Important Index

X1= Number of respondent who answered “Strongly disagree”

X2= Number of respondent who answered “Disagree”

X3= Number of respondent who answered “Neutral”

X4= Number of respondent who answered “Agree”

X5= Number of respondent who answered “Strongly agree”

The index with the most significant relative importance was positioned 1st and subsequently followed by the next most important.

#### 3.8.3 Factor Analysis

Analysing factors consist of a number of statistical tools used to deduce the summation of dimensions of a large set of autonomous components or items (Reber and Reber, 2001). Factor analysis is a data reduction technique used to establish categories or groups whose variables are independent (Pallant, 2005). In the process of using factor analysis, the basic constructs are developed to establish the linkages in the correlation set. The purpose of the factor technique was to provide solutions to the interpretations of correlations amongst a number of measurements undertaking on a set of measurable entities. Analysing factors construct the underlining pattern of components to be independent helping analyse the existing correlation using regression analysis (Sanchez and Cahill, 1998). On the contrary, because of the simplification process, critics are of the notion that some vital information might be lost (Sanchez and Cahill, 1998).The coefficient of regression tends usually to be biased through factor analysis (Green, 1997).

Due to the associated weakness of factor analysis, the study was inclusively subjected to certain required precautions as stated below; Bartlett’s test of sphericity, which ensures “testing the null hypothesis whether the original correlation matrix is similar to the identity matrix” (Leung, 2011). He went on to establish a fact that should the test fail to reject the null hypothesis, then by definition the correlation matrix can be termed as an identity matrix whereby the hypothesized construct correlates between them does not correlate with others. The Bartlett’s test of sphericity is important if p < 0.050 for the data set (Leung, 2011). A significant required test is the Kaiser-Meyer –Olking’s (KMO) in the measurement of efficiency of variables falling between 0 and 1 respectively. The KMO serves a checker to correlations that are partial within variables that are insignificant (Field 2009). Components closer to 1 depicts that the correlated configurations are condensed, robust, and more distinct and such analysis should bring consistency (Field, 2009). “KMO value above 0.500 is adequate, 0.500 to 0.700 is average, 0.700 to 0.800 is good 0.800 to 0.900 is very good and value above 0.900 is considered superb (Kaiser, 1974)”. Nonetheless, factor analysis must not be used when the KMO is less than 0.500 and hence another approach should be employed. With aim of getting an accurate factor analysis, there must exist a form of correlation amongst them, but not to an extent of perfection (Field, 2009).

Added to the aforementioned, the eigenvalue plays a huge role when it comes to factor analysis. The eigenvalue criterion represents the criteria to establish the total numbers of factors to dispose or still maintain (Conway and Huffut, 2003). Child (1990) indicated that an initial eigenvalue higher than 1 is indicated as noble extraction. Nonetheless he suggested that factors which have it eigenvalue to be less than 1 should be excluded. There are two forms of factor analysis, they include the confirmatory factor analysis and the exploratory factor analysis. The confirmatory analysis is used to observe how sound a postulated factor module is. It is also used to determine how appropriate a factor model is in a new sample population or a different sample population. Suhr (2003) postulated that the confirmatory analysis can be used to identify parameters that defines the model. On the other hand, the exploratory factor analysis (EFA) is used to identify latent constructs. It is used to identify the minimum quantity of interpretable factors by examining the dimensional level of the measured technique. This helps in clarifying the existing relationships among group of variables (Klin, 2013). In using EFA the first to consider is to select the factor extraction method that is deciding on the amount of factor to retain or drop. After this is done, the factor rotation is selected and the result is then interpreted (Klin, 2013).

#### 3.8.4 Ethical Issues

This study was conducted to meet any research requirement available. It will first have to pass the requirement of the academic institution. To the source of data, such as construction companies, an introductory letter was sent from the school`s research committee to introduce the study and put forth the purpose and essence of the study. Any prospective source of information was first given a brief on the need for the study and was made to decide either to take part in the study or refrain from it. Respondents selected in this study were assured that the findings of the study were for academic purposes and seek to solve problems presented. Names of respondents were therefore omitted from the data collection instrument.

### 3.9 CHAPTER SUMMARY

The chapter described the philosophical point of the research, research strategy, and research design adopted for the study. It spelt out the mechanisms for selecting the sample, collection of data and ended with the provision and explanation of statistical procedures employed for sample selection, explained the mechanism used to design the data collection instrument and concluded with the procedures used in analysing the data.

## CHAPTER FOUR

## DATA ANALYSIS AND DISCUSSION

### 4.1 INTRODUCTION

In this chapter, details of the analysis and results obtained from literature and survey and all empirical evidence were discussed. In analysing the data set, content analysis was used on the information gathered from professionals in the construction industry. The respondents profile from the questionnaire were then described and the outcome of reliability analysis for all measured scales reported. Descriptive statistics was performed on the first section (A) of the research topic. This was then followed by mean score and one-sample t-test on individual measurement constructs of the component of cost models in-use for cost prediction (section B). Sections C and D comprising barriers and drivers of non-traditional cost estimation models respectively were analysed using principal component analysis (PCA). Lastly , relative important index (RII) and then one-sample t-test was used to analyse the magnitude of agreement to the strategies for the utilization of non-traditional cost estimation models. Accuracy in a survey result according to Rea and Parker (1997) and Aday (1996) stems from how great the rate of response is in a field survey, because it depicts the quality of the survey data. Eighty-three (83) questionnaires generated using the kish formula were distributed to firms within the scope of the study, however analysis was performed on seventy-one (71) questionnaires retrieved from the respondents.

### 4.2 RESPONDENTS CHARACTERISTICS

Respondent’s demographic information was analysed by descriptive statistics by employing the international Business Machine Statistical package for Social Sciences (IBM SPSS) Statistics version 24.0. This section of demographics is made of four individual questions which are position in organization of respondents, academic qualification, professional qualification and professional experience. The analysis and results in table 4.1 affirms the fact that the respondents are qualified enough and have the appropriate requisite experience to provide credible responses due to their level of experience in the construction industry and expertise in the field of this research.

Table 4.1 details the characteristics of the survey respondents with 56.3% representing Senior Quantity Surveyors and 43.7% representing Assistant Quantity Surveyor. According to the results from the survey, 32.4% of the respondents were MSc holders, 64.8% of the respondents representing the highest were BSc holders and the least represented of the population were HND representing 2.8% of respondents. Professional qualification of the respondents referring to Table 4.1 indicates that majority of the population were Members of GhIS which had a percentage of 94.4% of respondents and 5.6% of respondents were Fellow members of the GHIS. Finally, 49.3% of respondents had between 5-10 years of working experience, 25.4% of respondents had between 11-20 years of working experience and 25.4% of the respondents had more than 20years of working experience, this range of years were the least represented in the survey. From these results, it suggests that the data gathered was of good quality and this informs the credibility of the data collected from these professionals.

Table 4.1: Respondents characteristics

Abbildung in dieser Leseprobe nicht enthalten

**Source: Field Survey (2017)**

### 4.3 FREQUENCY OF PREPARING INITIAL COST ESTIMATE

Respondents were required to indicate the frequency at which they prepare initial cost estimate to clients. Analysis helped to determine how conversant respondents were to the preparation of initial cost estimate. Referring to the bar chart (fig 4.1), it is observed from the ranking that majority of respondent which is 47.9% of the population with a frequency of 34 respondents ‘often’ prepared initial cost estimate. The second ranked is ‘More often’ representing 26.8% of the population with a frequency of 19 respondents. Ranking third was ‘less often’ representing 14.1% of the population and 10 respondents. The least is ‘Not often which had frequency of 8 and a representing 11.3% of the population.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4.1 Frequency of Preparing Initial Cost Estimates

### 4.4 KNOWLEDGE ON COST MODELS IN-USE FOR COST PREDICTION

Respondents were asked to rank the various cost estimation models in use for cost prediction in the industry. Extensive literature was reviewed based on the cost estimation models which exist. The five-point Likert scale rating was used in establishing the importance of these indicators with a success criterion set at a mean value greater than 3.0. The Likert scale was designed with a scale of ** 1= Not knowledgeable; 2=Less knowledgeable; 3 = Aware of but do not know much about it; 4=Knowledgeable; 5= Very knowledgeable. ** Giving the respondents a choice to choose from these scales. Mean score ranking was used to examine the variables ranking them according to the most significant. The analysis was to determine the importance of the variables henceforth “one sample t-test” was used to establish the significant of the variables. This is usually used to determine whether the sample mean is relatively different from the hypothesized mean. In using the sample t-test the hypothesis is normally set as;

Abbildung in dieser Leseprobe nicht enthalten

The Ho in the equation means “null hypothesis” Uo is also the population mean and Ha is the alternative hypothesis. Usually in one sample t-test the mean of the test group as well as the level of freedom of the test group is usually an approximation of the sample size. Again the ‘P’ value which is usually the test value considered are described in the test (Ahadzie, 2007). The mean for the standard error and standard deviation have been indicated in the table below. Uo was fixed at an appropriate level of 3.0. The importance level was also fixed at 95% in agreement with conventional risk ranks. Centring on the five-point Likert scale rating, a variable thought to be dire if it had a mean of 3 or more. Most variables had a standard deviation less than one (1) demonstrating there was variability in the respondents understanding of these variables.

#### 4.4.1 Reliability of data analysis

To check the consistency of the results, it was appropriate to use Cronbach’s coefficient alpha, this analysis was conducted on the variables of the research. “Cronbach’s coefficient alpha reliability test” is rooted in how consistent the internal test is. This can be indicated by the degree of homogeneity of the variables by the measurement set. In table 4.2, the 13 variables had an alpha coefficient of 0.850 indicating that, the items relatively had a high internal consistency.

Abbildung in dieser Leseprobe nicht enthalten

#### 4.4.2 Data analysis

From table 4.3, not all the 13 constructs had a mean value above the set criterion of 3.0. The highest and lowest means of each construct with their standard deviation are as follows:

Traditional models had Bill of quantities as its highest mean among its indicators as 4.83 and a standard deviation of 0.447 ranking as the first indicator with Story Enclosure method as its lowest mean having 2.77 with a standard deviation of 1.34, ranking fifth.

Non-Traditional models had Regression models as its highest mean among its indicators as 3.13 and a standard deviation of 0.893 ranking as the first indicator with Fuzzy logic as its lowest mean having 1.45 with a standard deviation of 0.732, ranking eighth. In using Standard deviation, it checks the internal consistency in the data collected in order to generalize the results. Standard deviations value of less than 1.0 indicate consistency in agreement among the respondents to the reported results (Stevens, 1996). In the SD column, not all the values are less than 1.0. These suggested that respondents gave low rating to some of the cost models according to their level of knowledge.

Table 4.3 Cost Estimation Models In-Use for Cost Prediction

Abbildung in dieser Leseprobe nicht enthalten

**Source: Field Survey (2017)**

#### 4.4.3 Data analysis

Moving from left-to-right from table 4.4, the following are observed, *t* -value ("**t**" column), degrees of freedom ("**df**"), and the statistical significance (*p* -value) ("**Sig. (2-tailed)**") of the one-sample t-test. If p < 0.05, it can be concluded that the population means are statistically significantly different and if p > 0.05 the difference between the sample-estimated population mean and the comparison population mean would not be statistically significantly different. The upper and lower values (confidence interval) which is 95% confidence interval for μ, you can be 95% self-assured that the interval contains μ. In other words, 95 out of 100 intervals will contain μ upon recurrent sampling. However, all the variables had a positive p-value.

Table 4.4 One-Sample Test

Abbildung in dieser Leseprobe nicht enthalten

**Source: Field Survey (2017)**

#### 4.4.4 Discussion of Cost Estimation Models in-use for Cost Prediction

Cost estimation models for cost prediction involve several techniques. The study however sought to find out from respondents the most dominant techniques use for cost prediction in the Ghanaian construction industry. The traditional and non-traditional models were the techniques respondents responded to. These were ranked based on the frequency of their utilization as indicated in table 4.3. According to the study, Bill of quantities, Superficial Floor Area method and Unit method which are ranked first, second and third respectively were the three dominant traditional methods for cost estimation in Ghana. This is in agreement with the study by Ashworth (1999) and Bari (2010) which postulated that the dominant traditional cost estimating tools used in cost estimation are bill of quantities Superficial floor area and the Unit method. A study by Alorgli (2010) in Ghana on cost estimation techniques opined that the most utilized cost modelling technique is the Storey enclosure which is in disagreement with this current study. This however suggest that a particular technique can change overtime paving way for new and modern techniques to be used.

Furthermore, the study proceeded with the non-traditional techniques which ranked Regression, Elemental floor area and Resource based techniques as first, second and third respectively. Regression models are the frequent non-traditional model used for cost prediction by most cost estimators, this is because it easily understood and simple to use (Ashworth, 1999). Among the non-traditional cost estimation models according to the research, regression models are the well-known models respondents are aware of. Regardless of its awareness level, none of the respondents utilizes it for cost estimation.

### 4.5 BARRIERS TO THE UTILIZATION OF NON-TRADITIONAL COST ESTIMATION MODELS

Respondents were required to tick in the spaces provided in the questionnaire in relation to the ranking provided, ** where 1 - Not Significant; 2 –Less Significant; 3-Moderately Significant; 4 -Significant; 5-Very Significant. ** Due to the large number of reliant variables involved in the study, there was a higher likelihood that some variables have their end results in the same underlying effects is high. Factor analysis technique was adopted to do the data reduction.

**Factor analysis**is a technique that is used to reduce a large number of variables into fewer numbers of factors. This technique extracts maximum common variance from all variables and puts them into a common score. It is recommended by literature that the value of the KMO value must be greater than 0.5. From table 4.5, the KMO value is 0.513 which is greater than 0.5 suggesting that the factors are appropriate for analysis. From table 4.6 the alpha coefficient for the 23 items is 0.784, suggesting that the items have relatively high internal consistency.

Table 4.5 Reliability Statistics

Abbildung in dieser Leseprobe nicht enthalten

Table 4.6 KMO and Bartlett's Test

Abbildung in dieser Leseprobe nicht enthalten

The extraction of variables is indicated in table 4.7. According to the table not all variables qualify to be included in the factor analysis because some of the variable had their extract value lower than 0.5, but the rule suggests that only extract value which is more than 0.5 are significant which proves its validity and essence. Models utilization is best for large construction projects obtained the highest extraction factor of 0.916 which comparatively is the closest to the initial set factor of 1. The following were not significant; ‘The degree of sophistication is seen as too needless for an average project’, ‘Lack of comfort in using model based estimating tools’ and ‘Lack of industrial patronage ‘with the least extraction factor of 0.266, 0.374 and 390.

Table 4.7 Communalities

Abbildung in dieser Leseprobe nicht enthalten

**Extraction Method: Principal Component Analysis.**

Guttman Kaiser Rule suggests that factors to be retained should have an eigenvalue greater than one (1). On the other hand, “Cattell scree test” argues that all components after those starting the elbow should be excluded. Both Catell scree test and the Guttman Kaiser rule were used. In applying this, the principle guiding the extraction of components suggest that 4 components are to be extracted for barriers to the utilisation of non-traditional cost estimating data set. This is shown in figure 4.2 where the components extracted were having an eigenvalue higher than one (1).

Table 4.8 Total Variance Explained

Abbildung in dieser Leseprobe nicht enthalten

Extraction Method: Principal Component Analysis.

Abbildung in dieser Leseprobe nicht enthalten

Fig 4.2 Scree plot

It can be indicated from the table that, eigenvalue factor of 0.5 is significant since the close an eigenvalue towards 1 the more significant that variable is. Therefore, B1, B4, B5 had eigenvalues below 0.5 which means those factors were insignificant.

#### 4.5.1 Data Analysis

According to table 4.9, four characteristics were generated after the twenty-three variables were factor analysed. Components were given distinctive names grounded on the connection that occurred among them. The four factors are being; Inefficient Techniques, Inadequate training of Professionals, inaccuracy of Cost data and Lack of Understanding.

Table 4.9 Rotated Component Matrixa

Abbildung in dieser Leseprobe nicht enthalten

In other to define the number of extracted variables, Guttman Kaiser Rule was used. The rule suggests that factors to be retained should have an eigenvalue greater than one (1). In applying this, the principle guiding the extraction of components suggest that 4 components are to be extracted for barriers to the utilisation of non-traditional cost estimating data set. This is shown in table 4.8 where four components extracted were having an eigenvalue higher than one (1).

#### 4.5.2 Discussion of Barriers to the Utilization of Non-Traditional Cost Estimation Models.

According to table 4.9, four characteristics were generated after the twenty-three variables were factor analysed. Each component was given a distinctive name grounded on the connection that existed among them. The four factors are as follows; Inefficient Techniques, Inadequate training of Professionals, unavailability of Cost data and Lack of Understanding. These factors are however discussed below;

##### 4.5.2.1 Component 1 – Inefficient Techniques

The first component extracted accounted for 30.05% of the total variance. The component was termed inefficient techniques. Generally, the techniques for effective utilization of non-traditional cost estimating models remains inadequate due to the extent of technological and mathematical knowledge that are required in its usage. Hence the need for QS to update their knowledge, since any technique that would be adopted influence the level of accuracy of a given cost model’s results. Although, quantity surveying professionals rarely use technique they are unfamiliar with bearing in mind the needed results (Skitmore, 2010). Ashworth (2010) opined that the lack of interest in developing cost models by professional add up to the non-application of newer techniques for cost estimation. It must be noted that, with the advent of modern designs and its accompanying complexities makes the use of non-traditional cost models more useful, hence, to achieve the expected clients’ satisfaction there is the need for adaptation of the required techniques.

##### 4.5.1.2 Component 2 – Perception of model techniques

The second component extracted accounted for 16.93% of the total variance and termed perception of model techniques. Perception in this case has to deal with how practitioners regard the use of these cost models. Such perceptions are drawn from the misunderstanding, and inadequate data in cost model development. On a whole some critics of non-traditional cost models hold a strong perception that these models are just for academic exercises, more of theories while in practice it might not yield the needed results. In addition, the techniques are considered not defining all the cost drivers and may be only useful for large projects. Bowen et al., (1982) considered outcomes of models’ techniques as inefficient owing to the absence of accurate historical data for its development. Hence the need for a rigorous validation processes where the weaknesses pertaining to these models will be known. This will help stakeholders to make the necessary adjustment in it usage.

This however shows the existing perceptions on cost modelling by quantity surveyors with respect to the application and accuracy and of their outcome.

##### 4.5.1.3 Component 3 – Unavailability of cost Data

The three variables which were loaded on the third component are model development requires large data to ensure confidence, lack of reliable cost indices and unavailability of cost data. The third component extracted accounted for 9.91% of the total variance and this has been termed unavailability of cost data. The availability of cost data facilitates the development of cost models. Cost models require historical cost data to ensure its development (Skitmore, 1998; Kissi et al., 2017). The results of the model highly depend on the quality of the available data. Aloysius (2010) indicated that there is lack of data management among firms in the Ghanaian construction industry. Lack of data management in firms makes it difficult to acquire data for the development of cost models. Corbett and Rowley (1999) stated that the main service of a professional Quantity Surveyor (QS) is to forecast to the clients the probable cost of their proposed project. However, the accuracy of the cost model depends on the quality of cost data used in preparing the cost forecast.

##### 4.5.1.4 Component 4 – Lack of Understanding

This component accounted for 7.41% of the total variance. The variable which loaded into this factor are lack of understanding on cost modelling concept, unfamiliarity of cost model techniques by professionals, resistance to change and Unstable inflation rate on the Ghanaian market. Understanding the concept of cost modelling significantly contributes to its utilization (Ashworth, 2004). Bowen contends that the primary reason for the low utilization of cost estimation modelling techniques for cost estimation is due to lack of understanding. It is worth to note that unfamiliarity and lack of understanding of model techniques are barriers to model utilization. The studies revealed that quantity surveyors mostly use techniques they are familiar with, lack of interest in adopting new techniques and resistant to change were among the barriers for model utilization.

Bowen contends that the primary reason for the low utilisation of cost estimation modelling techniques for cost estimation is due to lack of understanding. Fortune et al., (1996) posited that unfamiliarity and lack of understanding of model techniques are barriers to model utilisation. The studies revealed that quantity surveyors mostly use techniques they are familiar with, lack of interest in adopting new techniques and resistant to change were among the barriers for model utilisation in the Ghanaian construction industry.

### 4.6 DRIVERS FOR UTILIZING NON-TRADITIONAL COST ESTIMATION MODELS

Respondents were required to tick in the spaces provided in the questionnaire in relation to the ranking provided, ** where 1= Disagree; 2= Strongly Disagree; 3. Neutral; 4= Agree; 5= Strongly Agree. ** Due to the large number of reliant variables involved in the study, there was a higher likelihood that some variables have their end results in the same underlying effects will be high. Factor analysis technique is adopted to do the data reduction. This type of analysis usually requires a large sample size before it can be stabilized. The analysis also depends on the correlation matrix of variables. The sample size indicated its reliability to employ the factor analysis approach. In Statistical Package for Social Sciences (SPSS), the Kaiser-Meyer-Olkin measure of sampling adequacy (KMO-test) is a measure of the shared variance in the items (Beavers

*et al.,*2013). SPSS is used to check inter-correlation using the KMO test and Bartlett’s test of spherity whilst multi-collinearity is checked by the determinant of the correlation matrix (Osei-Hwedie, 2010). Reliability statistics of results were checked by using the Cronbach’s Alpha. Cronbach's alpha is a measure of internal consistency reliability, that is, how closely related a set of items are as a group. Composite reliability measures indicators with the idea that they have different loadings. It is recommended by literature that, the value of the KMO value must be greater than 0.5. From table 4.10, the KMO value is 0.592 which is greater than 0.5 suggesting that the factor is appropriate for analysis procedure. From table 4.10, the alpha coefficient for the 17 items is 0.893, suggesting that the items had relatively high internal consistency.

Table 4.10 Reliability Statistics

Abbildung in dieser Leseprobe nicht enthalten

Table 4.11 KMO and Bartlett's Test

Abbildung in dieser Leseprobe nicht enthalten

Table 4.12 shows the extraction of the variables. Kim and Mueller (1978) stated that all variables are adequate to be factor analysed if none of them had an extract value less than 0.5. According to the variables obtained, they were significant and therefore validated its inclusion and essence

Table 4.12 Communalities

Abbildung in dieser Leseprobe nicht enthalten

**Extraction Method: Principal Component Analysis.**

In defining the number of extracted variables, Guttman Kaiser Rule was used. The rule suggests that factors to be retained should have an eigenvalue greater than one (1). In applying this, the principle guiding the extraction of components suggest that 4 components are to be extracted in other to aid in the management innovations of the data set. This is shown in table 4.13 where four components extracted were having an eigenvalue higher than one (1).

Table 4.13 Total Variance Explained

Abbildung in dieser Leseprobe nicht enthalten

Extraction Method: Principal Component Analysis.

Table 4.14 Rotated Component Matrixa

Abbildung in dieser Leseprobe nicht enthalten

#### 4.6.1 Analysis and discussion of Extracted Components

According to table 4.14, four characteristics were generated after the seventeen variables were factor analysed. Each construct or component was given a unique name based on the relationship that existed among them. The four factors are as follows; Efficient Cost Estimation, Cost Advice, Risk Management and Improved Estimation Process. Following are the discussions on the various factors;

##### 4.6.1.1 Component 1 – Efficient Cost Estimation

The first component extracted accounted for 42.20% of the total variance. The component has been termed efficient cost estimation. The variables loaded into this component are increased accuracy, reliable estimated budget, improved estimation process, efficient cost estimating, cost control in construction projects, more convenient cost analysis, increase customer satisfaction and Cost reduction in preparing project proposals. Efficiency in estimation is very vital since initial estimated cost when relied on can make or mar a whole project. It is of importance to use techniques and methods that best provides confidence in an estimated project. From the analysis, it is concluded that Efficient Cost Estimation is the major driver for utilising non-traditional cost estimation model. Ashworth (1998) posited that a Cost model when developed is able to provide efficiency during the construction process. Ferry et al., (1999), opined that in order to avoid inaccurate estimates and estimation problems, cost should be modelled to evaluate the design. Efficiency was also revealed in Fortune (1996) which stated that among the traditional and the non-traditional estimation methods, the non-traditional methods have the highest percentage of accuracy reliability and value. Having efficient cost estimation methods helps in building a strong and loyal client base and helps to make the process transparent which strongly reinforces the competence of the professional.

##### 4.6.1.2 Component 2 – Cost Advice

The two variables which were loaded on the second components were Elimination of cost estimation errors and Project leaders and stakeholders’ ability to consider more options. The third component extracted accounted for 15.03% of the total variance and this has been termed cost advice. Cost models provide information on probable cost of a project. The initial cost provided is relied on by prospective clients to make informed decisions; this decision is however centred on the cost information provided in the cost model. This implies that advice on cost to a particular project is dwelt on a good cost estimated model. A study by Castagne (2008) suggests that cost models are decision making tools used by cost advisors to forecast building cost at the pretender stage. Cost modelling helps in the provision of cost advice and decision making in a more reliable and informed manner by considering variable factors (Boussabaine 1998).

Jagger et al., (2012) however postulated that Cost models provide a decision aid for the client which if applied could be used at the briefing stage of projects in evaluating the affordability of alternate options. The above discussion postulated that, when professionals provide cost estimates through the use of cost models, there exist the tendency of providing expert advice on cost to the client for consideration.

#### 4.6.1.3 Component 3 – Risk Management

This component accounted for 11.78% of the total variance and encompasses four variables. The variables which loaded into this factor are risk management in estimation, cost information within the design process, identifying major cost drivers in estimation Provision of reliable cost estimate. The Identification and quantification of the impact of risk on cost as well as the methods employed to minimise risk is one of the most important factors worth considering at the initial stage of every construction project (Fidgen, 1999). Risk is mitigated in the midst of reliable prediction, additionally it is possible to estimate likely totals for the risk related to the project as a whole thereby giving the client a true perspective of the total risks estimated in a project. Again (Fidgen, 1999) maintained that risk estimated in a project is performed specifically to identify the likely of contingency fund required for a project. This means that the scope of a project is well defined if a project is well estimated. Cost estimators should consider risk management as one of its drivers owing to the fact that these identified risks and uncertainties helps management procedures that eliminates risks in the provision of cost estimates (Skitmore, 2008).

##### 4.6.1.4 Component 4 – Improved Estimation Process

The fourth component extracted accounted for 9.71% of the total variance. The component has been termed Improved Estimation Process. The variables loaded into this component are Cost advice in an informed manner, improved organizations image and speed in estimation process. When cost estimations provided to clients are of the best results, it helps save the integrity of the organization and professionals of the firm. Ashworth and Skitmore (1982) emphasized on the need of attaching great importance to any estimation method employed when predicting cost estimate Traditional methods to estimation by Wu and Cheng (2005); Lowe et al (2006) and Marzok et al, 2008 affirmed the huge importance of estimation model to estimate cost of projects depicting a substitute to the current approach and reduce uncertainty to the best minimum level (Kim, et al., 2005). At the design phase of a project, projects cost can be ascertained using cost models. But it turns out to be more difficult in ascertaining the objective in association with building cost estimation (Ashworth, 2010). Ashworth (2010) emphasized on traditional building cost estimation strategy ending up ineffective in time, since undertakings have gotten to be bigger and refined. The traditional system of cost estimating has been substituted with computer-based systems that are simpler, saves time and more accurate. Computer based estimating models depending extensively on database from past project information for its development has significantly improved cost estimation process in the industry.

### 4.7 STRATEGIES FOR UTILIZING NON-TRADITIONAL COST ESTIMATION MODELS

Respondents were tasked to rank in their opinion, how these eight identified strategies for utilizing non-traditional cost estimation models from literature could facilitate the utilization of cost estimation models in the Ghanaian construction industry. Respondent was required to tick in relation to the rankings provided, where ** 1= Disagree; 2= strongly Disagree; 3. Neutral; 4= Agree; 5= Strongly Agree. ** The data were analysed by ranking the strategies for utilizing non-traditional cost estimation models to show those which respondents deemed most essential to the least indispensable. The results gathered were the collective responds from Consulting firms and Construction firms. Based on the information collected, relative importance indices (RII) of the respondents was computed to deduce their rankings. The analysis was further improved by using one-sample t-test to check the importance of the variables. The one sample t-test is usually used to know if a sample mean is considerably diverse from a hypothesized mean.

From table 4.15, observation made after the analysis revealed that, ‘Improvement of data management in firms’ had RII value of 0.961 and was 1st ranked. ‘Involvement of key stakeholders in cost estimation modelling study’ was ranked 2nd with RII value of 0.955. ‘Introduction of cost estimation modelling development in higher institutions’ was ranked 3rd with RII value of 0.927. The fourth ranked variable was ‘Enhancement of publicity about cost estimation modelling benefits’ with RII value of 0.904. ‘Organize cost modelling workshops for QS professionals’ was ranked 5th with RII value of 0.896. ‘Clarifying the perceptions about cost models’ was 6th ranked with RII value of 0.879. ‘Development of framework for cost model utilization’ was 7th ranked with RII value of 0.839. The least ranked (8th) was ‘Government support’ with RII value of 0.741.

Table 4.15: Strategies for Utilizing Non-Traditional Cost Estimation Models

Abbildung in dieser Leseprobe nicht enthalten

**Source: Field survey (2017)**

Moving from left-to-right from table 4.16, the following are observed *t* -value ("**t**" column), the degrees of freedom ("**df**"), and the statistical significance (*p* -value) ("**Sig. (2-tailed)**") of the one-sample t-test. If p < 0.05, it can be concluded that the population means are statistically significantly different and if p > 0.05 the difference between the sample-estimated population mean and the comparison population mean would not be statistically significantly different. The upper and lower values (confidence interval) which is 95% confidence interval for μ, you can be 95% self-assured that the interval contains μ. In other words, 95 out of 100 intervals will contain μ upon recurrent sampling.

Table 4.16 One-Sample Test

Abbildung in dieser Leseprobe nicht enthalten

#### 4.7.1 Discussion on strategies for the Utilization of Non-Traditional Cost Estimation Models

According to table 4.15, five key strategies were generated after the mean score ranking test on the eight variables. These factors including; Improvement of data management in firms; Stakeholders involvement in the study of cost modeling; Introduction of cost modeling development in higher institutions; Enhancement of publicity on cost modeling benefits; and Introduction of cost modeling development in higher institutions are discussed below.

**4.7.1.1** The first significant strategy ranked by the respondents was; Involvement of key stakeholders in cost modelling studies. Generally, the acceptance of a technique or method for a professional practice is widely engineered by the key members of the professional body. This means that if major stakeholders in the industry pushes for the introduction of new and best approaches in solving estimation issues, there would be a way for the introduction of non-traditional techniques for cost estimation. A study by Ashworth (1999) suggested that the issue of accuracy in cost estimation is paramount because the success and failure of every construction project lies in the estimated budget obtained by the client. It therefore proposed that, stakeholders in the industry should adopt innovative techniques to provide accurate estimate for successful completion of construction projects. This is however, a wakeup call on the stakeholders to stand for change to the old system of estimation. The findings of this research affirm to the findings by Ashworth confirming that if key stakeholders explored the outstanding benefits of cost modelling as revealed in this study and promote its study and development among members; estimation accuracy could be achieved.

**4.7.1.2** Improvement of data management in firms ranked second by the respondents.

Cost modelling involves mathematical and algorithmic representation of cost for a proposed project, however this representation of cost is largely dependent on historical cost data. It is therefore worth noting that, without a strong management of data among firms, the quest for utilization of cost modelling techniques cannot be realised. Skitmore (2001) opined that the availability of accurate and relevant data for the development of cost modelling improves its utilization. Moreover, Oblender (2003) also suggested that with the availability of historical data, cost models can be developed to help achieve value for money through the optimisation of cost.

In a similar study by Boussabaine et al., (1999) emphasis was made to client’s information and data availability are the important factors required to develop cost models. According to Kissi (2016) the Ghanaian construction industry lacks good historic data management to be used for early cost estimation and minimization of project risks. Unfortunately, the lack of available data inhibits the interest in developing cost models (Shehatto 2013), however the improvement of data management among construction firms could give a different environment where there would be enough data for cost model development.

**4.7.1.3** The third strategy scored by the respondents was the introduction of cost modelling development in higher institutions. Respondents according to them believe that the sophisticated and complicated nature of developing cost models requires the introduction of its studies at the academic level. This means, institutions that offer estimating oriented course should inculcate the study of cost model development in their curricular to aid in the introduction at the classroom level. This will equip the individual professional before he or she comes out to practice, by so doing will help eliminate the non-familiarisation and the lack of understanding to these techniques. A research by parametric cost estimation association (Department of defence ,1999) opined that model development introduced in the curricular of higher institutions has helped in its development. Un-doubtfully, this is revealed in the utilization of cost modells in developed countries.

**4.7.1.4** Enhancement of publicity about cost estimation modelling benefits ranked fourth by the respondents. Publicity for any product in effect achieves success for the product. Publicity in this content is making known the techniques and its utilization to the industry professionals. The publicity in effect will detail all aspect of the various techniques to aid in understanding of the techniques. Earlier discussion in this study however revealed the benefits of utilizing cost modelling techniques and hence the need to give it the necessary recognition. Bledsoe (1992) opined that the positive benefits of cost modelling demands a better way to communicate to the industry. The study by Smith et al., (2000) suggested that, if the positive effect is sold to the industry, it will maximize its appraisal. It is therefore believed that strong publicity in the Ghanaian industry will facilitate the utilization of cost models among cost estimation professionals.

**4.7.1.5** The fifth strategy ranked by the respondents was Organization of cost modelling workshops for QS professionals. Apparently cost model development tends to be distinct in terms of the techniques adopted and the type of project (Oberlender, 2001) hence the need for workshops to train professionals on the use and development of these techniques. Consequently, the best avenue for evaluating needs and the challenges of critical tools and techniques before they are used can be embraced by the creation of workshops for professionals (Wang and Wang 1996). Organizing workshops according to the respondents would provide better understanding to the various available techniques for cost estimation model development.

### 4.8 CHAPTER SUMMARY

This chapter covered the analysis of data and discussions of its outcome. It showed the data analysis and the discussion results obtained from the field as well as the methods and the research survey. The results obtained from this study showed the consistency in results obtained by other, studies on strategies for the utilization of cost estimation models in the Ghanaian building industry.

## CHAPTER FIVE

## CONCLUSIONS AND RECOMMENDATIONS

### 5.1 INTRODUCTION

This chapter presents an overview of the findings regarding the achievement of the research aim and objectives. Subsequently it filed proposed strategies to be adopted by the industry professionals for the utilization of cost estimation models based on the main findings. Additionally, it outlines the contributions of the research to knowledge and industry and finally limitations of the research as well as the recommendations made for both industry and future studies.

### 5.2 RESEARCH QUESTIONS

Four main research questions were proposed;

1. What is the level of awareness of cost estimation models with regards to construction industry in Ghana?

2. What are the barriers of cost estimation model utilization with regards to construction industry in Ghana?

3. What are the drivers of cost estimation model utilization in the construction industry?

4. What are the strategies that will help utilize cost estimation models in construction industry in Ghana?

### 5.3 ACHIEVEMENT OF RESEARCH OBJECTIVES

The study aimed at proposing strategies for the utilisation of cost estimation models in the Ghanaian construction industry. In achieving the stated aim, four objectives were established for the research study. Explained in the following sub-sections were the ways in which the objectives helped in attaining the stated aim.

#### 5.3.1 The first objective; To determine the level on awareness on types of cost estimation models in the construction industry;

To attain this objective, an extensive literature was reviewed based on the types of cost estimation models in the construction industry. In achieving this objective, quantitative analysis was used to examine the understanding of professionals. The researcher employed questionnaire where respondents were required to rank based on their understanding. Results were analysed using mean score and one-sample t-test. The Cronbach’s coefficient alpha was used to test consistency of the results which suggested that the 13 items have relatively high internal consistency having alpha coefficient of 0.847. From the responses gathered from Quantity surveyors within the construction industry, it was reviewed that the first ranked traditional cost estimation model used was the bill of quantities with a mean value of 4.83. This was attained because most quantity surveyors claimed they have more experience in preparing Bill of quantities and most clients require them. The second traditional cost estimation model used was the superficial floor area method with a mean value of 4.38. Superficial floor area method is measuring the aggregate floor region of all stories between outside dividers without conclusions for other items. This results also discovered that the least practiced cost estimation model is the Story Enclosure method. The Story Enclosure method has had little application in industry in view of the volume of work included and the deficiency of distributed expense information for its application.

Considering the non-traditional cost estimation models in use for cost prediction in the quantity surveying field, the most practiced cost estimation method was the Regression models which obtained a mean score of 3.13. Results agreed with Ashworth (1994) which stated that the most popular, useful and applicable technique, however is the multiple regression analysis. The second ranked traditional estimation method used is the Element based floor area model with a mean score of 2.88. Further the least ranked non-traditional cost estimation model was the Fuzzy logic obtaining means score value of 1.45.

#### 5.3.2 The second objective; To identify barriers of cost model utilisation in the Ghanaian construction industry;

The second objective focused on identifying the barriers to cost model utilization in the Ghanaian construction industry. In accomplishing this objective, barriers of cost model utilisation were retrieved from literature as well as data from respondents to ascertain what they admit as barriers to cost model utilization to the industry. However, due to the large number of reliant variables involved in the study, there was a higher likelihood that some variables have their end results in the same underlying effects which are high, in that case factor analysis was adopted to analyse the data retrieved. The results indicated that all the variables representing the barriers of cost model utilisation in the Ghanaian building industry were significant because they had an eigenvalue of more than 0.5 except ‘The degree of sophistication is seen as too needless for an average project’, ‘Lack of comfort in using model based estimating tools’ and ‘Lack of industrial patronage’ with the least extraction factor of 0.266, 0.374 and 390 respectively.

The twenty-three variables were reduced to four components. The components generated includes; Perceptions on cost model techniques, Inefficient Techniques, Unavailability of cost data and Lack of Understanding.

#### 5.3.3 The third objective; To identify the drivers of cost model utilisation for cost estimation with regards to the construction industry in Ghana;

In other to achieve this study objective, there was an in-depth review of literature based on the drivers of cost model utilization for cost estimation. On the drivers of cost model utilization for cost estimation, seventeen factors were identified with seven from literature. This was buttressed by the information gathered from the primary data where the respondents agreed to some of them. Respondents ranked all the drivers of cost model utilisation for cost estimation according to their importance to them. Results were analysed using factor analysis. Factors were grouped into five components and were discussed extensively identifying the main drivers of cost model utilization for cost estimation. Results indicated that all the variables representing the drivers of cost model utilisation in the Ghanaian building industry were significant because they had an eigenvalue of more than 0.5. Also, the alpha coefficient for the 17 items was 0.893, suggesting that the items had relatively high internal consistency. The seventeen variables were reduced to four components. The components generated are as follows; efficient cost estimation, cost advice, risk management and improvement in estimation Process.

#### 5.3.4 The fourth objective; To propose strategies for the utilisation of cost estimation models for cost estimation in the Ghanaian construction industry;

The fourth objective centred on the proposal of strategies for the utilisation of non-traditional cost estimation models for cost estimation in the Ghanaian construction industry. The secondary data showed that there were proposed strategies by previous studies. From the various strategies for the utilization of cost estimation models, eight strategies were identified and analysed using Relative Importance Index. The primary data revealed five main strategies for the utilisation of cost estimation models for cost estimation in the Ghanaian construction industry which included; improvement of data management in firms, involvement of key stakeholders in cost estimation modelling study, enhancement of publicity on cost estimation modelling techniques, organizing workshops and introduction of cost estimation modelling development in higher institutions.

### 5.4 CONTRIBUTION TO KNOWLEDGE AND INDUSTRY

The study has added to the pool of information both to knowledge and industry in different ways. Below are the ways in which academia and practice have gained from the research.

#### 5.4.1 Contribution to knowledge

Outlined below are the contributions made to literature by this research.

- The study has helped to reveal the awareness level of Ghanaian Quantity Surveying professionals on issues underlying non-traditional cost estimating models;

- It again unearthed the barriers and drivers to the utilization of non-traditional cost models in the Ghanaian construction industry;

- Proposed strategies when practiced could affect positively the utilization of non-traditional cost estimation models.

### 5.4.2. Contribution to industry

To the industry, below are the contributions from the research.

- Strategies proposed for the utilization of non-traditional cost estimation models could serve as a guide to be used by Ghana Institution of Surveyors and other stakeholders;

- Institutions and Stakeholders will be in the position to educate their members who are the major drive to the utilization of non-traditional cost estimation models.

### 5.5 RECOMMENDATIONS

The principal aim of the study was to propose strategies for the utilization of non-traditional cost estimation models in the Ghanaian construction industry as a result of this research; the following recommendations were made:

- Professional bodies are encouraged to consider the results revealed by this research to improve on their estimation by gaining understanding on non-traditional cost estimation models;

- Professionals are advised to obtain training on how to use the non-traditional cost estimation models in the system. It is believed that when they obtain mastery over the several cost estimation models, it will ease their work;

- Professional bodies must create the awareness of the concept to their members since that is an important point of call for its implementation. This could be done through the organization of seminars and workshops;and

- Sensitization of quantity surveying students on the need to improve cost estimation process through the adoption of new techniques. By so doing, the upcoming professionals would be up to date with the newer techniques.

### 5.6 LIMITATIONS

The study encountered some limitations in its reach and presentation which caused some precincts in its accomplishment. Outlined are the limitations of the researcher;

- The principal limitation encountered during the course of this research was the difficulty in reaching respondents to make available the needed information for the research. However constant visitation to their offices and telephone calls was helpful in the exercise;

- Another limitation was professionals finding difficulty in relating to the research topic since they barely practiced them, the researcher however took his time to explain the various techniques to them which in turn helped in the appraisal of the questionnaires;

- Non-traditional cost models involving the use of sophisticated computer applications were not part of this study. The study only looked at cost estimation model techniques that require basic computer software application;

- Considering the fact that cost modelling involves a lot of techniques, concentrating all these in a study meant trying to cover all the techniques on cost modelling which opens up for further additional techniques to be followed.

### 5.7 FUTURE RESEARCH

The research looked at the awareness level, barriers and drivers for utilizing non-traditional cost estimation models as well as proposed strategies for its utilization. It is revealed from the research that professionals are not knowledgeable on most of the newer techniques. For future research, the following outlined recommendations have been proposed.

- The benefits of utilizing non-traditional cost model techniques in cost estimation;

- Framework for developing cost estimation models.

### 5.8 CONCLUSION

Reference to the above summary concludes that there is poor knowledge base on the awareness of non-traditional cost estimation models among Ghanaian quantity surveying professionals. This is a challenge to the utilisation of non-traditional cost estimation models. The challenge, however with the utilization is due to professionals’ preference as they tend to perceive the newer techniques cannot address the issues of inaccuracies in cost estimation. Perhaps addressing the issues of the poor awareness level, lack of information and lack of education on cost estimation modelling in the industry would provide informed decisions on the modelling techniques to the quantity surveying professionals.

Further, it can be concluded that, the strategies if adopted could significantly influence cost estimation in the Ghanaian construction industry which in effect would eliminate estimation errors which sometimes leads project abonement.

### 5.9 CHAPTER SUMMARY

This chapter consisted of a summary of the findings, which formed part of contribution to the body of knowledge and fills the research and knowledge gap regarding the strategies for the utilization of cost estimation models. Within the Ghanaian construction industry, there exist poor appreciation on the utilisation of non-traditional cost estimation models and its importance within the industry. The conclusion to this research has been presented and the limitations recognized as well. Finally, recommendation for future research studies were also noted in the study.

## REFERENCES

Ahuja, H. N. And Campbell, W. J. (1987) Estimating: from Concept to Completion, Prentice-Hall Inc., Englewood Cliffs NJ.

Ali A S. (2009) Cost Decision Making in Building Maintenance Practice in Malaysia. Journal of Facilities Management 2009; 7(4): 298-306.

Al-Khalil, M. I. and Al-Ghafly, M. A. (1999) Causes of Delay of Construction Projects in Saudi Arabia. Construction Management and Economics, 17(5), pp. 647-55.

Al-Momani, A. H. (1996) Construction Cost Prediction for Public School Building in Jordan.

Alorgli Michael (2015) “cost estimation model for cost analysis of concrete works at the initial stages of construction projects in Ghana”. A thesis submitted to the department of building technology.

Aloysius Maalinyuur (2010) A study into pre-design construction project cost prediction in Ghana

Amoah-Mensah, K. (1995) Building Estimating Manual for West African Construction Practise, 2nd edition, Parcom Ghana.

Approach. International Journal of Quality in Maintenance 1997; 3(4): 273-80.

Arafa, M. &Alqedra, M., (2011): Early Stage Cost Estimation of Buildings Construction Projects Using Artificial Neural Networks. Journal of Artificial Intelligence, 4(1), pp. 63-75.

Ashworth, A. (1986) The Source Nature and Comparison of published Cost Information, building Technology and management annual report.

Ashworth, A. (1988) Cost Studies of Building. First edition, Longman Scientific and Technical.

Ashworth, A. (1994) Cost Studies of Building. Second edition, Longman Scientific and Technical.

Ashworth, A. and Skitmore, R. M. (1982) Accuracy in Estimating, Chattered Institute of Building annual report.

Ashworth, A., (2010): Cost studies of buildings. 5th ed. Great Britain: Blays Ltd, St Ives plc.

Assaf, S.A. and Al-Hejji, Sadiq. 2006. Causes of Delay in Large Construction Projects, International Journal of Project Management, 24: 349-357.

Babalola, O. and Adesanya, D.A. (2009). An evaluation of the level of accuracy of mechanical service cost estimates in Nigeria. Construction Research Journal, 2: 12-19.

Beeston, D., (1987) “A Future for Cost Modelling”, Building Cost Modelling and Computers. ed. Brandon, P.S., E& F.N. Spon, London, pp 17-24.

Berthouex, P. M. (1972) Evaluating Economy of Scale. Journal of the Water Pollution Control Federation, Vol. 44(11), pp. 2111-2118.

Boehm, B., Abts, C. &Chulani, S., (2000) Software development cost estimation approaches-A survey. Annals of Software Engineering, 10(1-4), pp. 177-205.

Bowen, P. (1982a) An Alternative Estimating Approach. Journal of Chattered Quantity Surveyor, February, pp. 191-194.

Bowen, P. (1982b) Problems in Econometric Cost Modelling. The Quantity Surveyor, May pp.83 84.

Bowen, P. (1984) Applied Econometric Cost Modelling, Third International Symposium on Building Economics, CIB Working Commission on w-55, (proceedings) Ottawa, vol. 3 pp. 144-157.

Brandon,P. S. (ed.) (1987) Building Cost Modelling and computers, E. & F. N. Spon, London.

Brandon, P. S. and Newton S. (1986) Improving the Forecast. Chartered Quantity Surveyors May, pp. 14-26.

Brandon, P.S., (1982) Building Cost Modelling: New Directions, E & F.N. Spon, London.

Briand, L. C., Emam, K. E. &Bomarius, F., 1998. A hybrid method of software cost estimation, benchmarking and risk assessment. Kyoto, Japan, s.n., pp. 19-25.

Brook, M. (1998) Estimating and Tendering for Construction Work. Second edition, Butterworth-Heinemann.

Buchanan, J. S. (1973) Cost Models for Estimating. Royal Institution of Chattered Surveyors annual conference report.

Chan, D. W. M. and Kumranwammy, M. M. (1999) An Evaluation of Construction Time Performance in the Building Industry. Building and environment 31(6) pp. 148-64.

Clark, F. D. and Lorenzoni A. B. (1978) Applied Cost Engineering, Marcel Derkker Inc., New York.

Cliffs, NJ. Oblender, G. D. &Trost, S. M., (2001). Predicting Accuracy of Early Cost Estimates Based on Estimate Quality. Journal of Construction Engineering and Management, 127(3), pp. 173-182.

Construction Management and Economics, vol. 14 (4), pp. 311-7.

Crosby, A. R. (1981) Forecasting Industry Resources. Institution of Chemical Engineers in London, November 4. Department of defence.

Ferry, D. and Brandon, P. (1980) Cost Planning of Buildings. Granada.

Gould, F.E. and Joyce, N.E. (2000). Construction Project Management. New Jersey: Prentice Hall.

Gould, P. R. (1970) The Development of a Cost Model for the Heating, Ventilating and Air-Conditioning Installation of a Building. MSc Project Report, Loughborough University.

Gunayin, M. H. &Dogan, Z. S., (2004) A neural network approach for early cost estimation of structural systems of buildings. International Journal of Project Management, 2 April, Volume 22, p. 595.

Harries, F. and McCaffer, R. (2001), “Modern Construction Management”, Blackwell Publishing, Oxford.

Harris, F. and McCaffer, R. (2013). Modern Construction Management (7th ed.). Oxford: Blackwell Publishing.

Hegazy T, Ayed A. Neural Network Model for Parametric Cost Estimation of Highway Projects. J ContrEngManage ;124 (3): 210 -25, 1998.

Heng Li (1996) “Neural Network models for intelligent support mark-up estimation”, Journal of Engineering, Construction and Architectural Management, Vol. 3, 1996, pp 69 – 81.

Heng Li, Shen, Q. P., Peter, E. D. Love (2005) Cost Modelling of office buildings in Hong Kong: an exploratory study [website: www.emeraldinsight.com (accessed May 2007)].

Horner R M, El-Haram M A, Munns A. Building Maintenance Strategy: A New Management.

Horner, R. &Zakieh, R., (1993) Beyond Bridget-An integrated model for estimating and controlling reinforced concrete bridge construction costs and duration. Transportation Research laboratory, pp. 4-11.

Idoro, G. I. (2009). Causes of time overrun in construction projects in Nigeria. Construction Research Journal, 2: 74-87.

Jaggar, D., Ross, A., Smith, J., and Love, P. (2002), Building Design Cost Management. Blackwell Science, London.

James, W. (1954) A New Approach to Single Price-Rate Approximate Estimating. Chartered Surveyor annual report.

Ji, S. H., Park, M. & Lee, H. S., (2010): Data Preprocessing-Based Parametric Cost Model for Building Projects: Case studies of Korea Construction Projects. Journal of Construction Engineering and Management, August, 136(8), pp. 844-853.

Ji, S. H., Park, M. & Lee, H. S., (2011). Cost Estimation Model for Building Project Using Case-Based Reasoning. Canadian Journal of Civil Engineering, Volume 38, pp. 570-581.

Khosrowshahi, F., & Kaka, A. P. (1996). Estimation of project total cost and duration for housing projects in the U.K. Building and Environment, 31(4), 375-383.

Kim, G., Seo, D. & Kang, K., (2005). Hybrid Models of Neural Networks and Genetic Algorithms for Predicting Preliminary Cost Estimates. Journal of Computing in Civil Engineering, April, 19(2), pp. 208-211.

Kim, G.-H., An, S.-H. & Kang, K.-I., (2004). Comparison of construction cost estimating models based on regression analysis, neural networks, and case-based reasoning. Journal of building and Environment, October, 39(10), pp. 1236-1241.

Kintoye, A. & Fitzgerald, E., (2000): A survey of current cost estimating practices in the UK. Construction Management and Economics, Volume 18, pp. 161-172.

Kish, L., (1957). Confidence Intervals for Clustered samples. American Sociological Association, April, 22(2), pp. 154-165.

Koo, C., Hong, T. & Hyun, C., (2011). The development of a construction cost prediction model with improved prediction capacity using the advanced CBR approach. Expert System with Applications, July, 38(7), pp. 8597-8606.

Kouskoulas, V. and Koehn, E. (1974) Predesign Cost Estimation function for building, ASCE Journal of the Construction Division, pp.589-604.

Kwakye, A. A. (1997) Construction Project Administration in Practice. Addison Wesley Longman.

Lamudi (2015): Factors affecting the prices of houses in Ghana. Accessed online at www.lamudi.com.gh. One 2nd November, 2015.

Laryea, S., (2010). Contractor project estimates vs. Consultant project estimates in Ghana. Paris, Royal Institution of chartered Surveyors, pp. 1-17.

Leader, A. L. & Prasad, J., (1995). Causes of Inaccurate Software Development Cost Estimation. Journal of Systems and Software, Volume 31, pp. 125-134.

Leung, M.Y., Zhang, H. and Skitmore, R.M. (2008). The effects of organizational supports in the estimation process on the stress of construction cost engineers, Journal of Construction Engineering and Management, 134(2): 84-93.

Lock, D. (2003). Project Management (8thed.). Hampshire: Gower Publishing Ltd.

Maddala, G. (1992) Introduction to Econometrics. Second Edition, Macmillan Publishing Company, New York. USA.

Marr, K. F. (1977) Standards for Construction Cost Estimates. Transactions of the American Association of Cost Engineers, (PaperB-6), pp. 77-80.

McCaffer, R. (1975) Some examples of the use of Regression Analysis as an Estimating Tool. The Quantity surveyor, December, pp. 81-86.

McCaffer, R., McCaffrey, M. J. & Thorpe, A., (1984). Predicting the Tender Price of Building during Early Design: Method and Validation. The Journal of the Operational Research Society, 35(5), pp. 415-424.

McNeil, S. (1981) Three Statistical Models of Road Management Based on Turnpike Data, M.S. Thesis, Carnegie-Mellon University, Pittsburgh, PA.

McNeil, S. and Hendrickson, C. (1982) A Statistical Model of Pavement Maintenance Expenditure. Transportation Research Record, Vol. 846, pp. 71-76.

Milton, J. A., Naumann, J. D. &Wetherbe, J. C., (1984). Empirical investigation of systems development practices and results. Information & Management, 2 April, 7(2), pp. 73-82.

Mislick, G. K. & Daniel, N. A., (2015). Cost Estimation Methods and Tools. 1st ed. Canada: John Wiley & Sons, Inc.

Norusis, M. J. (2004) SPSS12.0 Guide to Data Analysis. Prentice-Hall, Englewood.

Oxford Advance Learners Dictionary of Current English, (1998) Special Price Edition, Oxford University Press.

Perara, S. & Watson, L., (1998). Collaborative Case-based estimating and design. Advances in Engineering Software, 15 December, 29(10), pp. 801-808.

Quantity Surveyors (2017) Practicing Firms of Quantity Surveying Division. Journal of the Quantity Surveying Division of the Ghana Institution of Surveyors, issue 3 July-Sept. 2017 pp. 30-32.

R.I.B.A (1954) Royal Institute of British Architects, Rules for Cubing Building for Approximate Estimates, Ref. d/1156/54.

Raftery, J. (1984a) Models in Building Economics: A Conceptual Framework for the Assessment of Performance CIB Working Commission w55 Third Inter National Symposium on Building Economics. Ottawa vol. 3 pp. 103-111.

Raftery, J. (1987) The State of Cost/Price Modelling in the UK Construction industry: a Multicriteria Approach, Building Cost Modelling and Computers, E.& F. N. Spon, London.

Raftery, J. (1991) Models for Construction Costs and Price Forecasting, Surveyors Publications.

Samuel, Laryea (2010) Contractor projects vs consultant projects in Ghana: the construction, Building and Real Estates Research Conference of the Royal Institution of Chartered Surveyors, Dauphine University, Paris,2/3September 2010[http://reading.academia.edu/SamLaryea/Papers/313584/Contractor_project_estimates_vs._consultant_project_estimates_in_Ghana(accessed October, 2011)]

Shehab, E. M. & Abdalla, H. S., 2001. Manufacturing cost modelling for concurrent product development. Robotics and Computer integrated manufacturing, 17 August, pp. 134-353.

Singh, K., 2014. A Review on Software Cost Estimation Models and Techniques. IJETI International Journal of Engineering & Technology Innovations, September, 1(3), pp. 17-26.

Skitmore, R. M. and Ng, S. T. (2003) Forecast Models for Actual Construction Time and Cost, Building and Environment, vol. 38 (8), pp.107583. [http://www.ce. cmu.edu/pmbook/05_cost_Estimation.html (accessed May 2007)].

Skitmore, R.M., (1987). “The Influence of Professional Expertise in Construction Price Forecast”, Department of Civil Engineering, the University of Salford UK.

Southwell, J. (1971) Building Forecasting. Royal Institution of Chattered Surveyors Annual Conference Report.

Smith D, (2000) Construction Cost Estimation. Journal of Operational Research Society, June, 4(1), pp. 74-87.

Surveyors (1998). Guide to Elemental Cost Estimating & Analysis for building Works. The Association of South African Quantity Surveyors, pp. 1-41.

Surveyors (2013). Guide to Elemental Cost Estimating & Analysis for Building works. Africa Association of Quantity Surveyors, pp. 1-109.

The Institution of Surveyors, Malaysia (1998). “Manual for the preparation of Elemental Cost Analysis (ECA)”, Building Cost Information Centre, 2nd edition (with amendments).

Thomas, U. R., smith, G. R. And Cummings, D. J. (1995) Have I Reached Substantial Completion? Journal of Construction Engineering and Management, 121(1), 121-9.

Wiles, A. J. (1976) Cost Model for a Lift Installation. The Quantity Surveyors Annual Report. Wilson, A. J. (1978) Cost Modelling of Building Design. Chartered Surveyor, B and QS Quarterly.

Yaman, H. E., 2007. A building cost estimation model based on functional elements. Journal of Operational Research Society, June, 4(1), pp. 74-87.

**[...]**

- Quote paper
- Bismark Agyekum (Author), 2018, Exploring the Utilization of Non-traditional Cost Estimation Models in Ghana, Munich, GRIN Verlag, https://www.grin.com/document/500464

Comments