1.1 Problem definition
1.2 Objective and construction of the paper
2 THE EXPERIMENT IN ORGANIZATIONAL RESEARCH
2.1 General facts and the term of the experiment
2.2 Types of experiments
3 LABORATORY EXPERIMENTS
3.1 General facts about laboratory experiments
3.2 Data collection and data processing
3.3 Data analysis and data interpretation
3.4 Possibilities and limitations of laboratory experiments
4 THE EXPERIMENT FROM LYNNE G. ZUCKER 1977: THE ROLE OF INSTITUTIONALIZATION IN CULTURAL PERSISTENCE
4.1 Classification of Lynne g. Zucker
4.2.1 The Ethnomethodological Approach
4.2.2 General facts of the experiment
4.2.3 The Transmission Experiment
4.2.4 The Maintenance Experiment
4.2.5 The Resistance to Change Experiment
5 CHALLENGES OF EXPERIMENTAL ORGANIZATIONAL RESEARCH
6 CONCLUSION AND EVALUATION
1.1 Problem definition
This seminar paper with the topic ‘The appropriateness of laboratory experiments for the organizational research’ was made in the frame of the seminar ‘Current Agendas in Institutional Theory’. There are various forms and methods in organizational research. In addition to the observation, interviews and surveys inter alia there are also experiments. Since Galilei (1564 - 1642) the experiment counts to the methodological component of the natural sciences (cf. Stein 1990:7) Nevertheless, in the social sciences the experimental procedure find much later recognition and input. For this prejudices expressed by John Stuart Mill (1846) among others against the use of experiments in the social sciences have helped. He had the view that the necessary control in the experiment is not sufficient to provide (cf. ibid.).
The analysis object of the institution-sociological research is the organization and their relationship to social and institutional environment (cf. Mayrhofer/Meyer/Titscher 2010:119). Comparatively, the experiment is in sociology, the least used method of research (cf. Atteslander 2010:177). Furthermore, laboratory experiments which serve to define and test business situations are the exception rather than the rule (cf. Stein1990:1). Why is that? In Mayrhofer/Meyer/Titscher (2010:121) will be noted that it is the complex issue apparently does not readily possible, that the relationships in such laboratory experiments to represent valid. Moreover Atteslander (2010:177) has the opinion that this has ethical, theoretical and practical research backgrounds. It must have high ethical barriers to be overcome if people manipulated in closed rooms or on success and/or failure of given and questions or tasks are to be measured (cf. Atteslander 2010:177). People are not laboratory mice (cf. ibid). For this reason: in this paper should address the question whether laboratory experiments are suitable for the organizational research?
The experimental approach provides a recommended method, viewed from research economic grounds (cf. Mayrhofer/Meyer/Titscher 2010:121). It is important to be aware that individual behavior is a part of the organization and therefore of the organizational analysis (cf. Mayrhofer/Meyer/Titscher 2010:123). In the following point the objectives and construction of the work are described.
1.2 Objective and construction of the paper
This paper deals with experiments and especially with laboratory experiments. One aim is to identify the advantages and disadvantages of laboratory experiments and what to consider in these experiments.
The main objective of the work is to find out whether laboratory experiments are suitable for the organizational research. These points are to be achieved by the following procedure. At the beginning of the work the author wants to describe some general facts about the experiment in the organizational research. Here are the different variables will be described: independent variable, dependent variable, confounding variable. This point should also be mentioned the various forms of experiments.
Because this work is focused on the laboratory experiments, the next section is about this type of experiment. At first some general facts about the laboratory experiment are enumerated. Furthermore, it should be received the data collection/data preparation (includes operationalization, experimental design, control of confounding variables) and data analysis/interpretation (includes statistical evaluation of the experiment, internal/external validity). Finally at this point the possibilities and limitations are presented for this experimental method.
An example of a laboratory experiment is the study of Lynne G. Zucker. This study is presented in this seminar paper. Zucker deals with the micro-level of institutions and can classify historically to the beginnings of neo-institutionalism. Therefore, at this point some facts should be presented to the neo-institutionalism. Zucker has the opinion that institutions are characterized by the feature durability. According to Zucker the resistance of certain practices is not linked to an internalization or sanction of extrinsic and intrinsic nature (cf. Ortmann/Sydow/Türk 1977:128). Institutionalization is a phenomenon of its own kind for Zucker. This may result in practices and/or elements of knowledge in a social situation and hence can be an element of objective reality (cf. ibid.). Zucker examined in her study the effect of different degrees of institutionalization on the durability in organizations (cf. Weik/Lang 2003:229).
After these points the challenges of an experimental organizational research will follow. Here will be discussed on things that should be observed in an experiment. In the last point in this seminar paper a conclusion and evaluation should be given. Here also the experiment of Zucker should be valued.
2 The experiment in organizational research
This section deals with general facts, the concept and the various forms of the experiment.
2.1 General facts and the term of the experiment
Francis Bacon (1561-1626) required the introduction of experiments in science (cf. Hader 2010:340). Bacon wanted to achieve, that the researchers work themselves empirically and that they draw scientific conclusions by they own observations (cf. ibid.). The experiment (as well as all other forms of social research) influenced the subject of investigation (cf. Atteslander 2010:177). Because most research studies carry experimental features, it is difficult to provide a clear definition for the term ‘experiment’ (cf. ibid.). According to Atteslander (2010:177) investigations should only be defined as an experiment, if a high level of control for a social situation exists. The experiment is intended to represent the strongest form of testing hypotheses. This applies for hypotheses as statements for explanatory and predictive nature (ibid.). Experiments are often understood as a specific form of the study design (cf. Kühl 2009:534). The experiment serves to validate earlier theoretically defined statements to specified conditions (cf. Atteslander 2010:178).
There are different variables in an experiment - independent variable, dependent variable and confounding variable (cf. Kühl 2009:534). Following is a brief explanation of what each variable is:
- Independent variable: This is planned by the experimental director and it is altered with intent. The aim is a reaction of the dependent variables (cf. ibid.).
- Dependent variable: The response of the dependent variable is observed on the proposed change of the independent variable. The effect is predicted in the hypothesis (cf. ibid.).
- Confounding variable: This is a variable that distorts the influence of the independent variable on the dependent. The experimental procedure aims to influence potential effects of confounding variables can be controlled or completely neutralized (cf. ibid).
In an experiment, individual condition factors (independent variables) changed by the researcher to find out what effects (dependent variables) thus arise (cf. Kühl 2009:534). It cannot be determined and taken into account all the possible confounding variables effect in the context (cf. ibid). The experiment with a randomization (random allocation of investigation units and the monitoring of the situation) to control or experimental groups is the only way to remove potential influences in a sample size on average (cf. Hader 2010:341; Kühl 2009:354). Therefore the experiment in the causal-scientific paradigm is the best way in the search for causality (cf. Kühl 2009:534).
To explain the concept of the experiment two definitions will follow. Here it should be noted in passing that there is a variety in the use of the term experiment (cf. Hader 2010:339). According to Greenwood the experimental method in sociology is explained as follows: “An experiment is the proof of a hypothesis which seeks to hock up two factors into a causal relationship through the study of contrasting situations which have been controlled on all factors except the one of interest, the latter being either the hypothetical cause of the hypothetical effect.“ (Grennwood 1945:28 in Stein 1990:11; Greenwood 1972:177 in Atteslander 2010:179)
The second definition, which should contribute to the understanding of the concept of experiment, is: "The experiment we understand as a repeatable observation under controlled conditions, taking one or more independent variables are manipulated so that a possibility for checking the assumptions made, ie the assertion of a causal connection, is given in different situations.”(Zimmermann 1972:37 in Atteslander 2010:179 f.)
2.2 Types of experiments
There are different types of experiments. In the following some of the different types of experiments are described briefly. It is the most speech of laboratory, field and quasiexperiments (cf. Hader 2010:341).
Laboratory experiments: These take place in controlled and thus artificial conditions (cf. Hader 2010:341). In laboratory experiments, it is possible to monitor almost all circumstances and to keep it stable (cf. Hader 2010:342). Thus, for example, faults can be avoided by a third party (cf. ibid.). In a laboratory experiment, a matter or transaction under simplified plan and clean conditions is examined (cf. Atteslander 2010:181). Other characteristics of laboratory experiments will follow in the next bullet.
Field experiments: In this experiment, the test subjects are in a natural environment (cf. Hader 2010:342; Kühl 2009:536; Zimmermann 1972: 194). This means that there is no extraction of the object to be examined occurs from its natural environment (cf. Atteslander 2010:181). Again, the conditions - independent variables - are manipulated by the researchers (cf. Kühl 2009:536). In field experiments, the requirement must be fulfilled that two different groups can be studied in their real environment (cf. Atteslander 2010:181). Thereby a group is exposed to the supposed causal factor (cf. ibid). The control of all conditions and the recordings of all the reactions of the subjects are difficult in field experiments (cf. Hader2010:342). Advantages of field experiments are: the highly realistic and a better basis for a generalization of the results found (cf. ibid).
Quasi-experiments: Here, the independent variable will be consciously manipulated by the experimental director (cf. Kühl 2009:535). The participating subjects cannot be assigned by random selection to the respective experimental groups (cf. ibid). This experiment changes the described experimental design. It can, for example, missing the randomization or the effect of an influence quantity is easily determined in retrospect (cf. Hader 2010:342). Quasi-experiments can take advantage of normal operations in reality (cf. ibid.).
In addition to these experiments, there are also other molds of experiments in social research. In the following other forms are listed and briefly described.
Simulation game: On first sight such experiments resemble the laboratory experiments (cf. Kühl 2009:535). In this case the independent variable is manipulated and the dependent variable is observed (cf. ibid). During the simulation game the subject is given a social situation but the actors' behavior is determined unspecified (cf. Atteslander 2010:183). It involves possible feasible simulation of social realities (cf. ibid). Therefore, the complexity of the simulated reality is mostly high influenced (cf. Kühl 2009:535). The simulation game is also less affected by the demands of standardization and measurability (cf. ibid).
Simulation: Here, a model of a particular system is used (cf. Atteslander 2010:182). Thereby exists an understanding of the key variables, the interdependencies and the behavior of the components (cf. ibid). The model starts while variables manipulated individually or in groups. There are the effects of this manipulation are reviewed to the entire model and then it can be reached conclusions on the reality (cf. ibid). Simulations are suitable for items which are excluded manipulation in reality (cf. Atteslander 2010:182f.). Simulation technology is closely related to the development of computers (cf. Atteslander 2010:183). Reasons for this are that there are no real actors required (cf. ibid.). Even the program language is more accurate than the natural language. Therefore, the simulation is mostly used in political sciences (cf. ibid.).
Crisis experiment: These are usually a variety of the field experiment (cf. Kühl 2009:536). The researcher creates a crisis situation for the subjects. Crisis experiments are primarily used in qualitative social research. The reason for this is that there are difficulties in the control of confounding variables (cf. ibid.).
Experiments with an ex-post-facto-design: These are studies where a social process has been completed (cf. Atteslander 2010:181). The development of this social process is to use a causal factor as adopted traced (cf. ibid).
- Arbeit zitieren
- Carolin Luckner (Autor), 2013, Modern Organization Theory. Current Agendas in Institutional Theory, München, GRIN Verlag, https://www.grin.com/document/283000