Using the concepts of reliability and validity, critically examine how confident Human Resource professionals should be in using interviews as an effective selection technique.
Interviews are the most common selection technique for recruiting. Notwithstanding, this method has been frequently criticised. It is the task of this essay to evaluate how valid and reliable they are. Therefore, firstly validity and reliability have to be defined to form a basis to start with. Secondly, different types of interviews, beginning with the traditional type, will be presented and analysed. Thirdly, attempts to standardise and improve the interview in terms of reliability and validity will be critically examined. Finally a short outlook will be given, to show that not everything can be related to be reliable and valid. This text will analyse only selection interviews as the topic refers mainly to these. Furthermore it will only deal with personal interviews of employees, which excludes, for example, telephone screening. Every selection method has weaknesses and therefore it will become clear that interviews are not absolutely valid and reliable, but nevertheless to fill certain vacant positions they will be still irreplaceable.
Reliability and Validity are the two key characteristics that interviews have to have to be a suitable method for selection. They measure if the chosen methods provide consistent results and if they adequately measure the characteristic they are looking at.
“Reliability means that the selection methods, tests and ensuing results are consistent and do not vary with time, place or different subjects”. Or as Cowling puts it: “Reliability is a measure of the consistency with which a predictor continues to predict performance with the same degree of success”. That means that, for instance, two interviews at a different time and place, with different interviewers and questions but under otherwise same conditions and with the same applicants will bring the same result; namely the best candidate should still be the best and the interviewees who failed should still fail. It is also possible to maintain the conditions, the applicants and the structure but to change the other parameters of the assessment. By comparing the results, information about the reliability can be gained. However it is difficult to conduct these tests due to several constraints. It is considered to be nearly impossible to guarantee equal conditions for each sequence, as well as to provide sets of questions with different formulations but in a similar context. Furthermore the applicants have to be willing to take part a second time. These are only a few factors out of a bundle of problems which arise by testing reliability. Yet the measurement of validity is not easier.
The two concepts are connected: Reliability is a prerequisite for validity, this means that it is necessary, but not sufficient to ensure validity. This is easily understandable because: “if a test is so unreliable that it produces two different estimates of a person’s present behaviour, how can we believe that it gives a good estimate of future behaviour” ? “In the selection context, validity refers to the extent to which performance on the selection device/test is associated with performance on the job” or again Cowling: “Validity purports to measure how far a correct prediction of success in employment has been made.[...] Validation then consists of analysing the extent of the match between predicted performance and eventual performance”. Validity describes for instance the case when an employee is selected in an interview and eventually constitutes the best choice out of all the other applicants (who have taken part) for this special job. There are plenty of restraints concerning the verification of the validity (assuming reliability would be possible),however in this context it is suitable to list only two: one would have to ask the same candidates again and again and it is not sure if they are willing to participate. Consequently it is hard to examine if interviews are reliable and valid recruitment methods, although the two concepts set standards which are useful to build confidence in this selection method.
Interviews vary in many ways, nonetheless one can distinguish three main forms: The individual interview, where each participant has to compete with only one other speaker, or sequential interviews which take the form of a series of individual interviews. Besides those we can find panel interviews in which several persons conduct the interrogation (Cowling defines it as meaning that the applicant is faced with three or more interviewers). We can therefore imagine different situations with a different number and position of the interviewers: the departmental manager in small firms, a personnel officer with technical assistance, panels of senior executives sitting together, large committees in some of the public services, and variations and combinations of all these possibilities.
The traditional interview has questionable features in terms of validity and reliability. Normally it is not structured. The person who conducts it, asks his/her own questions which are different for every applicant. We will first examine the one-to-one situation, which has considerable weaknesses. There are reasons not to believe in the consistency of the outcome: It is carried out by a human being and as such there is always a certain amount of inconsistency in the judgement, behaviour and mood. Personal communication always bears its weaknesses and for that reason cannot be a perfect judgement method. The model of the communication process shows that there are natural obstacles which cannot be avoided, such as the perceptual filter between receiving, sending and decoding a message. Therefore two interviewers will never interpret and assess information in the same way, and even the same interrogator will reveal fluctuations in interpretations of data and assessment over a period of time. Hence it cannot be secured that an examiner rejects an aspirant only because he is tired or did not like to work on a warm summer day, but would recommend him on any other occasion. Other factors play an important role as well. Cowling, for example, cites a study where judgements were also affected by the sequence of interviews; if a manager interviewed three or four inferior candidates, the average one received a favourable rating. An interruption of any kind during the interview may be Disastrous for the consistency. The candidate becomes insecure which would possibly not have happened otherwise. He is furthermore worse off than another applicant without obstruction. Likewise one can discover deficiencies in validity. Especially in a judgement situation the halo and horns effects are to be considered because it is basically a meeting between strangers and every person is biased in a way or another. Some argue one might overcome this by long dialogues like one hour or even longer ones as we can see at Toyota where 75 minute talks (because they do not want the traditional 20-minute any more) are conducted. This might possibly cure the halo effect because the questioner has to overcome its first impressions and spend a considerable time with the other person and not only the first 5-10 (in which the halo effect manifestates). The horn effect however cannot be avoided. Many interviewers know the content of the application form and the curriculum vitae, and therefore tend to ask questions to confirm their often prejudiced first impressions. In this case it is not important how long the interview lasts. Additionally the particular interview situation can not be neglected. It is an “artificially distorted and entirely stressful situation”, a situation which hardly will appear in the job the questioned individual is applying for. Extreme opinions even suggest that “the only kind of validity that the interview can confidently be said to have, is to test whether people can cope with the special and unusual conditions of the interview”. That refers to the argument that important areas for a suitable performance in the following employment cannot be tested in a discussion. Furthermore, it is questionable if the demonstrated abilities (e.g. quick thinking and responding) can be repeated in very different circumstances at work, over a long period of time. Team work or hand-to-eye co-ordination for instance is not testable in a ‘confrontive’ situation. It is necessary to prove those traits with other methods. Nevertheless there have been attempts to improve the suitability of interviews. One concept refers to the role of the interviewer. Cuming mentions rules for interviewers to adopt during their time with the candidates. They should do a minimum of talking, ask open questions, remain impartial, be aware of their first impressions and the halo effect and take notes. Additionally the length should vary (adapted to each individual) and the questioning has to be exhaustive. These measures do not seem to be convincing with the respect of reliability and validity. They could be even counter-productive. Considering that, a different length of interrogation automatically leads to an unequal treatment of the candidates, since one obtains more time to present himself than another. This is then followed by distortions of results, validity and reliability. Furthermore, it is insecure if it is possible to override the halo or the horns effect. Notwithstanding, this concept contains positive ideas as well. The obligation to take notes might produce a more careful and comprehensible judgement. However because of all those reasons individual interviews in the traditional way are far from fulfilling the standards of a good selection process.
 Torrington, Derek, Hall, Laura & Taylor, Stephen, Human Resource Management (fifth edition), Pearson Education Limited, Harlow, 2002, p.196
 Storey, John & Sisson, Keith, Managing Human Resources and Industrial Relations, Open University Press, Philadelphia, 1995, p.122
 Tyson, Shaunt & York, Alfred, Human Resource Management (third edition), Made Simple Books, Oxford, 1996, p.94
 Cowling, A.G. and Mailer, C J B, Managing Human Resources, Edward Arnold Ltd., London, 1986, p.29
 There are various techniques to verify this: test-retest method, etc. cannot be listed here due to the limited extent of this essay. Not all of these tests are suitable for interviews.
 Berry, L.M., Psychology at Work, 2nd Edition McGraw-Hill, International: Psychology Series : Singapore, 1998, p.128
 Recruitment, http://wehner.tamu.edu/mgmt.www/W-Boswell/Overheads2.ppt
 Cowling, 1986., p.27
 there are diff. Kind of validity, but due to the extent of this essay it is not possible to explain Content validity, constant validity, construct validity and predictive validity
 Torrington, Derek, Hall, Laura & Taylor, Stephen, Human Resource Management (fifth edition), Pearson Education Limited, Harlow, 2002, p.246
 Cowling, 1986, p.44
 Cuming, Maurice W., The Theory and Practice of Personnel Management (sixth edition), Heinemann Professional Publishing, Oxford, 1989, p.97
 Recruitment, http://wehner.tamu.edu/mgmt.www/W-Boswell/Overheads2.ppt
 Buchanan, David & Huczynski, Andrej, Organizational Behaviour – an Introductory Text (third edition), Pearson Education Ltd, Harlow, 1997
 Tyson, 1996, p.98
 Cowling, 1986, p.43
 Torrington, 2002, p.103, p.242, p.386
 Tyson, 1996, p.98
 Storey, 1995, p.111
 Buchanan,1997, p.57
 Berry, 1998, p.107
 Tyson, 1996, p.98
 Cuming, 1989, p.102,103