Human Resources Management
March 20, 2006
When your boss tells you the time is ripe to take over the world, the correct retort is, "Yes sir, I'll get right on to recruiting an army of followers." Identifying and attracting hopeful candidates is as easy as sticking up posters in the surrounding countryside, but prudence suggests narrowing the legion of candidates to a choice selection instead of just hiring every goon who shows up at the corporate headquarters. Because picking only the best to join you is a taxing effort, selection methods have been developed to ease this task. Understanding job selection methodology may mean the difference between a new world order and another humiliating defeat.
As always, the prerequisite to getting what you want is knowing what you want. With a good job specification you will know what kinds of knowledge, skills and abilities to screen for. Some selection methods will suit your needs better than others. Each method could be evaluated using five factors: reliability, validity, generalizability, utility, and legality. After exposition of these important factors we can take a look at a few of the most used selection methods and see how they fare under scrutiny.
Reliability. If a selection method ranks a candidate highly intelligent yet malevolent, and you test the person again just to make sure, you would expect the results to match. If the retest comes up showing a benevolent being of standard intelligence, the method may not be reliable.3 Unless a selection method gives consistent results with little error, it is not very useful.1 Some candidate qualities are harder to measure than others; physical strength is simple to assess, whereas emotional strength can be quite tricky.
Validity. Closely related to reliability, validity is the factor that tells whether a selection method's emphasis realistically corresponds to the actual job. Does it give grounds for evaluating all the important aspects you are looking for? Does it measure irrelevant aspects?1 Validity can be further broken into five kinds, such as criterion validity, where the selection test results are compared to successful existing employees' results or to the candidates own results later on. Another subtype is face validity, which just estimates how valid the method appears to be. For more detailed information on the subdivisions, please see the reference.4 Validity needs to be measured by experts to be credible, and this may be expensive. Unfortunately, the only way to really know you are selecting the right candidates is to make sure your methods are validated correctly and thoroughly.
Generalizability. The least understood of the factors and the hardest to pronounce, generalizability is the extent to which the results gained from a selection method can be extended to other purposes. If someone needs to be evaluated for a job now, and for another one later, time and money can be saved if some of the first selection results can be generalized for the second selection. Several studies are said to suggest that, given sufficient data to prove the connection, test results can remain valid at later dates, for different tasks, or for different requirements.5
Utility. This is simply a measure of relevance or usefulness. For example, while cognitive ability is usually a desirable quality, it has only a little importance for cannon fodder troops. Doing an expensive evaluation of mental proficiency would not be cost-efficient, since it is not one of the key qualities the job requires. If a selection method only improves your final cadre's makeup by a scant few percent, it has low utility and is not very useful.
Legality. Laws and rules abound governing how employee selection is allowed to be done. In certain jurisdictions regulations can be very interfering, and care must be taken to abide by them lest a flock of lawyers descend upon you. Particularly, the selection methods should be fair and biased toward minorities, you should get the candidate's informed consent for anything suspicious your selection methods will do, and avoid any probing that is very personal and not strictly related to the job. Of course, not probing deeply enough might also give trouble, if the candidate turns out to have illegal tendencies exploitable in the new position you have granted.6
Now we are equipped to examine a few of the more popular job selection methods: interviews, written application materials, references, testing, and work samples. More comprehensive lists and evaluation of methods can be found in the references.1, 2, 5, 7, 8 Among the more exotic methods mentioned there are polygraph tests, drug tests, handwriting analysis, and personality profiling.
Interviewing can be done in many ways and remains a very important selection method. This is curious, since it ranks quite badly in all five factors. Even though subjectiveness can be reduced by increasing the number of interviewers, results tend to lack reliability. Unless a very specific system of interviewing is used, proper validation is difficult to procure, and this would nullify the primary reason why interviews are used: flexibility and adaptation toward each candidate. While an interview can be a decisive, and therefore a very important, part of the selection process, it takes comparably more effort than some of the other methods - interviews take time, especially if candidates are numerous. Furthermore, unless the interviewers are very careful to document the process duly, slighted candidates may pursue legal action with cries of "discrimination!"
Applications and biographical data are an improvement over interviews in that they are faster to go through and should be quite reliable sources of information. By consciously looking for certain qualities among the resumes and cover letters the process can feel quite valid. The good cost-benefit ratio means this method has high utility. Decisions made based on this method, however, still need to produce due documentation to prove fair objectivity.
Asking for references is another often used method, although perhaps it should not be: how do you know the referrer is reliable? Sending useless employees to the competition with shining recommendations is deliciously evil and rarely punished. A few actual studies show that reliability and validity of references are "frequently questionable and inaccurate".8 Concerning legality, there is a chance of a referrer getting sued for libel for giving negative references, and on the other hand not asking for references during the selection may result in being sued for negligence.
Testing for desired qualities is a much preferred selection method. If the qualities sought are inherently quantifiable and the tests well-designed, the results will be reliable and valid. Some qualities are also easy to generalize on, such as types of intelligence. Considering how useful tests can be compared to the time and money needed, they have fairly good utility ratings. Of course legal trouble hounds this method as well, as some requirements may seem discriminatory against one or another gender, or possibly against a race or species. Proof at hand of the necessity of the tested qualities safeguards against litigation.
Another highly preferable selection method is evaluating samples of the candidate's work. While this discriminates against students who may not have worked yet, courts do not care much and so the legality rating of requesting work samples is very good. They also are a reliable measure of the candidate's skills, if the samples can be proven genuine. Criterion and content validity are particularly well-rated.1 The evaluation of the samples is a fairly simple task but yields much useful information, meaning high utility as well.
A strong combination of these selection methods, chosen to fit your job specification, stacks the odds highly in favor of a successful venture. Some selection aid packages are even available commercially, although the cheap ones evidently skimp on proper validation. With all these resources at hand, an optimal selection of lackeys can become a reality, bringing the corporate vision another step closer to becoming true.