22. april 2021

Manipulation of test results

For as long as psychometric test tools have been used for recruitment, one question has been central: Can test results be manipulated?

For as long as psychometric test tools have been used for recruitment, one question has been central: Can test results be manipulated?

Research has shown that the more the test person has at stake, the greater the likelihood of manipulation. In so-called “high-stake” situations – like recruitment, when the candidate wants the job – it is therefore essential to identify and address possible attempts to cheat or influence test results in a certain direction. The good news is that research has led to the development of a number of methods for detecting various types of test manipulation that are actually better than even trained professionals!

In test theory, the term response styles is used to refer to the response patterns that can influence test results in a particular direction. In this article, you can learn how to uncover the most common response styles in the personality and behavioural test, People Test Person (PTP).

 

Faking

One of the most common forms of test manipulation is when a person actively tries to influence the result. In cognitive testing, the term cheating is most often used, e.g., if the candidate tries to circumvent the rules by using aids or getting help from others. In personality tests, we talk about faking when candidates try to present themselves in a certain light, e.g., by covering up their weaknesses. If a person consciously exaggerates their own abilities, it is called Malingering or Impression management, whereas a more unconscious form of exaggeration is referred to as Self-deceptive enhancement.

It is difficult to prove that a person has deliberately faked a personality test, but a low Credibility score in PTP can be a good indication. In particular, a low Concordance score indicates that the person has been inconsistent in their responses, which may be a result of faking. This hypothesis can be further supported if the person has spent a lot of time responding, as that may indicate that the person has thought (too) much about their answers.

 

Social desirability

Social desirability is a response style in which the person provides the answers that are most socially acceptable. In the recruitment situation, this can be seen when the candidate adapts their answers according to what is expected in the job. It is challenging to identify, since most people consider what others want from them to some extent. If a person has provided a very socially desirable answer, it is therefore difficult to determine whether this is due to a desire to influence the test or evidence of a person who adapts their behaviour to what others think of them.

In particular, if Realistic self-evaluation under Credibility in PTP is low, it may be a sign that the candidate has tried to present themselves in a somewhat positive and socially desirable light.

 

Acquiescence & Opposition

Another response style is the tendency to uncritically answer “yes” to most (if not all) questions in a test, referred to as Acquiescence. Closely related to this response style is Opposition, which conversely refers to answering “no” to most questions.

In PTP, both response styles are controlled for using so-called “reversed keying”, where both high and low levels of a given attribute are asked about using reversed formulations. A person who answers “yes” to a question about details being very important, for example, but then also answers that details are unimportant, will have answered quite contradictory – and if this pattern is very distinct, it will result in a low Concordance score under Credibility in PTP.

 

Positional set

Positional set refers to responding very extremely by often choosing the most extreme response categories, but unlike Acquiescence and Opposition, responses may go in either direction. Conversely, some individuals are generally very neutral in their responses. This presents challenges because two people with the same level of a personality trait may score differently, depending on their preferred response style – the so-called “positional set”.

PTP has taken this into account by minimising the number of response categories so that there is less need to answer in degrees of agreement or disagreement. In addition, this response style is directly expressed in the attribute Decision tempo, where many sure answers (yes and no) will result in a high score, while more doubting answers (don’t know) will result in a low score. The latter also appears as ranges in the profile, highlighting the areas where the person has been most in doubt.

 

Random responding

Finally, a person may also choose to answer the test questions randomly (which on a cognitive test would be equivalent to guessing blindly). Random responding can be an expression of protest against the test situation or indicate that the person has had difficulty understanding the questions. The former will typically be paired with a relatively low amount of time spent (which may also indicate a superficial response), whereas a high amount of time spent may indicate a lack of understanding – especially if the person has completed the test in a non-native language.

Whatever the reason, random responding on PTP will typically result in a low Credibility score, in particular a low score on Concordance. In addition, random responding will reveal itself in a retest, as questions appear in a random, different order each time.

 

“Fake it till you make it”

Manipulation of test results is a contentious and controversial issue – and a delicate thing to accuse someone of. Therefore, you should make sure to use a test that can detect or account for different response styles and the most common types of test manipulation. Even the best methods are not without error, however, so if you suspect attempted manipulation, ask the person to complete the test again. Finally, you might also want to consider getting more references and comparing the description with your impression of the person – based on the personality profile, the feedback, and the interview.

More articles

A good job analysis, done quickly

A good job analysis is absolutely essential to finding the right candidate for the job and avoiding expensive bad hires. It can be difficult to

Untrustworthy test results?

In about 6% of all personality tests, the candidate’s responses are described as unreliable. When this happens, we generally recommend performing a retest. Learn more about how to assess the credibility of a test and how to correctly set up a retest.