What To Know About Personality Tests: Validity & Reliability, Psychometric Properties & Why This Matters For Hiring

You may have run into personality tests because you had to take one as part of being hired or being promoted, or you may have enlisted the services of a consultant to assist with hiring new employees.  Why? The purpose of using personality tests is supposed to measure individual characteristics about people so that companies can have a purportedly objective way to measure those characteristics.  Is this person going to be reliable? Will this person get along with the rest of the team? What are the strengths of this person?  That's the hope and dream for using these instruments.  Instrument, by the way, is how people in psychology refer to these questionnaires. 

Some of the favorite personality tests used for hiring amongst businesses are the Myers Briggs Multiphasic Test (MBMT), the NEO (aka the Big 5), the MMPI, The DiSC, the Enneagram, amongst others.  If you're a business person trying to decide if you want to use one of these tests or if you're required to take one, there are some things you should know about these instruments.

I have a PhD in quantitative psychology with specialized training in measurement theory and psychometric psychology.  I also taught personality theory at the university level. 

The first two things you should know about in regards to personality tests are the concepts of reliability and validity. 

Reliability is that you get the same results every time you use an instrument such as a personality test.  So, the ruler will give you an inch or centimeter every time you measure linear magnitude.  It would be awkward if it didn't, if you were building a house, for example.  So, if you're using a psychological instrument, you should get the same results every time.  Well.  Sort of.  There are times when you'd expect the measurement to be different.  If you were measuring something where you'd expect growth, the measurement would differ each time you took the measurement.  There are ways to test for reliability under those circumstances.  If you expected two people to have the same growth, you might compare their scores and see if they were the same or different.  If different, then there might be something wrong with the instrument. That's one way to evaluate measurements and personality tests.  There are others.

The other concept is validity.  Are you measuring what you think you're measuring?  Is the ruler measuring linear magnitude or volume?  There are different ways to test and think about validity.  One is face validity.  It looks like the thing you're trying to measure.  You think you're measuring linear magnitude.  Another aspect of validity is construct validity, which is whether the thing you think you're measuring exists.   (Construct is a general term used by psychologists to denote constructed concepts, like love or sociability that can be measured through indirect means. We usually use behaviors associated with the construct.)  The example that comes to mind is not fit for polite company.  A wooden box was supposed to absorb and deliver healing energy.  That energy is not likely to exist, so measuring it would not have construct validity. Another type of validity that's interesting to know about is predictive validity.  If you're trying to measure someone's personality to see if they will get along with others, does that instrument predict how well someone will get along?  Under what conditions? What are the limitations to this?

When looking at personality measures, more is needed to know if the instrument is reliable.  You also have to know if the instrument has construct validity.  How well established is the construct validity?  What about predictive validity?  Will the instrument predict what it is intended to predict?  Is there a better one?

When psychological instruments like personality tests are being tested for reliability and validity, the results of those tests are part of what's called psychometric properties.  There will be information on the reliability and validity, as well as who the instrument was normed on.  What are the cut-off scores if the instrument is supposed to be predicting something?  Often, instruments are tested for what's called invariance, which is how well the instrument performs under different scenarios, such as different people in different languages, cultures, ages, education levels, etc.  There may be differences due to population differences, so the instrument may only be valid for certain people.

Why does all of this matter?  If you're a business owner, you don't want to waste your money on a personality test used for hiring that isn’t effective. .  If you're the one taking the test, you want to make sure that you are fully informed about what you're taking and know some questions you can ask if you're in a situation where you may be asked to take a test.  This is a matter of resource efficiency, ethics, and ensuring you're not putting yourself in an unfair situation. 

If you have yet to go to graduate school to learn all about psychometric testing and the statistical basis of reliability and validity, ensuring that you are either getting your money's worth or that the personality tests being performed on you are accurate may seem daunting.  The personality tests themselves have an air of objectivity and expertise that may seem iron-clad.  The fact is that there is no measure that can't be re-examined for either reliability or validity.  I belong to one of the divisions in the American Psychological Association that studies psychological tests.  Instruments that have been around for a long time are still being studied.  Does this instrument have construct validity on a population that hasn't been studied yet?  The answer may be no.  There are questions on the commonly used intelligence tests for children that will be answered incorrectly if the child grew up in a warm or tropical environment because the common knowledge that the question presupposes is geographically and culturally based.  In an email newsletter I get, there is study after study looking at invariance.  So no, no instrument can't be re-evaluated for reliability and validity.  Data on the demographics on which the instrument has already been tested should be included.  If you don't fit in that pool, it may not work for you the way it should and give inaccurate results.  If the person administering the test doesn't know that, that's a red flag on their expertise.

A quick way to know if a person is an expert is to look at their training.  Having a general understanding of psychology or related fields is just not enough.  Being able to administer a particular test is the bare minimum.  They should also know about the psychometric evaluation of instruments in general.  Did they take statistical classes in graduate school that looked at that?  Have they had subsequent training and experience that would teach them how tests are constructed?  Have they had mentorship or supervision in administering tests?  If they belong to the APA, they may belong to Division 12, the psychological assessment division.  They might belong to the Division for Consulting Psychologists (Division 13).  There are others. 

Personality testing can have a significant impact on people's lives especially when it comes to hiring.  It can mean the difference between getting a job or promotion or not.  Knowing what you're getting or taking is vital.  There are ethical issues involved.  Make sure you know before you buy.