I get asked what I think of psychometric test and here, I would like to share my thoughts on it. It seems like something popular and well-used in the corporate world, but like every tool invented in human history, it has limitations and disappointment visits anyone who puts undiluted faith in it. While there are many psychometric tests used throughout the world, the most well-known is Myers-Briggs. So, while my opinions below regards psychometric tests in general, Myers-Briggs will be the go-to model if more specific examples are required.
Limitations of psychometric tests
I will not go into the vast literature on academic and scientific criticisms of psychometric tests in general and Myers-Briggs in particular but will limit my observation to my own experience and my own deductions. Anyone who would like to understand the position of the academic and scientific establishments are encouraged to run their own research on the many academic and scientific papers and articles available.
Self-perception
One comment I read from respondents about psychometric test is that it is creepily accurate. Of course it would be! We need to be reminded that psychometric test are based on answers given by the respondent, and no other information. As such, they are perceptions of how people see themselves. So if we are assessing the accuracy of the results that are only a reflection of how we see ourselves, its ‘accuracy’ is only on how well it reflects what we think of ourselves, not necessarily who we really are. It is like dabbing red paint on a canvas and be surprise at getting a red painting as a result. The question is whether that is how everyone else see you, not how you see yourself.
Even if we were to take them as sincere answers, how many people do we know who has an inflated self-image. As a teacher in primary school once told me, we are the easiest person to fool. Therefore, psychometric tests are useful in what it tells us of how the respondent sees themselves, but the ability to tell us about the real person is often not all that reliable. Having the respondent themselves assessing the validity of the answers only opens out the scope for confirmation bias.
Snapshot in time
Psychometric test takes place at a certain point in time and reflects the situation at that time. Things may change if it is held at a different time in different conditions. Putting aside the more spurious but possible scenarios where are the respondent is answering a test after a particularly emotionally disturbing time, results may differ for instance, before a promotion and after the same person had a taste of management. The test providers answer to this is of course to conduct tests on a more regular basis, providing them with a convenient stream of steady income. Still there often isn’t an enough metadata attached to the test results describing the person’s conditions so that we can understand the test results within the context that it takes place.
Gaming the system
When you think about it, we use psychometric test for purposes such as assessing a person’s aptitude for joining an organisation or promotion within it. The person’s career could likely be dependent on the test. What makes us think that the person is not gaming the system? Yes, they are psychometric tests that include additional questions within a long questionnaire to assess consistency, and therefore the reliability of the answers. That is a start in that the test designer acknowledge that consistency could be questionable: but would any test designer be able to guarantee or even put a figure to the reliability of the conclusions? I highly doubt so and would view any such reliability indicators as, well, unreliable especially if the algorithm is not fully explained.
So, does this mean every set of answers that seems best fit for that position is either really that one in thousand best fit or merely reflecting social intelligence in being able to read another person’s intentions? Without casting any aspersions to anyone providing ‘best fit’ answers, I think anyone who uses the test conclusions without verify them shouldn’t be upset if the conclusions turn out to be unreliable. After all, humans have evolved to fit into the group on which we depend for survival and as such, we have evolved to please group members, and that probably include giving answers in psychometric test that we think would please the testers.
Complexity of human conditions
A psychometric test fits human personalities into pre-set boxes. A test like Myers Briggs applies four labels to four separate personality areas, giving rise to 16 possible permutations. In the first place, are there only 16 personality types in the human world? Surely there must be as many personality types as there are people and I for one, will in my stubbornness highlight in my actions the deviations from the predicted behaviour of the personality type enforced on me.
It is normal for us to categorise things into boxes so that we can analyse them: we generalise otherwise we will not be able to draw out common traits for specific observations: that’s how the human mind works. Sometimes, though we over-generalise in our need to categorise. We end up trying to fit complex conditions into simplified box labels and in the process, we lose the complexities of being a human being. Also, the categories we give tend to be discrete rather than continuous. Now, most of us inhabit a spot in the spectrum rather sits on an on-off position. For instance, most of us are at neither extremes of being an introvert or extrovert it, but is a mixture of both. We can even be fully extrovert in certain circumstances and fully introvert at other situations. The human condition is so much more complex than being an introvert or extrovert. If we were to label people as either extroverts or introverts, we are failing to recognise situations where they react differently from the labels. Or worse, we may them damage by forcing the real person into believing that they are one or the other, with all its consequent dysfunctional insecurities. Either way, we fail to recognise their humanity and accord them the respect due to them as a person.
How then to use psychometric tests
All decisions are made based on information and results of psychometric tests are just another source of information out of the many that a manager rely on to make a decision. The psychometric test is neither the decision-maker nor the tool to make the decision. If the decision is to assess what kind of person is sitting in front of me, the assessment can only be made by me, using a brain built out of millions of years of social evolution, informed by holistic data from all sources and based on my personal value system, expectations and very personal experiences: definitely not something psychometric test will be doing.
As with all sources of information, we need to understand how the information is derived to determine the context that it would be used and the level of reliability. Understanding this, I use psychometric test results only as an indication of what the respondent thinks of themselves at that particular point in time. I may try to understand the context of those results sometimes by asking further questions about the circumstances which give rise to certain interesting answers in the test – it is those circumstances I am interested in, not the answers themselves.
As with all sources of information, they need to be verified and corroborated. Psychometric test is only one source of uncorroborated sauce. So I use it only to develop lines of questioning to elicit further information. That is as far as I use psychometric test results. A company I was in tried to cut costs by only requiring successful candidates to conduct the tests, so I had no test results to help me prepare my interview. What a waste of money to run a psychometric test to be placed on file waiting to be used at a time when everyone has forgotten the context of those answers. Being a junior manager then, I just shrug my shoulders and carried on preparing my interviews using any other information available.
So use psychometric test results to help you elicit further information. Beyond that it has little practical help to categorize human personality types and I for one has no hesitation to step out of the assigned personality type labels and see people as they really are instead of the pre-set boxes such tests forces people to be.
Comments