The Dangers of Stereotyping (and the misuse of psychometric instruments)

– Les Jones

Having recently carried out testing using KAI, MBTI and some other inventories, I have again been concerned about the tendency of subjects to read far more into the results achieved than is intended by the originators of the tests.

Most of us using KAI try to correct the misapprehension of many test subjects that a high score is better than a low one and that Adaptors are not creative. We stress that KAI is a measure of style not level and that Adaptors are as essential to the success of an organisation as Innovators. Unfortunately the evidence from UK and US companies does not help, as it shows that senior managers are consistently returning mean KAI scores between 109 and 114. I believe that this is a commentary on the promotion policies of UK companies who tend to see the style characteristics of the Innovator as more appropriate for senior management positions. Companies seem to persist in this strategy despite the fact that many senior managers, who are themselves Innovators, are concerned that not only do their companies promote too many Innovators but that they also recruit too many. I have found it useful to explain to UK managers that a sample of Japanese senior managers attending the Executive Leadership programme at CCL registered a mean score of 95; it helps them to get their own scores in perspective.

These problems with KAI scores are trivial compared with the misconceptions subjects seem to have about MBTI. This arises primarily from the fact that the test producers insist on labelling subjects purely by their categories and ignore the actual scores involved. Hence, I found the ridiculous example of two students both having been labelled as “F”; one had a score of 27 on “F” and 0 on “T”, while the other had equal scores on both. The latter subject also objected to having to opt for the “F” category when she was so obviously borderline and saw herself as clearly having characteristics of both. This scoring problem is then compounded by subjects stating with absolute certainty that they are, for example, ENTPs and backing this up by saying that they exactly fit the profiles in the brochure. Even if each of the categories had an accuracy of 70%, which is extremely unlikely for the vast majority of people, the combined accuracy of all four categories is likely to be less than 25%.

Because of the dangers of stereotyping and the apparent willingness of the subjects to see in test results what they want to see, it is most important that psychometric tests are used very carefully. Not only should all users be trained, but the limitations of each test should also be carefully explained so that subjects have no reason to have any illusions, either negative or positive, about the true meaning of their results.

As test users, we have a duty to avoid using generalisations such as, “Innovators do this” or “That is typical NT behaviour”. Used properly, some psychometric tests can be extremely helpful to both individuals and organisations. Used carelessly, they may be useless and possibly harmful.

Les Jones, 1997