IQ scores on supervised tests.
hriqtests.com, March 2018
It has been a question for years, what the average IQ (according to supervised tests) of high range candidates is. The answer to this question would also be of utmost importance for high-range tests’ norming, at least.
After running a data-gathering marathon, I think I can present a rather indicative picture on this issue. A total of 92 scores on supervised, widely used, tests (WAIS, SBIS, RAPM, FRTs, CCFIT etc.) has been gathered (from candidates that have tried at least one of my tests). Quite unfortunately, however, many testees receive, especially through Mensa, a “greater than” score. In addition, many of these tests have a ceiling at 135 sd15 or so.
Two histograms follow. The first one includes all reported scores. In case of “greater than x” values, x is set as the intended value. The second one includes scores that were reported using an exact value, excluding the “greater/less than” values. So, let’s have a look at the charts.
1. Table N1
Calculated mean of the following scores is 138.848 (sd=15).
2. Table N2
Calculated mean of the following scores is 140.486 (sd=15).
3. Conclusion – comments.
It seems that the average performance of high-range candidates on supervised tests is around 140 sd15, as was well suspected all these years. I would, according to my personal opinion, suggest that it might be about 2 points higher, due to many near-ceiling or at-ceiling scores.
Knowing the average performance is a fine start; not enough, however. Well designed high range tests and supervised tests aim at measuring the same thing, using however a totally different method. Time factor is one crucial difference. Timed tests require quite a high level of focus in a short time period (minutes to hours), whereas high range tests probably require less focus, scattered in a much wider time period (hours to days or months, using usually breaks). In addition, high range tests require more complex ways of thinking than timed ones. This happens in order to balance the loss of time factor, but may include other consequences as well.
In addition, other factors tend to influence one’s result, either on timed or untimed tests. Fatigue, ADHD, autism spectrum disorders, motivation, personality traits, illness of any kind or any other personal/social issues at the time of testing may severely affect one’s performance and lead to significantly different results between two or more testing sessions.
Correlation with supervised tests stays between all this phenomena, struggling to prove high-range tests’ validity. As explained above, “greater than” scores and candidates’ specific traits (in general or at the time being tested), tend to distort given values and blur the picture. So, I think that any high range test that correlates more than 0.6-0.7* with supervised tests (after sufficient data are gathered, preliminary values can be much higher), it’s a test that can be trusted. (* Something that is usually missed by ones not much acquainted with statistics is that r (Pearson r) is a value between -1 and 1, not 0 and 1. -1 reveals a totally negative correlation (like two things are measuring the opposite thing, to make it simple), 0 reveals no correlation (like measuring something “neutrally” different), whereas positive values show that two things tend to measure the same thing – the higher the value, the higher the correlation.)
To sum up, I think that everyday life helps us conclude the following: Reaching Mensa or slightly higher level makes one intellectually capable of almost everything. However, personality traits, motivation, hard work, social life, personal life and luck (in the sense of circumstances) are only a few of other factors that can help one find the road to happiness and success.