A statistical comparison of the OEJTS with other on-line tests
A comparative validity study of the Open Extended Jungian Type Scales was done using three other on-line tests that measured the same construct (Myers-Briggs Type Indicator alternatives). On the OEJTS website, at the end of the test, users were asked to volunteer to take a test on a remote website and report their results. Users were also asked to report their Myers-Briggs type if they knew it. From this data, the quality of the four different tests was assessed based on power to differentiate between self-reported types. The ability of the OEJTS to differentiate among individuals of known type was significantly higher than other tests. This is good evidence for the OEJTS being the most accurate on-line Myers-Briggs/Jungian type test.
3. Statistical type differentiation power
4. Inter-scale correlations
IntroductionThe Myers-Briggs types (also known as Jungian types) is one of the most popular personality systems. The Open Extended Jungian Type Scales is an open source measure of the types and was developed by selecting items that differentiated among individuals of known type. A good item should significant response differences between individuals on the dichotomy it is meant to measure and little or no response differences between individuals on the other dichotomies. Likewise for scale scores. For example, a good Introversion-Extroversion scale should show a large effect size between groups of self-identified introverts (Ixxx) and extroverts (Exxx) and little difference between groups of self-identified thinkers (xxTx) and feelers (xxFx), as so forth.
MethodsThe data was collected on the website of the OEJTS. After individuals completed the EOJTS they were asked if they knew their Myers-Briggs type already, and if so what it was. Then, they were asked if they would be willing to complete another optional survey. Those who did not were sent directly to their results. Those who did were instructed to open a test on a different website in a different tab, complete it, and then record their results in the survey. This was done in three separate phases, one for each remote test to be compared.
There are innumerable on-line Myers-Briggs tests, only three were chosen due to limits on subjects. The three were chosen on the basis of their judged popularity within Myers-Briggs enthusiast communities. The three chosen were the HumanMetrics Jung Typology Test, SimilarMinds Jung Personality Test and 16Personalities personality test.
With the data, for each scale, its ability to differentiate in among individuals of self-identified types will be calculated it terms of effect size for each dichotomy. Effect size reflect how different the scores of two groups are. Effect size will be calculated here as (group_mean_a - group_mean_b)/(0.5 * standard_deviation_group_a + 0.5 * standard_deviation_group_b).
Statistical type differentiation powerThe main measure of sale quality used here is effect size for scale scores between individuals of known types. The table below shows the effect size for score differences on the separate scales by dichotomy in individuals of self-identified type. Values greater than 2 are shaded orange, values greater than 1 are yellow and values greater than 0.5 are pink.
Inter-scale correlationsTo compare the content of the tests, the correlations between each scale in the three comparison samples were to computed.
OEJTS and SimilarMinds (n=996)
OEJTS and HumanMetrics (n=1076)
OEJTS and 16Personalities (n=449)
DiscussionBy the metric chose to be used here, the OEJTS is a clearly superior test. Its relevant scales have 25-40% more differentiating power than the next best scales on the IE, FT, and JP dichotomies, and has about equal differentiating power to its best competitors on the SN scale.
DataAll data used in this article can be downloaded here. Each file contains a codebook and dataset.