Reliability and Validity are two concepts that help to quantify the performance of any profiling tool.
Probably the most important question about about any personality test is: how well does it measure what it claims to measure? To be of any practical use, you need confidence that profiles accurately represent the actual personalities they describe.
There are two important statistical values that help to measure this level of effectiveness, known as reliability and validity. We'll take a moment here to see what each of these measures means in practice, before going on to see how they relate to Discus.
The first measure is a question of consistency: if we test the same person several times, we'd expect to get similar results each time. Reliability measures how much profile results change from one test to another (actually, we should mention that there are several different types of reliability, and the form we're discussing here is specifically test-retest reliability).
We'd certainly not expect results to vary wildly from profile to profile over time, so for an effective profile, we're looking for a high reliability score. Note, though, that people's attitudes and approach really do change over time, so we wouldn't expect to see a 100% reliability level from one profile to the next. (Indeed, understanding changes like this can be valuable in assessing a person's actual performance.)
A more difficult challenge is assessing the validity of a set of profile results: that is, do the factors in a profile measure the personality traits they describe. The usual approach in this context is to measure profiles against another well-established personality inventory. The preferred comparison in a case like this is Cattell's Sixteen Personality Factor questionnaire (or 16-PF for short).
Different profiling techniques measure different factors, so we wouldn't expect to find exact matches in the comparison. We can, however, look at correlations between factors in the two systems, to establish whether factors in one test demonstrate a consistent statistical relationship with those in the other.
You can find a full analysis of how Discus measures up in terms of validity and reliability in A Reliability and Validity Study on the Discus Personality Profiling System. This independently-produced paper is known informally as the Roodt Report after its author, psychologist Karin Roodt of Technikon Natal in South Africa. You can read the full contents of the paper at the link below:
Discus gives you online profiling, extensive reports and a host of free and unlimited services:
Reliability: A measure of how consistently a test measures a person's personality factors from one profile to the next.
Validity: A measure of how well a profiler reports on the specific factors it is designed to assess.
A Note on Scales: If you're unfamiliar with statistical reporting, it might be useful to mention that values are typically expressed on a scale from 0.0000 to 1.0000. If it's helpful, you can think about this scale as corresponding to the more familiar 0% to 100%.
About p-Values: The 'p-values' you'll find in the report describe the probability that a particular correlation occurred by chance, so lower values relate to more strongly established correlations. Conventionally a p-value of less than 0.0500 is seen as statistically significant.
Understand your work style, help build your CV, and much more with your own personal Discus report here.