Hello,

the Torrance Tests of Creative Thinking have a long history and have proven their psychometric quality over the time. If you find significantly higher test scores in one of the subtests, it does not mean that this subtest is "better" than the other one. It rather means that cognitive performance in one facet (e.g., verbal) was significantly higher than in the other one (e.g., figural).

thanks Julia Maria Maier. But what if the observed difference happened say in two versions of a newly developed test?

The question is if the two tests are calibrated in a way that allows a comparison between the two measures. There must be a sort of linking or reasonable approach to norming that would allow such a comparison in a valid way.

So, the best answer to your question is: It depends.

When the new tests are just pilot versions, a difference between the tests for a person doesn't mean anything because it is unclear if metrics are comparable across tests. But if these two tests are well developed and have undergone massive empirical scrutiny that includes a compelling approach to norming or linking of the tests, you could make comparisons as suggested by Julia. For example, the Berlin Structure of Intelligence test for gifted children (Jäger et al., 2006) includes subscales for numerical, figural, and verbal divergent thinking. The test has been normed with a representative sample for Germany and, hence, allows for within-person assessment of specific strengths.

Jäger, A.O., Holling, H., Preckel, F., Schulze, R., Vock, M., Süß, H.M., & Beauducel, A. (2006).BIS-HB: Berliner Intelligenzstruktur-test für Jugendliche: Begabungs- und Hochbegabungsdiagnostik–Manual.Göttingen, Germany: Hogrefe.

thanks a lot, Boris Forthmann that was very helpful. One more question if you don't mind: for instance, what if you replace the pictures in TTCT -verbal with another pictures that you think are more (culturally) meaningful for a context, can that interfere with the validity of the test?

It could interfere, but if the choice you make is more meaningful for the context of your study, for example, it might be defensible. Even if you don't do proper pilot testing or other psychometric investigations in advance. Of course, then you might also risk that your study does not work. Hence, I recommend to make sure that validity of such a newly adapted version can at least be checked within the same study (if that is the purpose for you to adapt the test).

To answer this question shortly: yes, it can. If you exchange the pictures, you actually exchange the items. This means you would have to prove anew that psychometric properties are met and norms are proper etc. Formally, this can be interpretet as developing a parallel version of the original subtest.

https://ucr.sljol.info/articles/10.4038/ucr.v3i2.71?utm_source=TrendMD&utm_medium=cpc&utm_campaign=University_of_Colombo_Review_TrendMD_0

Article Development of self-reporting assessment tool to assess pers...