Statistics & Individual differences

Representative tests are essential in all empirical sciences. For instance, good visual acuity, measured with eye doctors’ Snellen E test, is assumed to predict good acuity when other visual acuity tests are used. A test is useless if performance in a test does not correlate with the performance of similar tests.

Surprisingly weak correlations are found in vision research in both young adults (Cappe et al., 2014), in healthy aging (Shaqiri et al., 2019; Garobbio et al., 2024) , and schizophrenia research (Gordillo et al., 2023). For example, we found that older adults performed worse than young adults in a battery of visual tests and illusions. However, within each population, performance in the various tests was largely uncorrelated, despite good test-retest reliability. In other words, poor performance in one test does not indicate poor performance in a similar test. In even other words: test validity is not good.

Therefore, tests are less representative than often expected, or the intra-individual variability is significantly higher than previously assumed. We propose to rethink the traditional approach of using a single test. Batteries of tests are needed.

Publications

Individual differences
Individual differences in aging populations
Individual differences in EEG