Openness to Experience and Overexcitabilities, a Jangle Fallacy With Ethical Implications: A Response to Barry Grant

Abstract

A jangle fallacy is the assumption that measures are different by having different names, although they assess the same underlying construct (Gonzalez et al., in press). This is the case of openness to experience and overexcitabilities, which we affirmed before (Vuyk et al., 2016a, 2016b) and Grant (2021) critiqued. We respond to his claims, insisting that gifted education ought to embrace openness to experience and the Five-Factor Model (FFM) when referring to personality. Studying construct overlap is less common than developing new measures, yet they should be equally common; assessments of potential jangle fallacies include a close study of measure content, factor analysis, multitraitmultimethod matrices, incremental prediction on chosen outcomes, and replications (Gonzalez et al., in press). A multitrait-multimethod matrix was impossible as the Overexcitability Questionnaire-II was the only available instrument. We examined the content of measures, recommended future replications, and focused on factor analysis to determine dimensionality. Grant (2021) incorrectly insisted that in our study, only the model with separate openness facets and OEs could support our hypothesis of openness and OEs representing the same latent constructs. However, the model where each openness and OE pairs loaded as one dimension also represents the equivalence of constructs in factor analysis. We had to operationalize openness and OEs using available instruments, even with the OEQ-II flaws. Grant (2021) regards their common variance of 58% to 75% as supposedly low and highlights the 3% of psychomotor OE after we had already mentioned the overlap was not as expected. A nonpeer-reviewed technical report on IQ tests explains that variances, which are group scores, mean that individual scores on one measure might differ from scores on the other measure. Grant cited this as evidence that “OEs and OtE facet pairs are different constructs sharing common variance” (p. 12; italics added) which does not logically follow. All we can conclude from McGraw’s statement is that measures, even measures of the same construct might differ among individuals even with high correlations; McGraw did not state this is evidence that constructs differ. Openness studies present many forms of organizing the construct; one is the NEO which aligns seamlessly with the five OEs. Other Openness models would have less facetlevel alignment; for example, Woo et al. (2014) with facets of intellectual efficiency, ingenuity, curiosity, esthetics, tolerance, and depth, which do not match directly with every OE yet have a parallel to some and to the general concept. As such, the argument that our study is without merit because two facets and one OE do not directly correspond is flawed; it still seems to be a jangle fallacy. Note a recent example in a meta-analytic correlation of .84 among the disputed construct of grit and the FFM factor of conscientiousness; even though measures did not “perfectly” overlap, they were deemed close enough to represent the same construct (Credé et al., 2017). Grant (2021) erroneously accused us of engaging in researcher degrees of freedom, a questionable reporting practice where researchers select to publish only the models with significant results and leave failed models unpublished (Simmons et al., 2011). In our paper, every decision regarding models is explicitly stated following the six requirements by Simmons et al. (2011) to avoid this practice: We finished data collection with a well-powered N before conducting analyses. We had many more than 20 observations per cell. We reported all variables assessed. We reported failed models. As we did eliminate variables, we reported results with and without said variables. As our models did not include covariates at any stage we had none to report. Simmons et al. do not propose blind rule-following; they question selective and misleading reporting, which their proposed reporting requirements should prevent. This then leads to our moral and ethical stance, which Grant thoroughly criticized, but misunderstood. As psychologists, we are committed to the Ethical Standards for Psychologists, which begin with a familiar principle: strive