The HEXACO Personality Space Before and After Re-Rotation to Approximate the Big Five Dimensions

Abstract

We re-oriented the HEXACO personality dimensions to approximate the Big Five, using two measures of the Big Five as targets in a derivation sample and then in cross-validation samples. The HEXACO approximations of Big Five Agreeableness represented blends of HEXACO Agreeableness, Emotionality, and Honesty-Humility. The HEXACO approximations of Big Five Neuroticism represented blends of Emotionality with low Agreeableness and low Extraversion. The residual sixth dimension, unrelated to the Big Five, contrasted Honesty-Humility with HEXACO Agreeableness. We then examined, in additional samples, some correlates of the original and re-rotated HEXACO dimensions. In the original HEXACO factor space, Honesty-Humility was the strongest correlate of unethical behaviors (selfishness and cheating), participant age, and “assumed similarity” to a friend or partner. Upon re-rotation of the HEXACO factors, associations involving these variables were divided between Big Five Agreeableness and the residual sixth dimension. Sex differences were mainly associated with Emotionality but after re-rotation of the HEXACO factors were divided between Big Five Agreeableness and Neuroticism. We discuss the relative merits of the original and Big Five-targeted HEXACO dimensions with reference to the practical utility of Big Five Agreeableness and Neuroticism and the simplicity and theoretical interpretability of the original HEXACO factors.

Author contributions

Disclosure statement

Ashton and Lee have received royalty payments for non-academic use of the HEXACO-PI-R.

Notes

1 Ethics approval for the study by Bucher and Samuel was obtained from the Purdue University Institutional Review Board (#1708019616); informed consent was obtained via a consent form completed by participants prior to completing the questionnaires. + Ethics approval for the study by Horwood and Anglim was obtained from the Faculty of Health, Deakin University Research Ethics (#HEAG-H 140 2017); participants were students who completed personality assessments as part of a course learning experience, but only those students who gave explicit consent for use of their data are included in the dataset.

2 McCrae and Costàs validimax rotation is derived from Schönemann’s (1966) orthogonal Procrustes rotation, in which one rotates an obtained orthogonal loading matrix toward a target orthogonal loading matrix. + In the validimax rotation, however, one rotates the matrix of correlations between a set of obtained orthogonal components and some external set of variables toward a target correlation matrix. + In this way, one can align the original components more closely with the external variables; the “validimax” name refers to the idea of maximizing validity by maximizing convergent and minimizing discriminant correlations. + The resulting transformation matrix can then be applied to the obtained orthogonal loading matrix to obtain variables` loadings on the validimax-rotated components. + The re-rotated components can also be examined in relation to other external variables.

3 The varimax-rotated HEXACO components were all similar in size, with sums of squared loadings ranging from 2.36 to 2.82. + After each of the re-rotations, four components had sizes within that range, but the Big Five Agreeableness component was considerably larger (3.83 for the IPIP-AB5C-SF target and 3.52 for the IPIP-NEO-120 target) and the residual sixth component was considerably smaller (1.47 for the IPIP-AB5C-SF target and 1.39 for the IPIP-NEO-120 target).

4 The study by Moshagen et al. was exempt from research ethics evaluation. Participants gave informed consent before taking part in the study, which involved no deception.

5 Ethics approval for the study by Ashton and Lee (2016) was obtained from the University of Calgary Research Ethics Board (#14-0792); informed consent was obtained via a consent form completed by participants prior to completing the questionnaires. + The consent form did not include any provision for public posting of the dataset but researchers can obtain it from Michael C. Ashton or Kibeom Lee.

6 Ethics approval for the study by Lee and Ashton (2018) was obtained from the University of Calgary Conjoint Faculties Research Ethics Board (#3235); informed consent was obtained via a consent form completed by participants prior to completing the questionnaires. + The consent form did not include any provision for public posting of the dataset but researchers can obtain it from Michael C. Ashton or Kibeom Lee.

7 McCrae et al. (2008, p. 1157), in discussing similarity and assumed similarity in spouses’ personalities as measured by the NEO-PI-R, asked “Why is the Compliance facet [of NEO-PI-R Agreeableness] so much less important for selecting a mate than the Straightforwardness and Modesty facets?” + We think that the difference can be understood in terms of HEXACO A and H: + the former facet is chiefly associated with A, whereas the latter facets are chiefly associated with H. + Another NEO-PI-R facet that is strongly related to (low) HEXACO A, Angry Hostility, showed results comparable to those obtained for the Compliance facet.

8 An anonymous reviewer inquired about the association of the residual sixth factor with broad personality meta-traits. + We expect that this dimension would be roughly uncorrelated with those meta-traits, because it roughly represents an opposite-signed blend of A and H, whereas previous research has found the A and H dimensions to have same-signed projections in the space of those meta-traits (e.g., Strus & Cieciuch, 2021).

Additional information

Notes on contributors

Michael C. Ashton

Ashton and Lee wrote the draft manuscript; Anglim, Bucher, Horwood, and Samuel gave input that was then incorporated into a revised manuscript. The latter authors also conceptualized and collected the data of the Study 1 samples re-analyzed here.

Kibeom Lee

Ashton and Lee wrote the draft manuscript; Anglim, Bucher, Horwood, and Samuel gave input that was then incorporated into a revised manuscript. The latter authors also conceptualized and collected the data of the Study 1 samples re-analyzed here.

Jeromy Anglim

Ashton and Lee wrote the draft manuscript; Anglim, Bucher, Horwood, and Samuel gave input that was then incorporated into a revised manuscript. The latter authors also conceptualized and collected the data of the Study 1 samples re-analyzed here.

Meredith A. Bucher

Ashton and Lee wrote the draft manuscript; Anglim, Bucher, Horwood, and Samuel gave input that was then incorporated into a revised manuscript. The latter authors also conceptualized and collected the data of the Study 1 samples re-analyzed here.

Sharon Horwood

Ashton and Lee wrote the draft manuscript; Anglim, Bucher, Horwood, and Samuel gave input that was then incorporated into a revised manuscript. The latter authors also conceptualized and collected the data of the Study 1 samples re-analyzed here.

Douglas B. Samuel

Ashton and Lee wrote the draft manuscript; Anglim, Bucher, Horwood, and Samuel gave input that was then incorporated into a revised manuscript. The latter authors also conceptualized and collected the data of the Study 1 samples re-analyzed here.