Great Minds do not Think Alike: Philosophers’ Views Predicted by Reflection, Education, Personality, and Other Demographic Differences

Abstract

Prior research found correlations between reflection test performance and philosophical tendencies among laypeople. In two large studies (total N = 1299)—one pre-registered—many of these correlations were replicated in a sample that included both laypeople and philosophers. For example, reflection test performance predicted preferring atheism over theism and instrumental harm over harm avoidance on the trolley problem. However, most reflection-philosophy correlations were undetected when controlling for other factors such as numeracy, preferences for open-minded thinking, personality, philosophical training, age, and gender. Nonetheless, some correlations between reflection and philosophical views survived this multivariate analysis and were only partially confounded with either education or self-reported reasoning preferences. Unreflective thinking still predicted believing in God whereas reflective thinking still predicted believing that (a) proper names like ‘Santa’ do not necessarily refer to entities that actually exist and (b) science does reveal the fundamental nature of the world. So some robust relationships between reflection and philosophical tendencies were detected even among philosophers, and yet there was clearly more to the link between reflection and philosophy. To this end, demographic and metaphilosophical hypotheses are considered.

Appendices

APPENDIX: GREAT MINDS DO NOT THINK ALIKE

Prior Study

Participants. The prior study used LeiterReports.typepad.com to recruit 979 respondents starting in 2014. An a priori decision was made to exclude all incomplete surveys (n = 382) and patently insincere surveys (n = 3)—i.e., one participant reported that their ethnicity was “Handsome” and two participants reported that they were citizens of the countries “Texas” and “Narnia”. The final sample of 594 (106 identified as women, 483 as men, and 5 as other; 485 identified as White, 37 as Multiethnic, 23 as Asian, 22 as Hispanic or Latino, 4 as Black, 3 as Caribbean, 1 as American Indian or Native American, and 19 as other ethnicity).

Procedure and materials. Data for this prior study were collected in 2014 using Qualtrics in accordance with University of Colorado IRB protocol #13–0678 (Byrd, 2014)—see the Open Science link in the main text. Participants in this study completed measures of philosophical beliefs, reflection, class, and culture.

All items from the replication. Details about most of the measures in this study can be found in main text about the replication, unless otherwise stated in the main text or below.

Culture. All participants in the prior study were asked to report their country of citizenship. Table A1 shows how these responses were sorted as WEIRD or non-WEIRD countries based on earlier work (Klein et al. 2018; Yilmaz & Alper, 2019). This study barely recruited enough participants from non-WEIRD nations to detect smaller effects of country-based cultural differences. Without a reason to expect the replication to overcome this limitation, the citizenship question was not included in the replication.

Results. Many correlations from the prior study were smaller than anticipated (Figures A1, A2, A3, A4; Table A2). In fact, its correlations were as small as r = 0.08. Posthoc power analysis suggests that the prior had 62% power to detect such a small correlation (Faul et al. 2007). A sample size of 964 participants would have been required for 80% power to detect such small correlations (ibid.). This is one of the reasons that a larger replication was pre-registered and deployed.

Table A1 List of countries and their sample sizes from prior study

Fig. A1

Bivariate reflection-philosophy correlations from the prior study (3-item CRT in two left-most columns, N = 594) and the replication (3-item CRT in middle two columns and New CRT in right two columns, N = 705, right) with 95% confidence interval bands (grey). Red or blue lines indicates p-values less than 0.05 (prior study) or less than the Bonferroni corrected value of 0.001 (replication). All squared and cubic relations' 95% confidence intervals almost fully overlapped with these linear relationships

Fig. A2

Continued bivariate reflection-philosophy correlations from the prior study (3-item CRT in two left-most columns, N = 594) and the replication (3-item CRT in middle two columns and New CRT in right two columns, N = 705, right) with 95% confidence interval bands (grey). Red or blue lines indicates p-values less than 0.05 (prior study) or less than the Bonferroni corrected value of 0.001 (replication). All squared and cubic relations' 95% confidence intervals almost fully overlapped with these linear relationships

Fig. A3

Continued bivariate reflection-philosophy correlations from the prior study (3-item CRT in two left-most columns, N = 594) and the replication (3-item CRT in middle two columns and New CRT in right two columns, N = 705, right) with 95% confidence interval bands (grey). Red or blue lines indicates p-values less than 0.05 (prior study) or less than the Bonferroni corrected value of 0.001 (replication). All squared and cubic relations' 95% confidence intervals almost fully overlapped with these linear relationships

Fig. A4

Remaining bivariate reflection-philosophy correlations from the prior study (3-item CRT in two left-most columns, N = 594) and the replication (3-item CRT in middle two columns and New CRT in right two columns, N = 705, right) with 95% confidence interval bands (grey). Red or blue lines indicates p-values less than 0.05 (prior study) or less than the Bonferroni corrected value of 0.001 (replication). All squared and cubic relations' 95% confidence intervals almost fully overlapped with these linear relationships

Table A2 Standardized multiple regression coefficients predicting philosophical tendencies (-2 to 2 a la Bourget & Chalmers, 2014) from all measures in the prior study (N = 594). Each column is a separate multiple regression analysis