Revisiting the Relationship between Individual Differences in Analytic Thinking and Religious Belief: Evidence That Measurement Order Moderates Their Inverse Correlation

Abstract

Prior research has found that persons who favor more analytic modes of thought are less religious. We propose that individual differences in analytic thought are associated with reduced religious beliefs particularly when analytic thought is measured (hence, primed) first. The current study provides a direct replication of prior evidence that individual differences in analytic thinking are negatively related to religious beliefs when analytic thought is measured before religious beliefs. When religious belief is measured before analytic thinking, however, the negative relationship is reduced to non-significance, suggesting that the link between analytic thought and religious belief is more tenuous than previously reported. The current study suggests that whereas inducing analytic processing may reduce religious belief, more analytic thinkers are not necessarily less religious. The potential for measurement order to inflate the inverse correlation between analytic thinking and religious beliefs deserves additional consideration.

Atheists and Agnostics Are More Reflective than Religious Believers: Four Empirical Studies and a Meta-Analysis

“…Crucially, this association was evident despite the fact that religiosity and analytic thinking were measured in separate sessions. These results contradict Finley et al’s [ 24 ] hypothesis that the association between religiosity and analytic thinking requires participants to be put in an analytic mindset when reporting their religiosity. In each of these four studies, our religious belief scale was administered in a mass testing survey along with a variety of scales (presented in a random order for each participant; see S1 Text ).…”

“…As a procedure for identifying candidate studies, the primary author manually searched each article that cited any of the three original CRT/religiosity studies [ 4 , 12 , 13 ] via Google Scholar . Thirty-five studies were identified, including Finley et al’s [ 24 ] two conditions (which we treat as two studies for the purpose of this analysis), the four studies from the present paper, an unpublished study from a Master’s thesis [ 26 ], and 8 studies (across 2 articles, [ 28 , 29 ]) that were published after Finley et al Thirty-one of these studies found a statistically significant negative association between analytic thinking and religious belief, and four did not. However, two of these non-significant results ( r ’s = -.15 and -.23 from [ 22 ] and [ 29 ], respectively) were within the range of previous work but had relatively small samples.…”

“…However, two of these non-significant results ( r ’s = -.15 and -.23 from [ 22 ] and [ 29 ], respectively) were within the range of previous work but had relatively small samples. The only two strong exceptions are from Finley et al [ 24 ] and McCutcheon et al [ 36 ]. Arguably, three large-sample studies in which the correlation between CRT performance and religiosity was significant, but quite modest could also be considered partial exceptions [ 5 , 30 ].…”

“…These results were challenged in a recent paper by Finley, Tang, and Schmeichel [ 24 ]. Specifically, they hypothesized that the association between analytic thinking and religious belief depends on the order in which analytic thinking and religious belief measures are presented.…”

Anxiety‐induced miscalculations, more than differential inhibition of intuition, explain the gender gap in cognitive reflection

“…We bring evidence that it is important to distinguish between group differences in the CRT that reflect intuition inhibition, and can inform the dual process models, from the differences that do not tell much about group differences and hence are not informing any new developments of the models. Consider for example the finding that religious believers do not perform as well as nonbelievers on the CRT (Gervais & Norenzayan, 2012), (Finley, Tang, & Schmeichel, 2015), (Pennycook, Ross, Koehler, & Fugelsang, 2016). This result has important theoretical implications because it provides us with new insights about how religious believers process information, and challenges us to apply the dual-process model to the complex domain of religious cognition.…”