Search TESS:


Discomfort in Disagreement with God

Download data and study materials


Principal Investigator(s):

Benjamin A. Converse
University of Virgina
Home Page:

Nicholas Epley
University of Chicago
Home Page:

Sample size: 899
Field period: 4/8/2008 - 4/18/2008



In a series of correlation, experimental, and neuroimaging studies, including another TESS module (“With God on Our Side,” Converse & Epley), we have found that people may be even more egocentric reasoning about God’s beliefs than reasoning about other people’s beliefs. We propose, and test in this study, that one reason for the high correspondence between people’s own beliefs and their estimates of God’s beliefs is that people find it uncomfortable or threatening to believe that God disagrees with them. We therefore predicted that bolstering people against such a threat through self affirmation (namely, asking people to elaborate on why a particular value is meaningful to them; Steele, 1988) would significantly reduce the egocentric correlations between self and God relative to making people more vulnerable to the threat through a self challenge (namely, asking people to reflect on a time when they failed to live up to an important value). If this account were supported, it would also imply that the self-affirmation manipulation would not affect other targets similarly because it is not as important for people to be in step with the Average American as it is to be in step with an omnipotent, omniscient being.
Although this experiment replicated the basic finding of increased egocentrism when inferring God’s attitudes relative to inferring the Average American’s attitudes, the self-affirmation manipulation had no significant influence on participants’ responses.


1. Self-God correlations will be higher than Self-American correlations.
2. Self-God correlations will be lower following self affirmation than following self challenge.

Experimental Manipulations:

Each participant first selected from a list of personal values the value that was most important to him or her. In the self affirmation condition, participants wrote a few sentences to describe why this value was important to them and when in their life it had been important. In the self challenge condition, participants wrote a few sentences to describe a time when they had failed to live up to this value.

Key Dependent Variables:

(a) Own attitudes about same-sex marriage and abortion
(b) God’s attitudes about same-sex marriage and abortion
(c) Average American’s attitudes about same-sex marriage and abortion

Supplementary measures:
(d) Belief in God (yes / no)
(e) To what extent do you feel you have a personal relationship with God (in whatever way you understand God)?

Summary of Findings:

For each issue, the egocentric correlation was higher for God (rabortion = .53, rsame-sex marriage = .72) than for the Average American (rabortion = .44, rsame-sex marriage = .43), Steiger’s Zs > 3.0, ps < .01. We did not find any evidence consistent with the hypothesis that affirming the self would decrease egocentric correlations between self and God. Among believers only, where the motivational account would predict effects of affirmation to be the strongest, we did not find a significant difference in the Self-God correlations when indicating attitudes about abortion in the affirmation (r = .51) versus challenge (r = .53) conditions or when indicating attitudes about same-sex marriage in the affirmation (r = .71) versus challenge (r = .69) conditions. In exploratory follow-up analyses, we categorized responses based on the values that people chose to write about (e.g. social relations vs. all others). We did not find reliable effects of the affirmation manipulation within any of the subcategories of participants that we explored.


Epley, N., Converse, B.A., Delbosc, A., Monteleone, G., & Cacioppo, J. (2009). Believers’ estimates of God’s beliefs are more egocentric than estimates of other people’s beliefs. Proceedings of the National Academy of Sciences, 106, 21533-21538.

Copyright © 2014, TESS