Do religious people lie?

This blog has shown over and over again that religious people tend to be better-behaved than secular folks. Is this due to a tendency among church attenders to exaggerate, perhaps because they care more about what others think of them?

The Add Health Study asked participants if they never get sad. A fully honest people would not answer that they never get sad. I calculated the percentages of those who agree or strongly agree with the statement (sample size = 3,667, whites only):

Percent agreeing that they never get sad

Never attends 15.1
Less than once a month 12.0
Between once a month and once a month 12.7
Once a week or more 13.3

Those who never go to church have the highest percentage of people not telling the truth, but none of the differences is statistically significant. There is no evidence here that religious people are prone to give socially desirable answers.

Candor on surveys

ADD Health interviewers were asked the following about the study participant: "How candid was the respondent?" Answers ranged from "very candid" (1) to "not candid" (4).  Here are means by race/ethnicity:

Mean evasiveness score (N = 6,477)

Males
White 1.64
Black 1.89*
Mexican 1.82*
Puerto Rican 1.93*
Cuban 1.80 
Amerindian 1.65
Asian 1.66

Females
White 1.55
Black 1.82*
Mexican 1.65
Puerto Rican 1.89* 
Cuban 1.85
Amerindian 1.55
Asian 1.61

*significantly higher than white counterpart


For both sexes, blacks and Puerto Ricans were significantly more evasive than whites. Mexican American males were also less candid than their white counterparts. The largest gap is between whites and Puerto Ricans. It turns out to be one-third of a standard deviation (sd) for males and four-tenths of an sd for females--moderate differences. 

The numbers are consistent with another study I described which found that, compared to whites, blacks and Hispanics score higher on lie scales when participating in surveys. It might simply be due to mistrust of authorities. These results suggest that the true racial gaps in undesirable behaviors like crime might be larger than survey data indicate.     

Blacks score highest on "lie scale"

A reader raised the question about racial differences in honest self-reporting. I ran across this study of 980 children with a mean age of eight. It found that black children score the highest on the "lie scale" (with questions like, "I never get angry"--obviously dishonest if you answer yes), while Hispanic children scored the highest on concerns about living up to others' expectations (the "social concerns scale"). Whites scored lowest on the lie scale. The black-white gap is 37 percent of one standard deviation--a moderate difference.