Sunday, 13 April 2014

Religious do not lie for financial gain

Last year a study was published by Jason Childs titled "Personal characteristics and lying: An experimental investigation" in the journal Economic Letters (subscription required). This study was sensationalised in the press and on blogs by claiming that religious (those for whom religion was more important) people were more likely to lie for financial gain. Let me try set the record straight on this utter bollocks, after finding the time to actually read the paper.

Religious people had a very small tendency to lie for financial gain, but this very small tendency noted in the manuscript is not statistically significant (p-value 0.061). What is irritating about this is that in the conclusion the religious factor is given as much weight as the other factors which contribute to lying when the author says. (p-values in brackets added by me, lower value more significant)
"We also find that business majors (0.016), those to whom religion is more important (0.061), children of divorced parents (0.012), those who faced a low incentive (0.089), and those found the incentives salient (at all levels) (0.021) were more likely to send inaccurate messages."
I believe this above sentence is why some people automatically and erroneously jumped to the conclusion. Its not that the author is lying, its that the statistics are been misinterpreted.

If we want to continue taking the religious lying as fact, then there are other parts of the research that we should also consider fact as they are more statistically significant. Here is a list of these more important factors.
1) Women are more likely to lie for greater financial gain then men.
2) Children with divorced parents are more likely to be dishonest.
3) People who have studied a business major (i.e. economics) are more likely to be liars.

These are the finding from the study. The religious finding is a non-finding and is just overblown hype to sell. Apply the scepticism hat.