Given a situation where science contradicts religion, who is more likely to fall into the trap of cognitive dissonance and say, “Damn the science! Damn the evidence! My religion is true!”? There are a lot of variables that determine how we respond to such contradictions, and I could build a model that captures many of them, but that’s not very fun, and to be honest—the question I ask is not scientifically interesting. That being stated…

The other day I received a huge (~750 mb) dataset from World Values Survey. It contains numerical data for questions asked to respondents all over the world about sociopolitical positions. There are over 1,000 questions, and I’ve spent a short while trying to get intimately familiar with them. Throughout my perusal I’ve come across some questions that pique my interest. I decided to build a logit model in Stata, which tests probabilities of the dependent variable.

I want to get some administrative issues out of the way first. This is not a scientific study. You, the reader, are my peer reviewer. I performed no post-regression analysis to determine if the model fits the data. And I’m deeply aware that because the data ranges from 1981 to ~2014, it likely suffers from some kind of serial correlation. Do not take this experiment as gospel.

The Model Variables

• DV: RelSci_Conflict: Whenever science and religion conflict, religion is always right
  • Binary variable
  • 0: Disagree, 1: Agree
• IV1: SexB4Marriage: Justifiable: Sex before marriage
  • 1 through 10
  • 1 = Never Justifiable, 10 = Always Justifiable
• IV2: Prostitute: Justifiable: prostitution
  • 1 through 10
  • 1 = Never Justifiable, 10 = Always Justifiable
• IV3: Beat_Wife:Justifiable: For a man to beat his wife
  • 1 through 10
  • 1 = Never Justifiable, 10 = Always Justifiable
• IV4: Homosexuality: Justifiable: homosexuality
  • 1 through 10
  • 1 = Never Justifiable, 10 = Always Justifiable

Results

Refer to the table below:

First, we can throw Prostitute out of the model. There’s no significant statistical relationship between justifying prostitution and rejecting science in favor of religion.

But this model does suggest that the more we justify sex before marriage, the less likely we are to reject science if it contradicts our religions. The same is true for justifying homosexuality. The more we justify homosexuality, the less likely we are to reject science. But those are obvious. Religion teaches us that sex before marriage and homosexuality are wrong. So it makes sense that the more we reject the teachings of religious texts, the less likely we are to support religion if evidence comes along to contradict it. So let’s look at that other variable.

The coefficient for beat_wife is striking (no pun intended)! The more someone justifies spousal abuse, the more probable it is they will agree that religious teachings must be believed despite contradictory scientific evidence. It’s not a small effect either. If I convert that into odds, it’s four to one that someone who justifies spousal abuse will reject science. To make better sense of this here’s a graph showing the relationship.

As you can see, justifying spousal abuse and rejecting science go hand-in-hand.

To be honest I thought this would not have had such an effect. I was surprised because views on domestic violence are largely secular and dependent on other social and legislative/judicial factors. But then I remembered: Religious books are rife with justifications for and duty towards domestic violence. Furthermore, women are generally subservient to men in many Jewish, Hindu, Christian, and Islamic sects. So on further contemplation, it’s no surprise that people who think beating their wives is okay are more likely to cling to their religions when science tells us their religions are wrong.

As I stated earlier this experiment is not scientific, meaning it probably suffers from a lot of problems associated with early models, which should give some religious folk a get-out-of-cognitive-dissonance-jail-free card after reading this. But do me a favor, if you believe both that religion always trumps science and that it’s okay to beat your wife, you need to make a serious change to your worldview.

I’ll let you slide for rejecting science, but domestic violence is unforgivable. Don’t expect the same level of patience from me for spousal abuse.