While the phrase did not originate with him, Mark Twain once quipped, “There are three kinds of lies: lies, damned lies, and statistics.” Statistics have always been an interest of mine, and it was my offer of statistical support for Michael Spencer’s “Coming Evangelical Collapse” that resulted in me starting to write for Internet Monk.
Mark Twain wasn’t off by much in his quote. Statistics can be easily misinterpreted by the well meaning, or twisted to fit an agenda. Regardless of intent, much of what you will hear or read when it comes to statistics should be taken with a grain of salt. There is no better example of this than Church Divorce Statistics.
The most common mistake that I run into, is that people fail to realize that correlation does not imply causation. Or to put it more simply, just because two sets of information seem to move in tandem, does not mean that one caused the other. The best silly example of this, is the inverse correlation between the number of pirates and global warming. As the number of pirates has decreased over the centuries, the world temperature has gone up. Correlation? Yes. Causation? What do you think is more likely? That a decreasing number of pirates has caused global warming, or that an increase in world temperatures has led to a decrease in piracy? Of course we realize that both are ridiculous claims.
So here are two more claims of causation which have been made because of correlations. They too should also be taken with a large pinch of salt.
1. Living together before marriage leads to higher rates of divorce.
2. Increased church attendance results in lower divorce rates. (You may have recalled this one from last Saturday’s ramblings.)
It is a statistical fact that those who live together before marriage (co-habitation) have higher divorce rates than those who do not. I have heard this mentioned in several sermons: “You may have heard ‘Try before you buy’, but let me tell you that not only is living in sin wrong from a biblical perspective, it also increases your likelihood of divorce when you do get married.” I would always do a mental face-palm when I heard this, because this is a classic example of correlation does not imply causation. Is there something inherent in living together that causes higher divorce rates later, or is it, as I surmised, that people who have difficulty committing to each other before marriage have difficulty committing to each other after marriage?
Well we finally have an answer to this one. It turns out that neither the Pastors I was hearing, nor my assumptions were correct. It turns out that the age at which you move in together for the first time, whether before or after marriage, is the only significant predictor of divorce rates. Those who live together before marriage do so at a younger age than those who wait for marriage. The older you are when you start living together the less likely your marriage will end in divorce. Once you adjust for age, there is no significant statistical difference in divorce rates between those who live together before marriage, and those who wait until they are married before they start living together.
As for the relationship between Church attendance and divorce, here is what was written here last Saturday (and trust me, I am not trying to pick on Daniel here, it just happens to be the most convenient example):
How many times have we heard that, “Christians divorce at the same rate as non-Christians”? Not true at all, at least if you define “Christians” as those who actually go to church. One sociologist found that that 60 percent of those who claim Christianity but do not attend church have been divorced. Of those who attend church regularly, 38 percent have been divorced. Another prominent sociologist adds, ”‘active conservative Protestants’ who regularly attend church are 35 percent less likely to divorce compared to those who have no affiliation. Nominally attending conservative Protestants are 20 percent more likely to divorce, compared to secular Americans.” His other findings: Active Catholics are 31 percent less likely to divorce than secularists, while nominal Catholics are only five percent less likely. The biggest difference active faith makes (in this area) concerns Jews. Active Jews were 97 percent less likely to divorce than secular Americans, while nominal Jews were 53 percent more likely to do so. (Emphasis mine)
Did you note the causation conclusion that is being drawn? Those who attend church are “less likely” to divorce. Here is the problem with this statement. I would argue that these sociologists have the causation backwards. That is, it is much more likely that divorce causes a drop in attendance, rather than a drop in attendance causes divorce.
Think about it. You and your wife both attend church faithfully. You have your marriage difficulties and get divorced. The chance of both of you still attending the same church are pretty much nil. It is much more likely that one or both of you will stop attending church altogether. So, when they survey church attendance and those who are divorced, you no longer fit in the regular church attender group. I have seen this happen over and over again. (I should warn you though that citing examples, like I have just done, is another statistical faux pas. Just because I can name x number of couples who have stopped attending because of divorce, doesn’t make it statistically significant. The examples illustrate my point, they do not prove it.)
What does help prove my point is the Jewish statistics. “Active Jews were 97 percent less likely to divorce than secular Americans, while nominal Jews were 53 percent more likely to do so.”
If your worship options are limited because you live in a small community or belong to a smaller ethnic group, then divorce will have a much larger impact on attendance than if you have a church on every corner. In communities with a single synagogue, you have two choices, attend the same worship center as your former spouse or don’t attend at all. The result, active Jews who divorce become nominal Jews. So, active Jews will be shown to have a much lower divorce rate than nominal Jews, and the differences will be much more extreme that seen in the general population. This is borne out by the data.
However, for either the sociologists’ or my claims to be validated, we would need a longitudinal study that followed people over time. That way, we could see the actual results of attendance on divorce or divorce on attendance. Until then, when your Pastor quotes statistics from the pulpit, just remind him or her that there are “Lies, Damn Lies, and [Church Divorce] Statistics.”
Post Script: My undergrad degree was Economics, where one of my favorite text books was “How to Lie with Statistics.” An interesting read, and I see it used all the time!