How to Lie with Divorce Statistics


There is an old saying often attributed to Mark Twain that goes… “there are three kinds of lies:  lies, damned lies and statistics.”

There are all types of misleading or false examples about divorce statistics by people who have a particular point of view about what meaning we should draw from rising or falling divorce rates.

Among the most troublesome examples is provided by this website which attempts to demonstrate that most of the worst problems in society (for example, murder, rape, armed robbery, etc.) are all the result of divorce.   Below is a sample of one of the tables at this website.  In this graph the author asserts that the divorce rate causes the murder rate to increase.  However, all this graph really shows is that there is a correlation between the murder rate and the divorce rate.  In statistics a fundamental idea is that the “correlation” between two numbers does not translate into a “causal” relationship.  There are at least three hypotheses that can explain the correlation between rates of murder and divorce:

  1. Murder causes divorce.
  2. Divorce causes murder.
  3. Some other factor (mental illness or spouse abuse) causes both divorce and murder.

The only way to figure out what is causing murder or divorce rates from rising is to test many different hypotheses and control for some of the possible other factors that may be contributing to changes in the rates.  It is important to test whether rates in one period of time predict future rates.

There are plenty of real consequences of divorce that should concern us without suggesting that all of society’s ills are the result of divorce.

Alert: false graph about divorce and murder rates