Wednesday, December 21, 2011

Warning: 70.4% of Statistics Are Bad For Your Health

Column Quick Summary:
  • No evidence supports Generation Y having a higher divorce rate than Generation X.
  • Three problems with a comparison: Generation Y is larger, divorce rate may be calculated in a way that misleads people, and the concern about the divorce rate might be to construct a false framework.

I'm only 8/27ths kidding.

After my exchange with Stephanie Coontz, she brought to my attention the post, Generation Y Dislikes Marriages, But Likes Divorce, where a columnist states the following:

Whether the description of this generation is warranted or not, statistics reveal that Gen Y'ers are more likely to have shorter marriages that [I'm assuming this word is supposed to be "than"] Gen X'ers and are more likely get divorced than the previous generation.
Here's the trouble: I haven't discovered these statistics (and Stephanie mentioned that she hadn't either). To be fair to the columnist, her credentials are listed in the article:
Silvana D. Raso, partner and head of the family law practice at Schepisi & McLaughlin, P.A. where she counsels clients in all phases of matrimonial law.
Her assertion might be made based on observation, statistics that legal departments have access to, or statistics that may be in databases that are not accessible for the public. However, I couldn't find anything that defends her assertion.

Do Echo Boomers have a higher divorce rate than Generation Xers? A few things to consider any time you see a divorce statistic concerning Generation Y:

  1. Generation Y is three times larger than Generation X. If the same percentage of Generation Y married as Generation X (unlikely to occur), there would be a probability of a higher divorce rate due to size.
  2. I intensely dislike the way that the divorce rate is calculated by most statistics. Calculating the number of divorces divided by the number of marriages creates numerous false implications. For instance, find this stat (good luck): what is the probability that a marriage will last a lifetime? That stat, by definition, would include only first marriages, thus the number of divorces in a year is unhelpful.
  3. Finally, I think people use divorce statistics often as a way of creating false frameworks (this would be the why are we looking at the divorce rate). For instance, educated people are less likely to divorce than uneducated. Then, some people (after reading that statistic) might get degrees with the hope that they'll be less likely to divorce. People who think this way fail to understand the most important aspect of studying human behavior: what you study, you change, meaning that if people alter their behavior, due to statistics, the statistics may change in the future.
For future studies, I'll find an appropriate disclosure. People should always treat studies - including my own - with a healthy degree of skepticism. Even large sample sizes can turn out to be wrong; and predictions of any kind - most of all my own - can easily turn out to be wrong in the long run (except Keynes' prediction: "In the long run, we're all dead").