Family First research 'high school level' statistics

June 5, 2016

Categories: Skepticism , Tags: Religion, Conservatism

Family first have confused correlation and causation (opens new window) in a recent report, and stated that unmarried couples are a major cause of child poverty.

Charles Waldergrave from The Family Centre said the report was a misuse of statistics:

"You can't just correlate things and then start talking about causality, you just can't do it that way. The fact that married people and people in de facto relationships earn different amounts of money doesn't make it causal in terms of child poverty."

There's a leap from showing that two statistics follow similar paths when drawn on a graph, to being able to say that one of the two is caused by the other. It may be because of a third factor that both measurements are tied to, it might be a more complex case where lots of variables come into play, or it might be just a coincidence.

There's a great website where you can see spurious correlations, where it's obvious that there's no causation from one to the other, such as:

  • Cheese consumption vs Death from being tangled in bedsheets
  • US oil imports from Norway vs driver deaths from railway collisions
  • The length of the winning word in a national spelling competition vs the number of people killed by venomous spiders

All of the above correlations are at least as close as the Family First correlation, but it would be a mistake to claim that one factor causes the other.

You can create your own correlations online. Just search for "spurious correlation" in google, click the first link, scroll to the bottom of the page and click "Discover a correlation (opens new window)".