The Woozle effect

One of my main beefs with feminists is the abuse of statistics (there is “lies, damned lies, and statistics”—and then there is feminism), often in combination with the principle that a lie repeated often enough is eventually taken to be the truth. (For instance, I have repeatedly written about the 77 cents on the dollar lie.) As I have also observed in the past, if their claims actually were true, anyone with a brain and a heart would be a feminist, myself included—but they simply are not true.

Recently, I became aware that there was a name for (a subset) of this type of abuse: The Woozle effect, where a piece of (real or merely claimed) information is repeated and repeated, with less and less constraint, until even an originally true claim is turned into an outrageous lie. (Cf. also the “Chinese whispers” game.)

I would strongly encourage the readers to read at least a part of the (lengthy) page behind the link, which includes not only a discussion of the phenomenon in general, but also discusses a number of common feminist “statistics”, including in the area of domestic violence and rape, as well as some uses not necessarily related to feminism. To quote one example:

• Gelles conducted a study using police domestic disturbance reports as the source. He explained this very specifically as a way to locate clear examples of domestic abuse. 20 Families with known histories were found. There were 20 Families referred by a private Social Service agency, making 40 in total. Then as a control group, neighbours of these families were recruited, making 80 families in total where half had a known history of Domestic Abuse. He was not looking for a national or global sample. Gelles says “Of the eighty families, 55 percent reported one instance of conjugal violence in the marriage. This was not unexpected, since half of the couples were selected because we thought they might be violent.” The evidence is for a small group, selected only due to police reports and known incidents. The 55% also referred to both men and women as victims.
• Straus writing the forward to the book “The violent home” used the 55% figure but without qualifying it.
• Langley & levy then cited Gelles & Straus claiming “Estimates that 50 percent of all American wives are battered women are not uncommon”. Gelles & Straus made no such claim or inference in their work.
• Langley & levy, journalist writing a book, then applied the Woozle to the general population arriving at the figure of 28 to 36 million American Wives being battered annually. “The twenty-six to thirty million are roughly half of all married women.”.
• The 28 Million figure, published in the book “Wife beating: the silent crisis”, then received extensive media coverage, including claims that at least 7 other studies showed the same 28 Million figure to be valid. In accounting for the lack of previous knowledge of what was called “A conspiracy of silence by men” the US Government, Congress, The American Bar Association, police and FBI, were all referred to as having “Culpable Ignorance”.

And please: The next time someone makes a statistical claim, please stop to consider whether it actually is true, given in the correct context, interpreted reasonably, and carries the appropriate qualifications about the circumstances. Sometimes a healthy skepticism is all that is warranted; sometimes the claim is simply absurd and can be ruled out with a little own thought, as with claims of absurd proportions of false accusations of rape to unreported rape cases or the existence of 14 million child-porn websites; sometimes the interpretation turns out be highly dubious, as with boys and girls doing housework ; sometimes the difference between correlation and causation is not even remotely understood; …