Monday, October 11, 2010

Two Final Points About Bad Research On Bad Mothers

I sometimes despair of my ability to explain what is wrong with some crappy piece of research in general or with an Internet survey which you can take or not, as you please, in particular. But maybe the latter becomes clearer if we did a survey among just the readers of this blog on some topic about gender. It's pretty obvious that the results from that one would not reflect the general views of people out there (mostly because you are smarter, heh).

The same problem exists with all those Internet surveys. People who are willing to answer them may differ in some ways from people who are not interested in answering them. For instance, they may care more about the questions in that survey than the general population the survey tries to reach.

All this is preparation to ask you if the two posts below are of any use or not and if they are clear to you. That's the first point.

The second point has to do with that outrageous argument in both the BBC article and in the Fox And Friends discussion that 55% of mothers find it easier to bond with their sons.

That made me incredibly angry because the survey, flawed as it is, tells us nothing of the sort. What it tells us is that 15.26% of mothers with both sons and daughters in the survey bonded best with a child who is a son. The other children are not necessarily all daughters, by the way.

But if we wanted to give some sort of a total number, along the lines followed by both the BBC article and the Fox And Friends discussion, then that number would be around 7%, as only 46% of the mothers in the study had both sons and daughters. That number is naturally nonsensical, because parents without both sons and daughters cannot answer a question like that in any meaningful way. Still, we are given something more nonsensical by the BBC: False data applied to an inappropriate sample.

ETA: I found out that the sources of the errors are in the Netmums' own summary of the study. Examples:

A new survey by Netmums revealed that mums are much more critical of their daughters than their sons. Over 2500 mums took the survey and a staggering 88% admitted that they treated their sons and daughters differently despite thinking that this was 'wrong'.

The results showed that mums are twice as likely to criticise daughters. One in five admitted to letting their sons get away with more and more than half of all respondents (55%) even went as far as admitting they have stronger bonds with their sons.
Now, the net is full of bad surveys. But why some of them are picked for popularization without any scrutiny, well, that's worth thinking about.