The Huffington Post has an article about this study which interviews one of the study authors, Paul Irwing and also Janet Hyde who authored the earlier study that this one essentially attacks.
I am still trying to find the answer to the following question about the basic study: If another set of researchers took the same data and applied the same method, would they unavoidably get the same overall results?
So far I have not been able to find an answer to that question. It matters greatly, because of this:
In past studies on this topic, researchers would simply add up all the survey responses, according to Del Giudice. This led to imperfect results because of careless responses and misreadings. Through a sophisticated method called "structure equation modeling," the researchers claim they were able to remove this random error. When asked if he could translate this concept for a lay person, Irwing replied: "I teach courses on this and it takes me approximately 20 hours."Thus, the method which produces the new results (from old data, by the way) cannot be explained. Which makes it pretty nigh impossible to join in the conversation about what the new findings mean.
For example, think of possible interpretations for this:
"If you translate it into the simplest terms," said Irwing, "only 18 percent of men and women match in terms of personality profiles, and that's staggeringly different from the consensus view."What does this mean? That 18% of the respondents in the study scored on the same points on all the various scales? And if they did not, how different were their "personality profiles?" Completely different? A bit different on one or two dimensions?
It is this opaqueness that makes it so hard to talk about the findings.
The other bit in that interview I want to talk about is this:
Irwing thinks that some researchers in the past may have been biased in their methods, in order to reduce any gender difference. "It's for totally laudable reasons," he said. "People are very concerned, or were very concerned, that women didn't get equal opportunities, and that there was a lot of bias in selection processes."This sets Irwing and Del Giudice as the objective scientists and Hyde as the person who distorted her results because of political consequences. Hmm.
"People are afraid that studies like ours will turn the clock back," Irwing added.
Hyde is one of those people. "This huge difference is not only scientifically false," she said, "it has unfortunate consequences for places like the workplace and education and heterosexual romantic relationships."
But the authors stand by their results, and are currently drafting a lengthy response to Hyde's objections. "I think distorting science because of what you would like to believe, or because of what you think the political consequences are, is very dangerous," said Irwing.
The problem is that all researchers would like to believe something. It's not necessarily only feminists who have those beliefs.
For instance, if your study frame is evolutionary psychology, then you initially have accepted the idea that gender differences in personality are something that was caused by sexual selection a long time ago and is now retained in our stone age brains. So you like results that support that and dislike results that don't support that.
I'm not going to use a word as strong as "distorting" when I mention that studies can be written up in ways which support one view rather than the other. For instance, the Del Giudice-Irwing study uses US data which seems to over-sample both white Americans and educated Americans. The study write-up does not address the fact that the data is drawn that way. It simply applies the findings to all men and women, never mind the culture the self-reported results are drawn from.
Then there is this quote by Irwing, on women and men:
They're almost like "different species," Paul Irwing, one of the researchers, told The Huffington Post.This may be a subtle thing, but I don't think objective scientists describe their results that way. People who have an axe to grind might, perhaps.