Jim Dougan wrote: > One of my favorites is the negative correlation between socioeconomic > status and diagnosis with mental illness, with the conclusion that being > poor causes mental illness. Or, better yet, that the correlation is > evidence of discrimination against the poor (i.e., it is easier to put them > in a mental institution than to deal with poverty). Way back when I was an > undergrad, I had a sociology professor actually make that claim. Even at > that tender age, I was able to put together various reasonable alternative, > such as: > > -- higher stress associated with poverty
I know this isn't what Jim meant, but it illustrates something I see regularly from students, and am struggling to understand. I might say in class "There is a negative correlation between socioeconomic status and diagnosis with mental illness. You might assume that means that being poor causes mental illness. Can you think of any alternative explanations?" and a student eagerly responds, "Maybe it's because of the high stress of being in poverty". Now, in my book, that's not an alternative explanation to the "being poor causes mental illness" cause-and-effect claim, but rather an explanation of how that very real cause-and-effect relationship works (assuming that the "high stress" is not accidental to "being in poverty", but is actually caused by the poverty, of course). An alternative would be something like "being mentally ill causes you to be poor" (the directionality thing, as Jim suggested) or the reporting bias he mentioned, or a true "third variable". I find that in general students find it easier to generate explanations of how a cause-and-effect claim works than to come up with real alternative explanations, particularly of the "third variable" type. For example, I ask students to imagine that they've found that a group of cat owners has higher blood pressure, on the average, than a group of dog owners, and ask for alternatives to the "Owning a cat raises your blood pressure and owning a dog lowers it" explanation. While some students talk about "third variable" things like differences in where the owners live (city/country), many of them come up with things like "Maybe it's because if you own a dog you have to walk it" or even "It could be because petting a loyal dog lowers your blood pressure"*. I guess there's a grey area there, and maybe these DO look to you like alternative explanations to the cause-and-effect claim. I think the acid test is this: under these "alternative" explanations, if we take away the original purported cause - the dog or cat - the effect disappears. Not so with real third variable explanations, like the classic "toasters and birth rate" one. I think that identifying "third variable" explanations is a relatively difficult thing, especially when you're dealing with negative correlations, for which the relationships between the third variable and the other two have to be in opposite directions from each other. Yet it's clearly a basic skill without which you cannot understand behavior. I wonder how much detail and how much practice instructors go into on this. Certainly the brief illustration with a couple of boxes and arrows in the book is FAR from sufficient. * This one is also an example of another general tendency I've noticed in research methods courses at all levels: when you use meaningful examples, the students want to talk about the content - about cats and dogs, in this case - and not about the methodological principles you're trying to illustrate. Paul Smith Alverno College Milwaukee
<<winmail.dat>>
--- You are currently subscribed to tips as: archive@jab.org To unsubscribe send a blank email to [EMAIL PROTECTED]