13/07/2008

A Note on Sociologists' Most Beloved Concept

Sociologists love to speak of "socioeconomic status". I have no problems whatsoever with Wikipedia's definition (except for the grammar):
Socioeconomic status (SES) is a combined measure of an individual's or family’s economic and social position relative to others, based on income, education, and occupation. When analyzing a family’s SES, the mother's and father’s education and occupation are examined, as well as combined income, versus with an individual, when their own attributes are assessed.

[...]

A fourth variable, wealth, may also be examined when determining socioeconomic status.

I've always had a funny feeling that this concept is so popular not because it makes an awful lot of sense theoretically, which it doesn't (for example, a simple hypothesis is that you're more likely to steal if you're poor, but your education should have nothing to do with that), but because the components are so highly correlated that they're hard to seperate from each other.

Hmmm...

4 comments:

pj said...

Is the definition not better that 'socioeconomic status' is a term applied to any number of measures of social and economic variables. There is no unitary 'socioeconomic status'.

LemmusLemmus said...

I completely agree with your last sentence; that was my point! (Maybe I didn't express myself properly.)

When I said that I had no problem with Wikipedia's definition, I meant that that is how the concept is used in sociology, in my experience.

In empirical research using questionnaires, at leat two of the variables are asked about and then the results of these two questions are combined using some mathematical formula. In many contexts I have no problem with that, I just don't like people pretending that SES is some sophisticated theoretical concept.

pj said...

Oh no, I realise the point you were trying to make, I just thought the wikipedia definition was a bit too essentialist, implying that there might be some 'real' SES that we are trying, imperfectly, to measure.

LemmusLemmus said...

Depends, I guess, on one's definition of "measure". If you take it literally, yeah, there'd have to be something to measure. More generously, a measure could just be some mathematical construct you put into a regression as an independent variable.