Social Science anthologies that use statistics?

May 7, 2014 6:04 AM

I'm looking for a social science type book, probably an edited volume, that tackles an issue using lots of data & statistical tests / arguments to debate an issue. I am less concerned with the issue being debated, and more concerned with learning 1) what constitutes proper manipulation of data 2) what tests are appropriate and when 3) what proper inference is. I'm not afraid of math.
posted by MisantropicPainforest to Education (12 answers total)

Anything you find will be an utter morass. There are parts of the the discipline of the social sciences that absolutely want to empirically justify themselves and gain respect by duplicating the methods of the physical sciences, employing quantitative research methods and randomized control trials. These efforts begin with the underlying assumption that all human experience is quantifiable, measurable, and very importantly -- predictable.

There are also parts of the discipline that feel no need to emulate the methods (or earn the respect of) the physical sciences, because the social sciences is often concerned with people, cultures, meaning-making, and these folks believe in the fundamental ability of human beings to be unpredictable and uncategorizable (or, more to the point, nothing important comes out of "objectively" "externally" categorizing people, that can't be better understood by respecting and listening to people's own understandings of their experiences). These are folks who resonate with the humanities, philosophy, critical theory, etc...

Given that -- the only book that comes readily to my mind, because I've been reading it, is "Social Class: How Does It Work?" by Lareau and Conley. The introduction lays out the challenges of deciding whether or not social class, as a concept, is robust, measurable, etc. But to give you some perspective, the American Psychological Association endorses, 3 different conceptualizations of social class that can be helpful to keep in mind when working with clients. Anyway -- here is a quote from the book:

"It is particularly difficult to discern whether 'class' is a viable concept that has empirical meaning across a variety of subfields."

And

"It [this book] seeks to determine whether there is evidence to support the viability of the concept of social class across a number of spheres..." And "...the volume seeks to assess the varying definitions of social class that are used in empirical work and to elucidate important challenges facing the researchers doing empirical work."


As a definable concept, for the social sciences, it can fall through your fingers. Is it just income? Education level + income? (Ask the baristas with PhDs if their education necessarily means they're upper class.) What about kids, who don't have incomes? What if one of their divorced parents is rich and the other is poor -- what social class is the kid in? Is it family wealth? If you are a trust fund anarchist kid picking food out of dumpsters to eat, but you know you'll inherit wealth at age 25, are you currently categorized as poor?

Every decision an academic makes about what defines social class is based on their own purposes for defining social class in the first -- and there we have a philosophical problem -- even the empiricism of the social sciences is reliant on a human being who lives in a particular cultural and historic context.

Does the inability of social scientists to agree on a single, measurable definition of social class mean that social class doesn't exist?

Ask the McDonald's worker fighting for a livable wage and full time hours if social class doesn't exist.

And this is where we turn to the man on the street. Ask me what the markers were of social class in my life growing up and the first thing I think about is that "rich people have family photos that don't have white borders around them" because they had 35 mm cameras instead of instamatics and they had dark rooms, or they hired people to take professional photos of them. If the photos were in black and white, they were that much classier.

So there is a phenomenological experience of social class, whether or not academics agree on a definition.

And you will find this with many aspects of the social sciences.
posted by vitabellosi at 6:59 AM on May 7, 2014

Best answer: Andy Field's books are good. His website statistics hell.com is a good first stop too.
posted by k8t at 7:15 AM on May 7, 2014

Response by poster: To clarify: I'm an incoming PhD student in the social sciences, so I've already accepted the utility and goals of social science research. I'm more interested in applied methodology than philosophy of science.
posted by MisantropicPainforest at 7:16 AM on May 7, 2014

Best answer: This isn't an edited volume and might be too basic for what you want but Inequality by Design which is not edited, but was authored by the entire UC Berkeley Sociology department faculty does some of this. It is essentially a take-down of The Bell Curve, showing for a non-specialist audience how the analyses in the Bell Curve were flawed and re-doing them to show that when analyzed properly the data show a very different result. "Public Sociology" which includes making sociological findings accessible to non-sociologists is a big part of UC Berkeley's identity as a department, so this book is very much about teaching non-sociologists how sociologists analyze data.
posted by If only I had a penguin... at 7:29 AM on May 7, 2014 [2 favorites]

You might want A Quantitative Tour of the Social Sciences by Gelman and Cortina.
posted by madcaptenor at 8:05 AM on May 7, 2014

Maybe try the Handbook of Quantitative Criminology.
posted by _cave at 8:39 AM on May 7, 2014

Because so much social science is tied to causal claims, I recommend taking a look at Berk's Regression Analysis: A Constructive Critique, Spirtes et al.'s Causation, Prediction, and Search, and to a lesser degree Morgan and Winship's Counterfactuals and Causal Inference.
posted by Jonathan Livengood at 8:58 AM on May 7, 2014

Also not an anthology, but I really like Methods Matter by Murnane and Wilett.
posted by plantbot at 9:16 AM on May 7, 2014

I'm a social sciences PhD student and I was assigned Rethinking Social Inquiry for a methods class. It's been very helpful.
posted by pantarei70 at 12:01 PM on May 7, 2014

Have you taken any statistics or quantitative social science courses already? In college or elsewhere?
posted by clockzero at 12:13 PM on May 7, 2014

Response by poster: yes
posted by MisantropicPainforest at 12:46 PM on May 7, 2014

I am wondering if How to Lie with Statistics would interest you.
posted by Michele in California at 4:29 PM on May 7, 2014 [2 favorites]