Friday, July 19, 2013

Is There Enough Caffeine in the World?

I am in codebook hell.

I'm working on a rather complicated research project that I'll explain briefly below, but basically it requires me knitting together national survey data from the 2000, 2004, 2008, and 2012 presidential elections.  Not just that, but a pre-election survey and a post-election survey for each year (same people before and after) -- essentially, eight surveys.  National surveys.  With hundreds of questions each time.

How big are these?  The PDF of the 2004 survey alone is 754 pages.  I just happened to have it up right now.  Go ahead, look at it.  Feel my pain.  Or at least my need for caffeine.

My plan is simple.  Each pre- and post-election survey has a number of questions similar, if not identical, to those asked in other years.  I'll slowly but ever so surely create SPSS files of each year and then, with luck, merge them into a single file for analysis.  This is harder than it sounds.  Take political interest, for example.  A simple yet important concept.  I have to find the same worded question in all surveys.  In 2004, for example, it's VARIABLE 045057.  I gotta label each one appropriately (POLINT04 for the 2004 version, POLINT08 for 2008, you get it) and make double sure they are identical not only in wording but also in what we in the biz call response alternatives (4=high, 3, 2, 1, etc.).  And then you have to do a merge and, honest to God, hope for the best. 

Today I've mostly worked on 2004 data.  I'll pull from the bejillion available questions about a hundred or so I may or may not use to build multivariate models.  For each year.  Label 'em, merge 'em, analyze the hell out of 'em.  Oh, and write stuff too. 

It's all part of my research sabbatical.  I'm not teaching this Fall, making it perhaps my best teaching performance semester ever, a "do no harm" kinda thing.  The study?  Simply put, it has with the notion that a democracy relies on the consent of the losers -- and a closer look at those who expected to win, but didn't, and just how consenting they are after an election outcome.  There's gobs of theory to appease the academic gods, not to mention it'll play well in the press.

If it all works, that is.

No comments: