A bit off topic, but one of my favorite parts of any social science paper is when the authors describe how they developed the dataset. In this case, it's a clev

A bit off topic, but one of my favorite parts of any social science paper is when the authors describe how they developed the dataset. In this case, it's a clever linking of tax forms, Census Bureau surveys of business owners, and state unemployment insurance databases. It's a neat solution that the authors are justifiably proud of. But since it's considered passé to talk about your data too much at the expense of your results, it only gets a couple pages (pages 9-10) of the paper even though I'm sure they spent many man-months compiling and cleaning up the data.