Sunday, February 3, 2013

If you are doing runs late at night and exclaiming to yourself "holy shit - why am I seeing seven times as many engineers in this data as are supposed to be employed in the U.S.?"....

...make sure you remind yourself whether you're using your seven year stacked dataset or your annual dataset before you worry too much.

3 comments:

  1. That's like when I was working on the data for Zac's and my War and Presidential Greatness paper and, working late at night, I kept coming up with one too few presidents. The reason I finally figured out the next a.m.: Grover Cleveland's 2 non-consecutive terms.

    ReplyDelete
  2. What software are you using for your data analysis Daniel? I think one of the lead developers for [statsmodels](http://statsmodels.sourceforge.net), a statistics/econometrics library for the Python programming language, is a grad student at American. Skipper Seabold is his name.

    ReplyDelete
    Replies
    1. I use stata. I've definitely heard of Skipper but have never met him myself.

      I've used Python some... I can see why people like it, but I just don't have a programmers brain - I haven't done much with it outside of what is required for class and haven't done any stats with it.

      Stata is great for econometric analysis and it is always getting better. I have no complaints about the software... it's user error and forgetfulness that was confusing me last night :)

      Delete

All anonymous comments will be deleted. Consistent pseudonyms are fine.