Skip to main content

View Diary: Wall of Shame: Republicans Lose Intellectual Cover for Austerity/Spending-Cut Policies (87 comments)

Comment Preferences

  •  Lots more wrong (6+ / 0-)

    Go to the diaries here yesterday. (I'm too tired to dig out the links.) One key problem is often called "the unit of analysis.") As Konczal points out, they counted a few years for (maybe) NZ and averaged this with (maybe) 19 years of UK data. They also used "unweighted" means, where data from small countries - with small economies - counted the same as the larger countries. I believe the data they "missed" came from just post-WW2, where many countries with large debt/GDP fractions also had booming economies.  

    If there's fraud, in addition to honest mistakes, in their data analysis, it will show in where they chose the cutoff points.  It's long been argued that splitting data into high, medium, and low groups, especially with attention to cutoffs that amplify the differences authors want to claim, is bad practice.  If data points just below their 90% level, in addition to "Excel coding errors," show higher than 90%-level average GDP growth, it is entirely possible that they picked cutoffs opportunistically.  

    Your joke about Harvard students unable to count and MIT students unable to read is new to me and funny. But, if you can recall or research Cyril Burt, you'll see that there are data forensics that can be used to separate mistakes from misrepresentation. I have no doubt that Harvard will be taking a close look at what went on here.  

    Even if it turns out these were all honest mistakes, you have to be simultaneously saddened and also have a wry smile --- economists are supposed to be competent with data, right?

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site