comment by b

a thoughtful web.

Good ideas and conversation. No ads, no tracking. Login or Take a Tour!

comment by b_b

b_b · 4296 days ago · link · · parent · post: Mike Konczal: Researchers Finally Replicated Reinhart-Rogoff, and There Are Serious Problems.

I'm sure Fox News will be all over this just like they were with "Climategate"! We know they just want academic accountability.

But seriously though, this paper was obviously an attempt at academic fraud, given that no economist can go through even undergraduate without rigorous statistical training.

markup tips · 0

mk · 4296 days ago · link ·

It's difficult to see it any other way. Leaving out 5 countries from the average, weighting one year of New Zealand to 19 years of the U.K... this looks pretty foul.

In this day and age, all publications should come with the original dataset.

+discuss+discuss

–

user-inactivated · 4294 days ago · link ·

To be fair, they only left out one actual data point from their average. The rest of the countries show n.a. for the data point where the average is calculated from.

+discuss+discuss

–

mk · 4294 days ago · link ·

Yet because there are only 7 data points possible, it completely changes the result. This is a very high profile economic report coming from two very influential Harvard economists. Consider that they accidentally left out their -7.9 data point from New Zealand:

  Their 6 data points: ave: -0.07, s.d.: 3.83
  All 7 data points: ave: 0.26, s.d.: 3.67
  No NZ data point: ave: 1.43, s.d.: 1.73

Not only does removing the NZ data point instead make the average much higher, the data looks more reliable than the full dataset as the standard deviation has almost been cut in half. Leaving out one data point from such a small set has big consequences.