topic What is wrong with kurtosis in Geostatistical Analyst? in ArcGIS GeoStatistical Analyst Questions

What is wrong with kurtosis in Geostatistical Analyst?

MilanDjordjevic — Sun, 24 Nov 2013 12:14:21 GMT

When I compare values of descriptive statistics of my data from Excel and Minitab with statistics from Geostatistical Analyst (Histogram) there is a huge difference.

results from Excel and Minitab:
Variable Skewness Kurtosis
C1           -0.00     -0.82
C2           -0.82     -0.22
C3           -0.49     -1.03
C4           -0.21      0.49
C5           -1.45      2.10
C6           -0.26     -0.29

results from Geostatistical Analyst - Histogram:
Variable Skewness Kurtosis
C1            0.00      2.16
C2           -0.78      2.69
C3           -0.49      1.96
C4           -0.20      3.38
C5           -1.40      4.83
C6           -0.23      2.66

Those results shown in Geostatistical Analyst - Histogram are for the same data with no transformation performed. Why is there any difference?

Re: What is wrong with kurtosis in Geostatistical Analyst?

DanPatterson_Retired — Sun, 24 Nov 2013 12:54:32 GMT

check the actual formulae being used, there are variants, have a look at this as well
http://resources.arcgis.com/gallery/file/geoprocessing/details?entryID=BDF21676-1422-2418-34E6-A29C6BAC8A0B

Re: What is wrong with kurtosis in Geostatistical Analyst?

EricKrause — Mon, 25 Nov 2013 17:32:22 GMT

We use a simple kurtosis formula in Geostatistical Analyst. The formula is the fourth moment divided by the square of the second moment. We do not correct for small samples or calculate excess kurtosis.

It appears that the KURT function in Excel calculates excess kurtosis and corrects for small sample sizes. In general, the difference between these two calculations should be approximately 3. You can see from your data that if you add 3 to the Excel values, you will get close to the ArcGIS values (the remaining difference is due to small sample corrections in Excel).