1

I have two histograms shown below. enter image description here

enter image description here

As you can see, the two histograms are different in appearance. However, when I calculate the mean of the two histograms, they are pretty similar. The mean for the top graph is close to 0.500 and the Mean for the bottom graph is 0.478. I'm trying to figure out a way to capture this difference in some sort of measurement. The mean is obviously not too descriptive of the data. Is there any way to calculate a measurement that will more closely depict the difference in the two histograms data?

I'm sorry if this post is confusing. I'm having trouble wording me question. I had someone suggest to do a distance from middle measurement, but when I tried to look up what that entailed, I couldn't find anything.

cosmictypist
  • 143
  • 1
  • 6
  • 4
    You can get bigger differences than that with the appearance of two histograms on *exactly the same data*, just by changing where the bin starts, or by changing the bin-width. See examples [here](http://stats.stackexchange.com/questions/51718/assessing-approximate-distribution-of-data-based-on-a-histogram/51753#51753) (the first example involves shifting the data slightly to the left each time and keeping the bins constant but the effect is the same if you keep the data constant and shift the bins slightly to the right. The second example (small change of bin width) is pretty striking. ... ctd – Glen_b Aug 13 '15 at 12:38
  • ctd ... I realize you have the same bins though, which does point to a real difference. A 'distance from the middle' would be something like a mean deviation or a standard deviation. However, there's also differences in skewness and kurtosis. What you might use to capture the change depends on which aspects of the change are the most important to you. – Glen_b Aug 13 '15 at 12:43
  • @Glen_b I will look into the SD to see if it is a more descriptive measurement. I do have the same bins and I do believe there is some difference in the two datasets. Thank you for your help. – cosmictypist Aug 13 '15 at 12:49
  • @Glen_b the skewness is what I'm the most interested in. Should I still be looking into the SD? – cosmictypist Aug 13 '15 at 12:50
  • I guess there is a skewness measurement I could use – cosmictypist Aug 13 '15 at 13:06

0 Answers0