# 25th and 75th percentile according to wolfram alpha

I'm just studying my statistics exams and I came across something odd.

When I try to get the 25th or 75th percentile with wolfram alpha, it gives me another answer than I should get according to my textbooks.

according to my textbook, the answer should actually be 37 and not 39.5

Does anyone have any idea what wolfram alpha does different?

• – Henry Dec 13 '11 at 23:09

There are at least 9 [different definitions of empirical quantiles, see Wikipedia or the R manual (i.e., ? quantile). R computes the 25th percentile of your data as 37 if you specify type=6 (like Minitab and SPSS) and 39.5 if you specify type=5 (piecewise linear function):
> quantile(x=c(32,42,46,46,54), probs=0.25, type=5)