recommend
Recommends the number of intervals for a chi-squared goodness of fit test based on the supplied sample size. This does not assume anything about the distribution. An adjustment ensures that there are at least 3 intervals based on other theory.
On the Choice of the Number and Width of Classes for the Chi-Square Test of Goodness of Fit Author(s): C. Arthur Williams, Jr. Source: Journal of the American Statistical Association , Mar., 1950, Vol. 45, No. 249 (Mar., 1950), pp. 77-86 Published by: Taylor & Francis, Ltd. on behalf of the American Statistical Association Stable URL: http://www.jstor.com/stable/2280429