It's not a problem with the median, any more than it is a problem with the mean to say that 44.14 represents the data set. The problem is in the nature of the data set, which is not homogeneous and independent. It is highly autocorrelated. It doesn't matter which descriptive statistic you use. You're screwed.

For homogeneous and independent data, it matters very little whether you use mean or median, unless you have outliers you want to more or less ignore. Median is frequently much easier to get when in the field.


Last edited by denton; 03/16/14.

Be not weary in well doing.