How to analyse differences in groups over time? I have a data set which looks at the effect of diet on 3 different vitamin concentrations (eg. vitamin d, c, iron). There are two measurements for each vitamin - one at the start of the study and one after one month. These measurements are also categorised into male and female samples (10 of each).
Basically, I have to analyse the difference between the male and female vitamin concentrations as well as the differences found between the first measurement and last measurement (separately for males and females), and I'm having trouble figuring out which tests would be best.
Any help much appreciated. 
 A: Being two different populations, I'm not totally sure I follow your question. Males never become females, except for statistically negligible exceptions, so what kind of cross-effect are you looking to model?
It somewhat matters how many datapoints you have, but it seems you have datapoints that describe two distributions: before and after. These are for two distinct populations: male and female. You seem to be interested in studying the difference between these distributions. Getting those distributions from the datapoints is necessary, also.
One model family that fits both requirements, if you have several levels of vitamins you are measuring, is clustering. Many clustering algorithms will work for this, for example K-means. Basically, just cluster your before and after data, match clusters that are closest to each other, and describe how much they moved. Sometimes the clusters don't capture "movement" so much as changes in cluster populations, but it sounds like a good bet for your situation.
