Why is missing data a problem?

Why is missing data discussed here and not earlier? Why is it a problem with repeated measures?

Missing data is always a problem, but it is just discussed here.
In a between-groups design, it is okay to have a slightly different number of subjects in each group, so if one subject drops out in one of the conditions then that group has just one less subject. Now you want to look at how subjects change over time and the different scores between two or more conditions for each subject.
In a within-subjects design, it is okay to have a slightly different number of subjects in each group, so if one subject drops out in one of the conditions then that group has just one less subject. Now you want to look at how subjects change over time and the different scores between two or more conditions for each subject.