Statistical techniques do not magically eliminate dependence between measurements or allow you to obtain good results with poor experimental design. They merely provide ways to quantify dependence so you can correctly interpret your data. (This means they usually give wider confidence intervals and larger p values than the naive analysis.)
