Which type of data needs to be collected to conduct the study of the significant difference between statistics test scores of college students who have had early statistic exposure than those college students without? and what techniques are best for collecting this kind of data?