MAST10011 Chapter Notes - Chapter 7: Statistical Unit, Confidence Interval, Test Statistic
Paired comparisons: the two treatments are applied to pairs of experimental units which have been matched so as to be as alike as possible (even the same experimental unit at different times) Self pairing: measure(cid:373)e(cid:374)ts are take(cid:374) o(cid:374) a si(cid:374)gle su(cid:271)je(cid:272)t at two disti(cid:374)(cid:272)t poi(cid:374)ts i(cid:374) ti(cid:373)e ((cid:862)(cid:271)efore a(cid:374)d after(cid:863) a treatment or a cross-over design) or at two different places (e. g. left and right arms for a skin treatment). Matched pair: two individuals are matched to be alike by certain characteristics under control (e. g. age, sex, severity of illness, etc). Then the people in the pair are each assigned to one of two treatments at random. Calculate the confidence interval of the mean differences. Independent samples: the two treatments are applied to separate sets of experimental units randomly selected from the sample population. Consider the difference between samples (i. e. compare sample means/proportions/medians etc) Assume the underlying populations of both samples are normally distributed , null: u1=u2.