ST 307 Lecture 9: 9 Inference for Two means and paired data in SAS
Document Summary
Hypothesis tests vs confidence intervals - 2 major methods if inference. Make an assumption about the population, check if data contradicts. Give range of values we are confident contains the truth. Analyzing data roughly follows the 3 step process below: read in data, summarize data numerically and graphically, analyze data to answer questions of interest, just need appropriate proc. Observations are independent and identically distributed (random sample) Used to do ht and ci for mean(s) from normally distributed populations for: P-value here is not significant, meaning that we might be comfortable using either test. If there is a rough straight line, that indicates no large departure from normality. Reasonable amount of curvature have trepidation about our conclusions. So some kind of modifications of data and rerun analysi. The class statement in the proc ttest procedure defines groups for comparisons. Two observations on same or very similar person/unit. Inference for mean difference just like a one-sample test.
