This preview shows pages 7–16. Sign up to view the full content.
∙Suppose productivity is expensive to measure, so the firm chooses arelatively small sample of workers. The random sample isXi:i1,...,n, whereXiis the change in workeri’s productivity.7
This preview
has intentionally blurred sections.
Sign up to view the full version.
∙Here the null hypothesis might be that there was no effect onproductivity and the alternative that there was some effect (positive ornegative):H0:0and the alternative hypothesis isH1:≠0∙Of course, the firm is interested in whether0 or0 ifH0isfalse.8