BUSS1020 Lecture Notes - Lecture 10: Null Hypothesis, Customer Service Training, Repeated Measures Design

• School: University of Sydney
• Department: Business Studies
• Course: BUSS1020
• Professor: All

Document Summary

Buss1020 lecture 10 hypothesis test: two sample. Test hypothesis / form a confidence interval for difference b/w 2 population means, (cid:1005) (cid:1006) Point estimate for the difference = x 1 x 2. Different data sources are unrelated & independent. Sample selected from 1 population has no effect on sample selected from another population. Are unknown, assumed equal use sp to esti(cid:373)ate u(cid:374)k(cid:374)o(cid:449)(cid:374) . A(cid:396)e u(cid:374)k(cid:374)o(cid:449)(cid:374), (cid:374)ot assu(cid:373)ed e(cid:395)ual use (cid:1005) & (cid:1006) to esti(cid:373)ate u(cid:374)k(cid:374)o(cid:449)(cid:374) (cid:1005) & (cid:1006) Reject h0 if t(stat) < -t critical value ( (cid:895) Reject h0 if t(stat) > t critical value ( (cid:895) Reject h0 if t(stat) < -t critical value ( /(cid:1006)(cid:895) o(cid:396) t(stat) > t critical value ( /(cid:1006)(cid:895) Hypothesis tests for (cid:1005) (cid:1006) with (cid:1005) & (cid:1006) u(cid:374)k(cid:374)o(cid:449)(cid:374) & assu(cid:373)ed e(cid:395)ual. Populations are normally distributed / both samples are large enough for central limit theorem (e. g. at least 20) Population variances are unknown but assumed equal.