QM201 TA Section 1 Jieling Han Ch10 Hypothesis Testing 1. GOAL (USE µ AS AN EXAMPLE) Goal: Decide between two possibilities (about the population parameter) Null hypothesis H0: µ=µ0 (States the prior value of the parameter) Research hypothesis H1: µ≠ µ0 (Negates the null hypothesis) µ0 is called reference/stated value 2. HOW TO DECIDE (USE µ AS AN EXAMPLE) Confidence Interval (e.g 95% or 5% test level): Is µ0 in [ X tS X , X tS X ]? Yes, accept Ho o The sample average is not significantly different from the stated value µ0 (likely an estimation error) o A weak/not strong conclusion, since the probability of making Type II Error cannot be estimated / controlled. No, reject Ho and accept H1 o The sample average is significantly different from the stated value µ0 (likely a material error) o A strong conclusion, since the probability of making Type I Error can be estimated / controlled. 3. ERRORS IN HYPOTHESIS TESTING (USE 5% TEST LEVEL AS AN EXAMPLE) TRUTH DECISION NULL HYPOTHESIS: H0 RESEARCH HYPOTHESIS: H1 ACCEPT H0 CORRECT DECISION TYPE II ERROR The probability of making Type II Error CANNOT be determined. REJECT H0, AND ACCEPT H1 TYPE I ERROR CORRECT DECISION Using 5% test level / 95% CI, the probability of making Type I Error is 5%. Only 95% chance the CI contains µ, so 5% chance the CI doesn’t contain µ, which is the probability of making Type I Error. Ch10 Hypothesis Testing