33:630:385 Lecture Notes - Lecture 9: Statistical Hypothesis Testing

• School: Rutgers University
• Department: Marketing
• Course: 33:630:385
• Professor: Richard Schaeffer

Document Summary

Determines if mean for one group staisically difers from mean of second group: for descripive data, test typically used to compare two market segments, for causal/experimental data, test used to compare control and treatment group. Step 2: reject ho if t stat suggests so t stat > t crit. Or: reject if, if t stat farther from zero, treatment and control samples difer greatly and stronger t stat < -t crit evidence that treatment would afect outcome in enire populaion. Step 1: build alternate cross tab consistent with no relaionship. Step 2: compare actual data to alternate cross tab. Step 3: reject ho if actual data is suiciently diferent from alternate cross tab: reject if, larger x2, then treatment and control samples difer greatly and stronger evidence that. X2 > x2 crit treatment would afect outcome in enire populaion. Esimate of coeicient b yields causal efect observed in sample. Causal efect of x(treatment) = o2 o4.