categorical independent variable ◦ e.g., men versus women, control vs. experimental groups continuous dependent variable ◦ e.g, # times purchased, $ spent comparisons of means ◦ men 4.0, women 5.0 Marketing Research 7/17/2016 1 You have two groups and a mean (average) for each ◦ e.g., men = 4.0, ◦ women = 5.0 How do you determine the strength of the covariation? Marketing Research 7/17/2016 2 Situation #2 Situation #1 W W M W M W M W M W 1 2 3 4 5 6 7 8 Mean: M = 4, W = 5 M M MW 1 2 3 4 Mean: M = 4, Marketing Research 5 6 7 8 W=5 7/17/2016 3 How do you test the covariation? 1. “inter-ocular” test does it “hit you between the eyes?” does it look big? Marketing Research 7/17/2016 4 2. Formal t-test ◦ use the numbers in the sample ◦ scale by the “spread” [variance] ◦ (use the “standard error”) ◦ how many standard errors apart? Marketing Research 7/17/2016 5 Marketing Research 7/17/2016 6 Marketing Research 7/17/2016 7 2. t-test ◦ Formula: X1 - X2 S.D./sqrt of n [number of subjects] Marketing Research 7/17/2016 8 2. t-test (continued) ◦ compare observed “t” ◦ to t “critical” from table [A-4] d.f. = n [number of subjects] - 1 e.g., t [29] @ .05 = 1.699 [two tailed] t [29] @ .025 = 2.045 [one tailed] ◦ if t > t critical, difference in population Marketing Research 7/17/2016 9 dianaid Frequency 60 40 20 Mean =3.02 Std. Dev. =0.859 N =755 0 1.00 2.00 3.00 4.00 5.00 dianaid Marketing Research 7/17/2016 10 Group Statistics dianaid sex male female N 277 461 Mean 2.6136 3.2656 Std. Deviation .82723 .78627 Std. Error Mean .04970 .03662 Independent Samples Test Levene's Test for Equality of Variances F dianaid Equal variances as sumed Equal variances not ass umed 1.634 Sig. .202 t-test for Equality of Means t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper -10.697 736 .000 -.65207 .06096 -.77175 -.53239 -10.562 558.271 .000 -.65207 .06174 -.77334 -.53080 Marketing Research 7/17/2016 11 Marketing Research 7/17/2016 12 Marketing Research 7/17/2016 13 A useful rule of thumb is: ◦ the difference in standard deviations is seldom a problem until one is more than twice the other. In that instance, do a t-test using “separate” variance estimates. Marketing Research 7/17/2016 14 Independent Samples Test Levene's Test for Equality of Variances F dianaid Equal variances as sumed Equal variances not ass umed 1.634 Sig. .202 t-test for Equality of Means t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper -10.697 736 .000 -.65207 .06096 -.77175 -.53239 -10.562 558.271 .000 -.65207 .06174 -.77334 -.53080 Marketing Research 7/17/2016 15 Marketing Research 7/17/2016 16