Question: This problem uses an SPSS data set called HW5_2.sav. A survey was conducted to examine the effects of pharmaceutical drug advertising on physician practices. Self-administered questionnaires were delivered to randomly selected physicians in various practice types: general practice, family practice, and internal medicine. Physicians reported on their attitude toward drug advertising. Likert scale measurements were used in the survey (1=strongly disagree and 5=strongly agree; the higher the score, the more favorably disposed the respondent is to advertising). See page 469 for the data.

a. Test whether the physician types differ significantly in average attitude toward advertising. State your null hypothesis and the result of the test (use ?=0.05) (use either the GLM procedure or the one-way ANOVA procedure in SPSS).

b. Using dummy variables, create an appropriate regression model for this situation (use either reference cell coding or effect coding, but not both!). Interpret the regression coefficients for your model and verify that the omnibus F value is the same in regression as when you conducted the ANOVA in part a).

c. Use the Brown-Forsythe test to assess whether or not the assumption of homogeneity of variance is violated. Report and interpret your results.

d. Use Tukey’s HSD test to locate differences between pairs of means.