BRIEF ANSWERS TO PRACTICE PROBLEMS: SAS code given below! Practice Problem ------------------- (a) See page 536 for the form of the logistic regression model. (b) See SAS code for model fitting. Note beta_0_hat = -4.048 and beta_1_hat = 0.057. (c) Hosmer-Lemeshow test has a P-value of 0.5403, so we conclude the logistic model is appropriate. (d) The Wald test about beta_1 has a P-value of 0.0149, so at alpha = .05, we conclude ventricle size has a significant effect on the probability of abnormal brain wave behavior. (e) The estimated odds ratio is 1.059, so we estimate that the odds of having an abnormal brain wave reading increases by 105.9% (by a factor of 1.059) when the ventricle size increases by one unit. (f) Since the estimate of beta_1 is positive, a person with a smaller venticle size is estimated to have a lesser probability of abnormal brain wave behavior than a person with a larger ventricle size. (g) Run the SAS code below to see the plot (it is a monotone increasing curve). /* ************ SAS code *************************** */ data brain; input VENT EEG; cards; 53 0 37 0 63 0 25 0 60 0 58 0 56 0 59 0 50 0 58 1 70 0 68 1 50 0 59 0 51 0 76 0 74 1 62 1 41 0 65 0 50 0 94 1 73 1 72 0 45 1 56 0 56 0 75 0 76 0 78 1 50 0 68 0 47 0 66 0 42 1 76 1 57 0 65 0 51 0 83 1 51 0 80 1 70 0 68 1 49 0 56 1 58 1 58 1 64 1 60 1 57 0 54 0 58 0 63 1 61 0 70 0 40 0 51 1 58 1 70 1 57 1 84 0 58 0 51 1 57 0 85 1 50 0 48 0 67 1 62 0 65 0 ; run; PROC LOGISTIC DESCENDING DATA=brain; MODEL EEG = VENT / LACKFIT; OUTPUT OUT=NEW P=PRED L=LOWER U=UPPER; RUN; PROC SORT DATA=NEW; BY VENT; symbol1 i = join v=circle l=32 c = black; symbol2 i = join v=star l=32 c = black; symbol3 i = join v=star l=32 c = black; PROC GPLOT DATA=NEW; PLOT PRED*VENT LOWER*VENT UPPER*VENT / OVERLAY; RUN;