Data Set 5 This data set is discussed in Chapter 8. The study involves 117 persons in 39 matched sets, each strata containing 3 persons, 1 of whom is a case and the other 2 are matched controls. The disease variable is myocardial infarction status (MI). The exposure variable is smoking status (SMK). Four variables are involved in the matching, namely, age, race, sex, and hospital status.

Two additional independent variables, systolic blood pressure (SBP) and electrocardiogram status (ECG), are not involved in the matching.

Data set name: my.dat Matched Case set Person status SMK SBP ECG 1 1 1 0 160 1 1 2 0 0 140 0 1 3 0 0 120 0 2 4 1 0 160 1 2 5 0 0 140 0 2 6 0 0 120 0 3 7 1 0 160 0 3 8 0 0 140 0 3 9 0 0 120 0 4 10 1 0 160 0 4 11 0 0 140 0 4 12 0 0 120 0 5 13 1 0 160 0 5 14 0 0 140 0 5 15 0 0 120 0 6 16 1 0 160 0 6 17 0 0 140 0 6 18 0 0 120 0 7 19 1 0 160 0 7 20 0 0 140 0 7 21 0 0 120 0 8 22 1 0 160 0 8 23 0 0 140 0 8 24 0 0 120 0 9 25 1 0 160 0 9 26 0 0 140 0 9 27 0 0 120 0 10 28 1 0 160 0 Matched Case set Person status SMK SBP ECG 10 29 0 0 140 0 10 30 0 0 120 0 11 31 1 0 120 1 11 32 0 0 120 0 11 33 0 0 120 0 12 34 1 0 120 0 12 35 0 0 120 0 12 36 0 0 120 0 13 37 1 0 120 0 13 38 0 0 120 0 13 39 0 0 120 0 14 40 1 0 140 0 14 41 0 0 140 0 14 42 0 0 140 0 15 43 1 0 120 1 15 44 0 0 140 15 45 0 0 160 0 16 46 1 0 120 1 16 47 0 0 140 16 48 0 0 160 17 49 1 1 160 1 17 50 0 0 140 0 17 51 0 0 120 0 18 52 1 1 160 18 53 0 0 140 0 18 54 0 0 120 0 19 55 1 I 160 0 19 56 0 0 140 19 57 0 0 120 0 20 58 1 1 160 1 20 59 0 0 140 1 20 60 0 0 120 0 21 61 1 160 0 21 62 0 0 140 0 21 63 0 0 120 0 22 64 1 120 0 22 65 0 0 120 0 Matched Case set Person status SMK SBP ECG 22 66 0 0 120 0 23 67 1 1 140 0 23 68 0 0 140 0 23 69 0 0 140 0 24 70 1 1 120 0 24 71 0 0 140 0 24 72 0 0 160 0 25 73 1 1 120 0 25 74 0 0 160 0 25 75 0 0 140 0 26 76 1 0 160 0 26 77 0 1 140 0 26 78 0 0 120 0 27 79 1 0 120 0 27 80 0 1 120 0 27 81 0 0 120 0 28 82 1 0 160 1 28 83 0 0 140 0 28 84 0 1 120 0 29 85 1 0 160 0 29 86 0 0 140 0 29 87 0 1 120 0 30 88 1 0 120 0 30 89 0 0 140 0 30 90 0 1 160 0 31 91 1 0 140 0 31 92 0 0 140 0 31 93 0 1 140 0 32 94 1 1 160 1 32 95 0 1 140 0 32 96 0 0 120 0 33 97 1 1 160 1 33 98 0 1 140 33 99 0 0 120 0 34 100 1 1 120 1 34 101 0 1 120 1 34 102 0 0 120 1 Matched Case set Person status SMK SBP ECG 35 103 1 1 160 0 35 104 0 0 140 0 35 105 0 1 120 0 36 106 1 0 160 1 36 107 0 140 1 36 108 0 1 120 1 37 109 1 0 120 0 37 110 0 1 140 0 37 111 0 1 160 0 38 112 1 1 160 38 113 0 1 140 0 38 114 0 1 120 0 39 115 1 1 120 0 39 116 0 1 120 0 39 117 0 1 120 0 SampIe Questions A. Set up a data file than can be used to carry out conditional ML estimation of a logistic model for these data.
B. Use cor.ditional ML estimation to fit a model that allows for both confounding and interaction involving the variables SBP and ECG. Test for the significance of the interaction terms.
C. Use conditional ML estimation to fit a logistic model that contains main effects. What odds ratio dso you obtain for this model? Is there confounding due to SBP and or ECG?