The problem of multicolinearity is a persistent one, usually bringing devestation with it for analysts unable to detect/treat it. Perform a preliminary search of cases in which multicolinearity presents an issue for researchers/analysts, and explain potential sources of the multicolinearity (using rigorous reasons discussed in chapter 9.3).

Now consider the patient-satisfaction data from the Topic 2 assignment. Analyze this data for near-linear dependence by using Equation 9.3 to compute the variance inflation factor. Does this data exhibit multicollinearity?

Satisfaction Age Severity Anxiety

68 55 50 2.1

77 46 24 2.8

96 30 46 3.8

80 35 48 4.5

43 59 58 2

44 61 65 5.1

26 74 60 5.5

88 39 42 3.2

75 82 42 3.1

57 27 50 2.4

56 51 38 2.2

88 53 30 2.1

88 41 31 1.9

102 37 34 3.1

88 24 30 3

70 27 48 4.2

82 50 61 4.6

43 58 71 5.3

46 60 62 7.2

56 62 38 7.8

59 68 41 7

26 70 66 6.2

52 79 37 4.1

73 39 42 3.5

Equation 9.3 below

When there are more than two regressor variables, multicollinearity produces similar effects. It can be shown that the diagonal elements of the C = (X'X)-1 matrix are Cji = j = 1, 2, ..., P (9.3) 1-R