DRAFT
Table 3 contains the MNL results presented as relative risk ratios. Estimates are reported
for the never and irregular screeners categories, with regular screeners as the omitted
reference category.
INSERT TABLE 3 AROUND HERE
Age is a significant predictor of screening behaviour amongst the never screeners. Figure
2 shows the average predicted probabilities for women of different ages (within the target
age range). Amongst the never screening group, the relative risk ratio for the age square
variable is greater than unity (P < 0.01), indicating that age has a positive quadratic
effect. The age co-efficient was not significant for irregular screeners.
Having been born overseas or speaking a language other than English at home appeared
to increase the probability of belonging to the never or irregular screener groups, but
neither reached significance. Education only mattered amongst women whose highest
level of education was completion of high school between years 7 and 10. The odds of
screening irregularly are about 28% lower for a woman in this group relative to a woman
with a university degree (p < 0.10). None of the education variables reached significance
in the never screened group.
Women in higher income households were less likely to screen irregularly, compared to
women from households with an annual income of less than $10,000. Figure 2 shows the
average predicted probabilities for women in different income groups for the never and
irregular screeners. Four out of five income groups were statistically different from unity
(p < 0.10). Using the regular screeners as the base case, we find a relative risk ratio of
0.61 for the ‘$20,000 to $40,000’ household income group, suggesting that the odds of
screening irregularly are about 39% lower for those women compared to women with
household income less than $10,000. Further, we find some evidence of a negative
income gradient for the irregular screener group, with respective relative risk ratios
decreasing with each subsequent increase in household income. By comparison,