Is participating in a follow-up survey associated with having medical aid? A South African study compared whether participating in a 5-year follow-up interview was associated with having medical aid (a type of health insurance). Based on the setup of the problem in the original source, whether an individual was followed-up is the explanatory variable and receipt of medical aid is the response variable. The extraneous variable (“potential confounder”) is race.

(a) Table 19.5A tabulates the results for all subjects combined.
Calculate the prevalence of medical aid in the group that was followed up and the group that was lost to follow up. Describe the relationship between follow up and having medical aid.

(b) Subjects were subclassified by race. Table 19.5B shows the results for white participants. Describe the relationship in this stratum.

(c) Data for black participants are shown in Table 19.5C. Analyze the results for this stratum.

(d) Why did there appear to be a negative association in the crude data and no association within strata?