Infant survival. Data on the survival of 715 infants in relation to the amount of medical care received are as follows:

Crude data (clinics 1 and 2 combined)

Did not survive Survived Less care 20 373 More care 6 316 Data from Whittaker, J. (1990). Graphical Models in Applied Multivariate Statistics. Chichester:

John Wiley & Sons; Bishop, Y. M. M., Fienberg, S. E., & Holland, P. W. (1975). Discrete Multivariate Analysis: Theory and Practice. Cambridge, MA: MIT Press.

(a) Determine the proportion of infants who did not survive in each group. Report the association in terms of a risk ratio. What do you conclude from this crude comparison?

(b) The experience within the two clinics that contributed to the earlier table is shown in the following tables. Address the effect of receiving less care within each clinic. What do you now conclude?

(c) How do you explain the apparent contradictory results of the crude analysis in part

(a) and the stratified analysis in part (b)?

(d) Using a computer application such as WinPEPI or OpenEpi.com, test for interaction in the risk ratios. Show all hypothesis testing steps.

(e) Calculate the Mantel–Haenszel summary risk ratio associated with less care. Include a 95% confidence for the RR. Summarize your findings.