The following table compiles data from six studies designed to investigate the accuracy of death certificates. The causes of death listed on 5373 death certificates were compared to autopsy results. Of those considered, 3726 certificates were confirmed to be accurate, 783 either lacked information or contained inaccuracies but did not require recoding of the underlying cause of death, and 864 were incorrect and required recoding [261].

(a) Do you believe that the results are homogeneous or consistent across studies? Explain.

(b) It should be noted that autopsies are not performed at random; in fact, many are done because the cause of death listed on the certificate is uncertain. What problem might arise if you attempt to use the results of these studies to make inference about the population as a whole?

Death Certificate Status Confirmed No Recoding Total Hospital Accurate Change 1955-1965 2040 367 327 2734 1970 149 60 48 257 1970-1971 288 25 70 383 1975-1977 703 197 252 1152 1977-1978 425 62 88 575 1980 121 72 79 272 Total 3726 783 864 5373