Adams M M, Scherr P A, Branch L G, Hebert L E, Cook N R, Lane A M, Brock D B, Evans D A, Taylor J O
Channing Laboratory, Department of Medicine, Harvard Medical School, Boston, MA.
Public Health Rep. 1990 Nov-Dec;105(6):617-22.
In 1982-83, 4,485 persons ages 65 or older were identified by a household census in East Boston, MA: 3,812 (85 percent) of them responded to a health and social status questionnaire. Data on age, sex, and living arrangements for the 4,485 eligible people were analyzed with respect to final participation status and reason for refusal or reluctance. The health and social status of reluctant and ready self-respondents were compared, and respondents-by-proxy were compared with self-respondents. Total participation rates were similar for both sexes and all ages, but the likelihood of interview by proxy increased with age, as did the likelihood of nonparticipation due to unavailability. Living alone or with other participants favored participation, and living with refusers or other nonrespondents increased the probability of refusal. While reluctant and ready self-respondents differed in only one health variable and two social variables, respondents-by-proxy differed from self-respondents in most variables tested. These analyses suggest an absence of major differences between self-respondents and refusers. Therefore, nonresponse bias is not likely to have a major impact on interpretation of the data obtained from participants in this study.
1982 - 1983年，马萨诸塞州东波士顿通过家庭普查确定了4485名65岁及以上的老人：其中3812人（85%）回复了一份健康与社会状况调查问卷。对这4485名符合条件者的年龄、性别和居住安排数据，就最终参与状况及拒绝或不愿参与的原因进行了分析。对不愿参与但最终参与的受访者与愿意并实际参与的受访者的健康和社会状况进行了比较，还将他人代答的受访者与自行作答的受访者进行了比较。男女及各年龄段的总参与率相似，但他人代答的访谈可能性随年龄增长而增加，因无法参与而不参与的可能性也如此。独自居住或与其他参与者同住有利于参与，而与拒绝者或其他未回复者同住会增加拒绝的概率。虽然不愿参与但最终参与的受访者与愿意并实际参与的受访者仅在一个健康变量和两个社会变量上存在差异，但他人代答的受访者与自行作答的受访者在大多数测试变量上存在差异。这些分析表明自行作答者与拒绝者之间没有重大差异。因此，无回应偏差不太可能对本研究参与者所获数据的解读产生重大影响。
