Confidence in selected institutions in 1975: An attempt at replication across two national surveys

A comparison of responses to four attitudinal items included in two national surveys reveals significant differences between surveys in the marginal distributions of all four items. These discrepancies are attributed to differences in survey ‘context’ (i.e., the nature of the questions which preceded the items of interest in the interview schedules). Despite the differences in marginal distributions, however, the four items showed identical relationships in both surveys to five background characteristics of respondents in 18 of 20 tests. Although the generalizability of these findings to other attitudinal items is an open question, they do suggest the inadvisability of combining data collected in surveys using markedly different interview schedules into a time series of subjective social indicators. Contextual effects are apparently less critical in cross-sectional analyses of survey data.