Improving on Probability Weighting for Household Size
In: Public opinion quarterly: journal of the American Association for Public Opinion Research, Volume 62, Issue 3, p. 398-404
ISSN: 0033-362X
Examines the effects of poststratification & inverse-probability weights, both used as corrective measures, for household size in telephone polling. Focus is on how these tools rectify distortions in sample composition produced by nonresponse rates & a lower probability of individuals in large households being included in the surveys using random samples of households. Comparison of responses in 1988 national telephone polls & the National Election Study (N = 2,404 adult respondents to in-person interviews) vs 1990 US census data revealed that probability weighting corrected distortions but overrepresented large households in the telephone polls. It is concluded that poststratification is a more effective tool to reduce the predictable biases due to nonavailability/nonresponse & produce less variable final weights that fit the target population more accurately. 3 Tables, 2 References. T. Arnold