Using a dual-frame design to improve phone surveys on political attitudes: developing a weighting strategy for limited external information in Hong Kong

Abstract

In recent years, rapid increases in mobile phone ownership and decreases in landline users have led to potential biases in landline phone survey estimations. Mobile-only users have been found to be over-represented in many mobile phone surveys. A dual-frame survey, namely a combination of a landline and mobile phone survey, is proposed to solve this problem. The design of such a survey requires a more complex weighting procedure and thus additional benchmark information on phone status and usages for weighting. There is no consensus on a standard weighting method, but there is a general agreement that it should include: (1) a computation of base weight and (2) a post-stratification adjustment. Various weighting methods for a dual-frame phone survey of political attitudes using empirical data from a study on political attitudes in Hong Kong were investigated in this study. We found that the average estimator and the single-frame estimator methods are the best approaches for computing the base weight for a dual-frame survey, and that they provide similar estimations on political attitudes. No significant difference in estimation on political attitudes was found between using only gender and age for the post-stratification adjustment and including gender, age, education, and working status. Cell weighting and raking provided similar estimations.