Telephone Survey Versus Panel Survey Samples Assessing Knowledge, Attitudes and Behavior Regarding Animal Welfare in the Red Meat Industry in Australia.

Lauren M Hemsworth,Paul H Hemsworth,Grahame J Coleman,Maxine Rice

doi:10.3389/fpsyg.2021.581928

Abstract

Surveys are used extensively in social research and, despite a lack of conclusive evidence of their ‘representativeness,’ probability internet panel (PIP) surveys are being increasingly used to make inferences about knowledge, attitude and behavior in the general population regarding a range of socially relevant issues. A large-scale survey of Australian public attitudes and behavior toward the red meat industry was undertaken. Samples were obtained using a random digit dialing telephone survey (Computer-Assisted Telephone Interviewing-CATI, n = 502 respondents) and a PIP survey (PANEL, n = 530 respondents) to examine differences between the two samples regarding attitudes and behavior relating to livestock use and welfare. There was little difference in demographics between the CATI and the PANEL surveys apart from highest level of education. However, there were differences between the two samples in both attitudes and behavior toward the red meat industry after controlling for education levels. The PANEL respondents gave generally more conservative responses than did the CATI respondents in the sense that they were more positive toward the livestock industries and animal welfare within these industries. Differences were also found between the respondents of the two samples regarding behavior that relates to the red meat industry, both community and consumer behavior. PANEL respondents were less engaged in community behaviors performed in opposition of the red meat industry when compared with the CATI sample. The majority of CATI and PANEL respondents were red meat eaters and there was no difference between respondents of the two samples in relation to red meat consumption, however, there were fewer vegetarians and vegans in the PANEL survey. Possible reasons for the observed differences are discussed, however, a definitive answer will depend on further research to identify the specific psychological factors that differ between samples derived from different survey methodologies.

Highlights

A questionnaire was developed using an iterative process that began with questionnaires that had been developed by the Animal Welfare Science Centre (AWSC) for livestock industries including the pork, egg and red meat industries
A comparison of the computer-assisted telephone interviewing (CATI) sample with the PANEL sample on the composite variables showed that the PANEL sample gave generally more conservative responses than did the CATI sample, in the sense that they were more positive toward the red meat industry and animal welfare within these industries
We found differences between the two survey samples in both attitudes and behavior toward the red meat industry

Summary

Introduction

High marginal costs and low response rates have reduced the viability of random telephone surveys (RDD; random digit dialing) whilst increasing the viability of surveys delivered online (internet surveys) (Berrens et al, 2003; Li et al, 2004; Ansolabehere and Schaffner, 2014). Whilst telephone surveys reportedly generated higher participant response rates than online or mail delivery (Yu and Cooper, 1983) and data quality that was comparable to that obtained from face-to-face interviews (Groves and Kahn, 1979), it has become increasingly difficult to maintain participant response rates and, as a result, the costs of telephone data collection has risen considerably (Lavrakas, 1997; Holbrook et al, 2007; Chang and Krosnick, 2009). Ansolabehere and Schaffner (2014) compared internet, mail and telephone surveys and found response rates of 42.9, 21.1, and 19.5%, respectively, with completion times of 8.9, 11.8 and 14.3 min, respectively. The CATI survey achieved the highest response rate, and the highest cost and the longest completion time

Methods

Results

Discussion

Conclusion