The use of incentives in vulnerable populations for a telephone survey: a randomized controlled trial
1 Division of Clinical Epidemiology, Department of Medicine, McGill University Health Centre, Montreal, QC, Canada
2 Division of Pediatric Allergy and Clinical Immunology, Department of Pediatrics, McGill University Health Center, Montreal, QC, Canada
3 Department of Geography, University of Toronto, Toronto, ON, Canada
4 Departments of Epidemiology and Biostatistics, McGill University, Montreal, QC, Canada
5 Food Directorate, Health Canada, Ottawa, ON, Canada
6 Faculty of Applied Health Sciences, University of Waterloo, Waterloo, ON, Canada
7 Division of Allergy and Clinical Immunology, Department of Medicine, McGill University Health Center, Montreal, QC, Canada
BMC Research Notes 2012, 5:572 doi:10.1186/1756-0500-5-572Published: 19 October 2012
Poor response rates in prevalence surveys can lead to nonresponse bias thereby compromising the validity of prevalence estimates. We conducted a telephone survey of randomly selected households to estimate the prevalence of food allergy in the 10 Canadian provinces between May 2008 and March 2009 (the SCAAALAR study: Surveying Canadians to Assess the Prevalence of Common Food Allergies and Attitudes towards Food LAbeling and Risk). A household response rate of only 34.6% was attained, and those of lower socioeconomic status, lower education and new Canadians were underrepresented. We are now attempting to target these vulnerable populations in the SPAACE study (Surveying the Prevalence of Food Allergy in All Canadian Environments) and are evaluating strategies to increase the response rate. Although the success of incentives to increase response rates has been demonstrated previously, no studies have specifically examined the use of unconditional incentives in these vulnerable populations in a telephone survey. The pilot study will compare response rates between vulnerable Canadian populations receiving and not receiving an incentive.
Randomly selected households were randomly assigned to receive either a $5 incentive or no incentive. The between group differences in response rates and 95% confidence intervals (CIs) were calculated. The response rates for the incentive and non-incentive groups were 36.1% and 28.7% respectively, yielding a between group difference of 7.4% (−0.7%, 15.6%).
Although the wide CI precludes definitive conclusions, our results suggest that unconditional incentives are effective in vulnerable populations for telephone surveys.