The Impact Of Social Marketing On A Web-Based Behavioral Risk Factor Surveillance Survey

Loading...
Thumbnail Image

Files

umi-umd-1433.pdf (570.55 KB)
No. of downloads: 3868

Publication or External Link

Date

2004-05-18

Citation

DRUM DOI

Abstract

The potential exists for Web-based health survey methods to collect, analyze, and disseminate increasing amounts of health risk factor and health behavior data. There is a need to establish Web-based survey methodologies that produce results equivalent to current health survey modes. This study examined the impact a social marketing campaign had on a Web-based Behavioral Risk Factor Surveillance Survey. It had three dependent variables, response rate, response time, and item completion rate. An explanatory model of response was also developed using logistic regression analysis. Both qualitative and quantitative data collection methods were used in this study. Qualitative data were used to develop the social marketing promotional framework. Quantitatively, an experimental design was used, with a random sample of 847 participants randomly assigned to control and experimental groups. A Web-based version of the 2001 Behavioral Risk Factor Surveillance System questionnaire was adapted to create a 70 item Web-based Behavioral Risk Factor Surveillance Survey (WBRFSS). The control group had the WBRFSS delivered without any intervention. The experimental group was exposed to a social marketing campaign prior to delivery of the WBRFSS. Of the 770 who were mailed participation requests, 70 completed the survey representing an overall response rate of 9.1 percent. Response rates were 5.96 percent for the control group and 12.24 percent for the experimental group. Chi-Square analysis showed that the response rate for the experimental group was significantly higher than that of the control group. The mean response times for the control group was 96.09 hours (SD=49.41) and for the experimental group was 124.53 hours (SD=112.75).The item response rates were 99.50 percent for the control group and 99.82 percent for the experimental group. The t-test for independent means found no significant difference between mean response times or item response rates. The regression model included the dependent variable, response, and the independent variables, exposure to social marketing promotions, age, sex, ethnicity, county of residence, education, perceived Internet literacy, and availability of an Internet connection at home. The overall model was significant (p<.05). Exposure to the social marketing campaign promotions increased WBRFSS response by more than two-and-one-half times.

Notes

Rights