Rachel Tumin1, Kurt Johnson2, Doug Spence3, Reena Oza-Frank4. 1. Survey Research Division, RTI International, 3040 E. Cornwallis Road, Durham, NC, 27709, USA. rtumin@rti.org. 2. Division of Research Services, RTI International, 3040 E. Cornwallis Road, Durham, NC, 27709, USA. 3. Ohio Colleges of Medicine, Government Resource Center, 150 Pressey Hall, 1070 Carmack Road, Columbus, OH, 43210, USA. 4. Bureau of Maternal, Child and Family Health, Ohio Department of Health, 246 N. High Street, Columbus, OH, 43215, USA.
Abstract
INTRODUCTION: With shifting demographics and declining response rates, state and national health surveys are considering mixed mode approaches. Most states field the Pregnancy Risk Assessment Monitoring System (PRAMS) surveillance project, but few have studied the effect of encouraging online responses. METHODS: Like PRAMS, the 2016 Ohio Pregnancy Assessment Survey interviewed new mothers 2-4 months after delivery (n = 3382). Fielding included a traditional mailed paper questionnaire with telephone follow-up protocol and two experiments: a rotating web invitation added a web survey link at different points during the mail protocol, and a push-to-web protocol asked women to complete the survey online before mailing a questionnaire. This analysis examined the responses rates and tested for unweighted demographic differences using Pearson's chi-square. RESULTS: The unweighted response rate was highest with the traditional contact protocol (30.0%) and slightly lower among the rotating web invite (27.4%) and the push-to-web (25.5%) groups. Nearly two-thirds (64%) of push-to-web protocol respondents completed the survey online, with 70% of those web surveys submitted before the first paper questionnaire was mailed. Women who responded to the web versus mail surveys were similar on most characteristics, although in both experimental groups, women who completed the web version were more likely to be college educated. Among the push-to-web group, 60% of web and 36% of mail respondents had a 4-year college degree (p < .001). DISCUSSION: Given the potential for push-to-web to shift respondents to an online survey without greatly impacting response rates, researchers should continue to examine the utility of incorporating a web mode in surveys of new mothers.
INTRODUCTION: With shifting demographics and declining response rates, state and national health surveys are considering mixed mode approaches. Most states field the Pregnancy Risk Assessment Monitoring System (PRAMS) surveillance project, but few have studied the effect of encouraging online responses. METHODS: Like PRAMS, the 2016 Ohio Pregnancy Assessment Survey interviewed new mothers 2-4 months after delivery (n = 3382). Fielding included a traditional mailed paper questionnaire with telephone follow-up protocol and two experiments: a rotating web invitation added a web survey link at different points during the mail protocol, and a push-to-web protocol asked women to complete the survey online before mailing a questionnaire. This analysis examined the responses rates and tested for unweighted demographic differences using Pearson's chi-square. RESULTS: The unweighted response rate was highest with the traditional contact protocol (30.0%) and slightly lower among the rotating web invite (27.4%) and the push-to-web (25.5%) groups. Nearly two-thirds (64%) of push-to-web protocol respondents completed the survey online, with 70% of those web surveys submitted before the first paper questionnaire was mailed. Women who responded to the web versus mail surveys were similar on most characteristics, although in both experimental groups, women who completed the web version were more likely to be college educated. Among the push-to-web group, 60% of web and 36% of mail respondents had a 4-year college degree (p < .001). DISCUSSION: Given the potential for push-to-web to shift respondents to an online survey without greatly impacting response rates, researchers should continue to examine the utility of incorporating a web mode in surveys of new mothers.