JMIR Publications

ABSTRACT

Background:

Web-based survey data collection has been widely used due to its advantages, although attaining and retaining participants can be challenging. There are several factors associated with successful web-based survey participation; yet little is known regarding socioeconomic (SES) and racial/ethnic differences in web-based survey participation.

Objective:

The purpose of the study was to examine SES and racial/ethnic differences in participation in a web-based survey.

Methods:

We conducted secondary data analysis using data collected in the parent mixed methods study that enrolled racially and ethnically diverse parents of preschool children. We used two phases of web-based surveys deployed in Research Electronic Data Capture (REDCap): 1) screening questions including race/ethnicity information; and 2) full survey with a consent form. Study flyers were posted on local community boards as well as social media, online communities, and online blogs. Once potential participants submitted the screening questions, including their racial/ethnic information, the team sent the full survey link to potential participants who met study eligibility criteria. We calculated the proportion of racial/ethnic groups in each of the following areas: consent, partial survey completion, and total survey completion.

Results:

A total of 487 participants [236 non-Hispanic White; 44 Hispanic; 137 Black; and 70 Asian] completed screening questions and a total of 458 participants met study eligibility. There was a significant difference by race/ethnicity in informed consent for participation in the study. Compared with Black participants, non-Hispanic Whites and Asians were more likely to consent to participate in the study (OR = 1.73, p=0.02, 95% CI = 1.08, 2.78; OR=2.07, p=0.04, 95% CI = 1.04, 4.13, respectively). However, there was no racial/ethnic difference in both completion of demographic questions and partial survey completion rates. Lastly, there was a significant difference among racial/ethnic groups for total survey completion rate. Compared with Black participants, non-Hispanic White participants were more likely to complete both partial and total survey (OR = 3.8, p=0.001, 95% CI =1.73, 8.42). With respect of SES, participants with high school diploma or less were less likely to complete half of the survey as well as total survey (OR = - 2.53, p < 0.01, 95% CI = 0.02, 0.39; OR = - 1.87, p< 0.01, 95% CI = 0.50, 1.48, respectively).

Conclusions:

We found a significant difference among racial/ethnic groups as well as different education levels in web-based survey participation. Survey researchers need to consider the SES and racial/ethnic differences in web-based survey participation and develop strategies to address this bias in participation and completion in their research.