Evaluation of Different Recruitment Methods: Longitudinal, Web-Based, Pan-European Physical Activity Through Sustainable Transport Approaches (PASTA) Project

Abstract

Background: Sufficient sample size and minimal sample bias are core requirements for empirical data analyses. Combining opportunistic recruitment with a Web-based survey and data-collection platform yields new benefits over traditional recruitment approaches.

Objective: This paper aims to report the success of different recruitment methods and obtain data on participants' characteristics, participation behavior, recruitment rates, and representativeness of the sample.

Methods: A longitudinal, Web-based survey was implemented as part of the European PASTA (Physical Activity through Sustainable Transport Approaches) project, between November 2014 and December 2016. During this period, participants were recruited from 7 European cities on a rolling basis. A standardized guide on recruitment strategy was developed for all cities, to reach a sufficient number of adult participants. To make use of the strengths and minimize weakness, a combination of different opportunistic recruitment methods was applied. In addition, the random sampling approach was applied in the city of Örebro. To reduce the attrition rate and improve real-time monitoring, the Web-based platform featured a participant's and a researchers' user interface and dashboard.

Results: Overall, 10,691 participants were recruited; most people found out about the survey through their workplace or employer (2300/10691, 21.51%), outreach promotion (2219/10691, 20.76%), and social media (1859/10691, 17.39%). The average number of questionnaires filled in per participant varied significantly between the cities (P<.001), with the highest number in Zurich (11.0, SE 0.33) and the lowest in Örebro (4.8, SE 0.17). Collaboration with local organizations, the use of Facebook and mailing lists, and direct street recruitment were the most effective approaches in reaching a high share of participants (P<.001). Considering the invested working hours, Facebook was one of the most time-efficient methods. Compared with the cities' census data, the composition of study participants was broadly representative in terms of gender distribution; however, the study included younger and better-educated participants.

Conclusions: We observed that offering a mixed recruitment approach was highly effective in achieving a high participation rate. The highest attrition rate and the lowest average number of questionnaires filled in per participant were observed in Örebro, which also recruited participants through random sampling. These findings suggest that people who are more interested in the topic are more willing to participate and stay in a survey than those who are selected randomly and may not have a strong connection to the research topic. Although direct face-to-face contacts were very effective with respect to the number of recruited participants, recruiting people through social media was not only effective but also very time efficient. The collected data are based on one of the largest recruited longitudinal samples with a common recruitment strategy in different European cities.

Keywords: Web-based questionnaire; Web-based survey; longitudinal survey; mobile phone; multicentral; opportunistic sampling; recruitment.

Figure 1

Cities’ own Twitter and Facebook accounts to recruit and inform people. Left: A Twitter page from London (Source: PASTA consortium, 2016, Project profile [Twitter]). Right: A Facebook page from Antwerp (Source: PASTA consortium, 2016, Project profile [Facebook]).

Figure 2

Recruitment material. Left: T-shirts worn by recruiters during outreach activities in Barcelona. Right: Registration postcard distributed in the city of Vienna.

Figure 3

Number of participants who registered for the survey and who started and finalized the baseline questionnaire. (a) Attrition rate between the number of registrations and the number of started baseline questionnaires. (b) Attrition rate between the number of started and finalized baseline questionnaires.

Figure 4

Number of recruited (observed) and predicted participants in Antwerp (top graph), Barcelona (middle graph), and Vienna (bottom graph; based on a nonlinear model).