Testing the effectiveness of behavioural messages to increase recruitment to health research when embedded within social media campaigns on Twitter: A Web-based Experiment (Preprint)

BACKGROUND Social media is rapidly becoming the primary source to disseminate invitations to the public to consider taking part in research studies. There is, however, little information on how the contents of the advertisement can be communicated to facilitate engagement and subsequently promote intentions to participate in research. This study designed and tested the effectiveness of behavioural messages on participants’ intention to take part in a web-based study using simulated Twitter posts. OBJECTIVE Social media is rapidly becoming the primary source to disseminate invitations to the public to consider taking part in research studies. There is, however, little information on how the contents of the advertisement can be communicated to facilitate engagement and subsequently promote intentions to participate in research. This study designed and tested the effectiveness of behavioural messages on participants’ intention to take part in a web-based study using simulated Twitter posts. METHODS 1060 women were included in an online experiment and randomised to one of three experimental conditions; standard advert (N=360), patient endorsement advert (N=345), and social norms advert (N=355). After seeing one of three adverts, participants were asked to state: 1) their intention to take part in the advertised study, 2) the ease of understanding the message and study aims, and 3) their willingness to be redirected to the study website after completing the survey. Individuals were further asked to suggest ways to improve the messages. Intentions were compared between groups using ordinal logistic regression, reported in percentages (%), adjusted odd ratio (aOR) and 95% confidence intervals (95% CI). RESULTS Those who were in the patient endorsement and social norms-based advertisement groups had significantly lower intentions to take part in the advertised study, compared with those in the standard advert group (aOR: 0.73, 95% CI 0.55-0.97, p=0.029 and aOR: 0.69, 95% CI 0.52-0.92, p=0.009, respectively). The patient endorsement advert was perceived to be more difficult to understand (aOR: 0.65, 95% CI 0.48-0.87, p=0.004) and to less clearly communicate the study aims (aOR: 0.72, 95% CI 0.55-0.95, p=0.010). While the patient endorsement advert had no impact on intention to visit the main study, the social norms adverts decreased willingness, compared with the standard advert group (44.2% vs. 53.1%, aOR: 0.74, 95% CI 0.54-0.99, p=0.018). The majority of participants (395/609, 64.8%) stated that the messages did not require changes, but some preferred clearer (75/609, 12.3%) and shorter (59/609, 9.7%) messages. CONCLUSIONS The results of this study indicate that adding normative behavioural messages to simulated Tweets decreased participant intention to take part in our web-based case-control study, as this makes the Tweet harder to understand. This suggests that simple messages should be used for participant recruitment through Twitter.