Highlighting Consensus Among Medical Scientists Increases Public Support for Vaccines

Evidence From a Randomized Experiment

Sander L. van der Linden; Chris E. Clarke; Edward W. Maibach


BMC Public Health. 2015;15(1207) 

In This Article


We conducted a between-subject experiment in June of 2015. Participants (N = 206) were a diverse sample of American adults (56 % male, 18–75+, 45 % Democrat, see Appendix (Table 2) for a full description of the sample) recruited from Amazon Mechanical Turk (Mturk) – a platform which has shown to be more diverse and at least as reliable as other internet-based samples.[21,22] Parental information was not recorded, as we focused on promoting science-based vaccine attitudes among adults broadly. Respondents were offered a small reward ($0.40) to complete an online survey in which they were randomly assigned to one of four experimental conditions; a descriptive norm condition (n = 59), a prescriptive norm condition (n = 60), a combination of the two (n = 44) or a control group (n = 43). Drawing on expert-survey estimates,[3,4] participants were shown a pie-chart which either stated that; "90 % of medical scientists agree that vaccines are safe" (descriptive), "90 % of medical scientists agree that all parents should be required to vaccinate their children" (prescriptive) or a combination of the two (see Additional file 1: Figure S1 http://static-content.springer.com/esm/art%3A10.1186%2Fs12889-015-2541-4/MediaObjects/12889_2015_2541_MOESM1_ESM.tif). Participants in the control group received no statement. Approval from Princeton's Institutional Review Board (#7310) was obtained prior to the study. Participants also signed a written consent form.

After exposure to the treatment(s), all respondents answered the main survey questions. Perceived consensus was assessed with the following item; "to the best of your knowledge, what % of medical scientists agree that vaccines are safe?" (0 to 100 %). Perceived risk was assessed with the following item; "how concerned are you about the potential risks of vaccines? (1 = not concerned at all, 7 = very concerned). Endorsement of the autism link was assessed by asking people to what extent they agreed with the following statement; "there is scientific evidence for a causal link between vaccines and autism" (1 = strongly disagree - 7 = strongly agree). Public support for vaccines was assessed with 8 items, which were combined and averaged into a single measure to form a reliable index (cronbach's α = 0.96), example items include; "I believe that vaccines are a safe and reliable way to avoid the spread of preventable diseases", "I have already vaccinated my children or would do so if I had children" and "I would support policies that require people to vaccinate their children" (1 = strongly disagree - 7 = strongly agree). A full description of all measures used in the study is provided in the Appendix (Table 3). Results of the experiment were assessed through mean-comparisons (main effects) and mediation analysis (adjusted estimates) using STATA (StataCorp) v.13.