Vaccine hesitancy is low and messaging solves.

Our experiment revealed that all message framings (i.e. expert, dynamic social norms, safety, efficacy and pro-social messages) were equally effective and that unvaccinated participants overwhelmingly intended to get vaccinated. Regarding vaccination intentions, researchers have commonly used behavioural intentions to predict behaviours in experiments (33), and one meta-analysis examining 47 experiments found that a medium-to-large change in behavioural intention (d = 0.66) led to a small-to-medium change in behaviours (d = 0.36) (33).

Our experiment leveraged an online survey and employed cutting-edge risk communication techniques (25,26,34-38) to examine which messages could increase vaccination intention relatively early for vaccination roll-out in the Kingdom of Saudi Arabia. Most of the previous research in Saudi Arabia focused on identifying factors predicting vaccination intentions using cross-sectional surveys (18-21). In contrast, previous experimental research conducted in other regions had mixed findings (3437). For example, vaccine safety messages encouraged vaccination intentions (34) in Latin America (39), the United Kingdom (40), and USA (41), but not among unvaccinated adults in California (42) and adults in Poland (43). Social norms messages increased vaccination intentions among healthcare workers (44) and adults (41) in the USA but not in Poland (43) and the United Kingdom (45). Sharing expert opinions about the COVID-19 vaccine increased vaccine uptakes in the Czech Republic (32). Our framing experiment provides causal evidence to support communication campaign designs and adds evidence from non-Western societies (34-38,46).

The already high baseline intention rate and relatively small sample sizes of the unvaccinated in the framing arms may be the reason for not detecting statistically significant differences. However, descriptively, dynamic social norms and safety messages increased the vaccination intentions of unvaccinated participants, while messages using national and international experts and efficacy, and pro-social messages, reduced the vaccination intentions. Given previous research (34-41,44) and with a caution that only unvaccinated adults participated in our experiment, and we only found descriptive differences, public health authorities could attempt to use social norms and safety messages in their immunization communication efforts. Future research can test messages with physicians or doctors to guide risk communication by public health authorities, given previous research findings in the Czech Republic (32) and our insignificant findings about expert messages. Also, there may be other impactful message types that can be applied in the Middle East and North Africa, such as value-based messages or messages using different sources, narratives, visuals, and emotional appeals, which future research can examine (26).