Evaluating a large-scale online behaviour change intervention aimed at wildlife product consumers in Singapore.

Interventions to shift the behaviour of consumers using unsustainable wildlife products are key to threatened species conservation. Whether these interventions are effective is largely unknown due to a dearth of detailed evaluations. We previously conducted a country-level online behaviour change intervention targeting consumers of the Critically Endangered saiga antelope (Saiga tatarica) horn in Singapore. To evaluate intervention impact, we carried out in-person consumer surveys with >2,000 individuals pre- and post-intervention (2017 and 2019), and 93 in-person post-intervention surveys with traditional Chinese medicine (TCM) shopkeepers (2019). The proportion of self-reported high-usage saiga horn consumers in the target audience (Chinese Singaporean women aged 35-59) did not change significantly from pre- to post-intervention (24.4% versus 22.6%). However, post-intervention the target audience was significantly more likely than the non-target audience to accurately recall the intervention message and to report a decrease in saiga horn usage (4% versus 1% reported a behaviour change). Within the target audience, high-usage consumers were significantly more likely than lower-usage consumers to recall the message and report a behaviour change. Across respondents who reported a decrease in saiga horn usage, they cited the intervention message as a specific reason for their behaviour change significantly more than other reasons. Additionally, across all respondents, the belief that saiga is a common species in the wild decreased significantly from pre- to post-intervention. TCM shopkeepers, however, cited factors such as price and availability as the strongest influences on saiga horn sales. In sum, the intervention did significantly influence some consumers but the reduction of high-usage consumer frequency was not significant at the population level. We explore reasons for these findings, including competing consumer influences, characteristics of the intervention, and evaluation timing. This work suggests our intervention approach has potential, and exemplifies a multi-pronged in-person evaluation of an online wildlife trade consumer intervention