Abstract:
Confirmation bias is the tendency of information searchers to select and evaluate information that supports pre-existing attitudes favourably. The current dissertation investigates whether confirmation bias affects health-related search in online environments, where users share content and social tag clouds are the navigation interface for searchers. I assumed that when individuals search health-related issues, they are motivated to find accurate information (accuracy motivation), in contrast to defending their self-concept (defense motivation). To determine what information is accurate, I expect that searchers attend to internal, individual evaluations (prior knowledge, prior attitudes, and attitude confidence), and external, collective cues (tag popularity and source credibility).
Regarding the influence of individual evaluations, in studies 2 and 3, a linear influence of prior attitudes on the selection of blog posts (but not tags), and the evaluation of blog posts was found. In studies 2 and 3, I tested whether the influence of prior attitudes was moderated by confidence. I found that high confidence did affect the selection of blog posts but not tags in both studies, and confidence influenced the evaluation of tag-related blog posts.
Regarding the influence of the collective cues, tag popularity was manipulated in studies 1 and 3, where I found a main effect of tag popularity on the selection of tags, blog posts, and evaluation of content, showing that tag size influenced confirmation bias in a moderate to strong way. In the student sample (study 2), I found that high credibility reduced the influence of prior attitudes on the selection of tags and consequently blog posts. However, using a representative sample (study 3), no influence of source credibility was found. With respect to the searchers’ evaluation of content, credibility had no influence in study 2, but in study 3, under high source credibility and low attitude confidence, searchers evaluated content more favourably when content was attitude consistent.
In conclusion, the present dissertation shows that confirmation bias and individual evaluations guide information searchers in tag-based navigation, extending the literature which showed behaviour in social tagging environments follows semantic associations. The results are interesting for the construction of content aggregation or social tagging platforms, and practitioners who provide health-related online content. Practitioners and platform providers pay attention to their target audience, as this will either elicit accuracy or defense motivation. So, different strategies can be implemented when the aim is to reduce the influence of confirmation bias on information search behaviour.