Confirmation Bias Towards Treatments of Depressive Disorders in Social Tagging

How Confidence in Prior Attitudes, Social Tag Popularity, and Source Credibility Shape Confirmation Bias Toward Antidepressants and Psychotherapy in a Representative German Sample: Randomized Controlled Web-Based Study

The study examines whether people primarily want to confirm their prior attitudes in health-related information search, in an online environment using social tags for navigation. Participants were looking for information on the treatment of depression with antidepressants and psychotherapy. They were randomly assigned to two groups with either high or low credibility of the community who provides social tags, and two groups where participants' confidence in prior attitudes was heightened or lowered, and to two groups where either antidepressant tags were more popular or psychotherapy was more popular. The investigators measured attitude change toward the treatments and also navigation behavior.

In health-related, Web-based information searches, people should select information in line with expert (vs nonexpert) information, independent of their prior attitudes and consequent confirmation bias. This study aimed to investigate confirmation bias in mental health-related information searches, particularly (1) if high confidence worsens confirmation bias, (2) if social tags eliminate the influence of prior attitudes, and (3) if people successfully distinguish high and low source credibility. In total, 520 participants of a representative sample of the German Web-based population were recruited via a panel company. Among them, 48.1% (250/520) participants completed the fully automated study. Participants provided prior attitudes about antidepressants and psychotherapy. The investigators manipulated (1) confidence in prior attitudes when participants searched for blog posts about the treatment of depression, (2) tag popularity -either psychotherapy or antidepressant tags were more popular, and (3) source credibility with banners indicating high or low expertise of the tagging community. The investigators measured tag and blog post selection, and treatment efficacy ratings after navigation.

Confidence in prior attitudes (high vs. low: recalling situations in which participants were confident or uncertain about their thoughts)

Credibility of the source (tagging community: experts - many years of professional experience vs. novices - students in the first semester)

The relative size of treatment tags in a tag cloud was either larger for antidepressant treatments or psychotherapy treatments.

Participants thought back of situations in which they were either confident or doubtful about their own knowledge. This should elicit a mindset where participants are more or less confident about their own prior attitudes.

The source credibility of the community that allegedly collected and labelled the blog posts was either high or low in terms of expertise. Either experts (high credibility) or first semester students (low credibility) did allegedly collect blog posts. This was indicated by banners on top of the navigation platform in the internet browser.

The investigators constructed a questionnaire to measure the attitudinal preference of psychotherapy over antidepressant treatments of depressive disorders. On a 7-point likert scale, participants rate the degree of efficacy of antidepressant and psychotherapy treatments, on 8 items (e.g. item 1: "Antidepressants/Psychotherapy are/is effective in treating depression."). An index score for the degree of preference of psychotherapy is calculated by subtracting the average antidepressants score from the average psychotherapy treatment rating score for each participant. To analyse if attitudinal preferences predict the number of clicks on social tags and blog posts, the treatment preference score is entered in a logistic regression as predictor. Ratings are inquired at the beginning of the 1 hour study (prior attitudes), and at the end of the study (attitude change).

Through study completion, an average of 1 hour. Prior to and after information search phase in the study.

Both, psychotherapy and antidepressant tags can be clicked on, and are counted respectively. An index score will be calculated for each participant subtracting the sum of clicks on antidepressants from the sum of clicks on psychotherapy, to analyse if clicks are associated with the treatment preference measured by prior treatment attitudes.

Both, psychotherapy and antidepressant blog posts can be clicked on, and are counted respectively. An index score will be calculated for each participant subtracting the sum of clicks on antidepressants from the sum of clicks on psychotherapy, to analyse if clicks are associated with the treatment preference measured by prior treatment attitudes.

Inclusion Criteria: Online Population - Internet Browser, Representative Sample of Germans with respect to age and region Exclusion Criteria: No Internet Browser

Individual level data will be anonymized and appended to the publication at the Journal of Medical Internet research.

The completely anonymized data will be available for researchers and the public who access the published paper at the Journal of Medical Internet research.

Schweiger S, Cress U. How Confidence in Prior Attitudes, Social Tag Popularity, and Source Credibility Shape Confirmation Bias Toward Antidepressants and Psychotherapy in a Representative German Sample: Randomized Controlled Web-Based Study. J Med Internet Res. 2019 Apr 23;21(4):e11081. doi: 10.2196/11081.