2.1.3. Dependent variableAnecdotal information use was measured with the question “Haveyou read someone else's commentary or experience about health ormedical issues online,” with respondents choosing from the options“Yes”, “No”, “Don't know” and “Refused.” We included only the datawith the responses Yes and No. Thus, our two categories of dependentvariable include respondents who consumed anecdotal information andthose who did not.2.1.4. AnalysisA binomial logistic regression analysis was performed using SPSS21.0 to predict membership of the responder in one of two categories:anecdotal information user or non-user. The Pew data included finalweights to adjust for effects of the sample design and to compensate forpatterns of nonresponse that might bias results. This investigation in-cluded those in the analysis. We then developed a logistic regressionmodels to assess the characteristics associated with anecdotal in-formation use. For the model, fit indices, pseudo R-squared, effect sizeestimates, logistic regression coefficients and their significance, corre-sponding odds ratios and their confidence intervals were calculated