Fig. 1 | Scientific Reports

Fig. 1

From: Using stated preference methods to design gender-affirming long-acting PrEP programs for transgender and nonbinary adults

Fig. 1

Mean preference weights with 95% confidence intervals (CIs). Bars represents mean preference weights and error bars present 95% confidence intervals. Estimates are from conditional logistic regression models that included the primary (first) program choice vs all other program choices as the dependent variable and dummy-coded variables for each attribute level as the independent variables. Dummy coding estimates the preference weights of a given attribute level relative to the omitted level of that attribute. A more positive weight indicates a stronger preference towards a given attribute level.

Back to article page