Purpose: The results of mixed logit modeling (MXL) of discrete choice experimentation include parameter estimates, standard deviations of the parameter estimates, and their associated standard errors, when heterogeneity exists. The purpose of this analysis was to provide a graphical presentation of willingness to pay (WTP) estimates, and the associated heterogeneity, using results from a mixed logit model (MXL) in a discrete choice experiment (DCE).
Methods:
Results from a DCE determining societal preferences for the human papillomavirus (HPV) vaccines were used. The DCE was comprised of 10 choice sets with 7 attributes. The specific attributes and levels were selected based on current vaccination policy and vaccine characteristics. Preferences for each attribute were determined using a MXL. Apart from cost, each of the levels of the attributes was specified to have normally distributed coefficients. The cost coefficient was specified to be fixed, such that the distribution of WTP for each non-cost attribute (which is the ratio of the attribute's coefficient to the cost coefficient) is normally distributed. Density plots of the distributions of WTP for each non-cost attribute were then plotted. Results: The figure presents density plots for WTP for 4 attribute levels: ‘never requiring a booster', 'requiring a pap smear every 5 years', ‘vaccinating girls only' and ‘vaccinating neither boys nor girls'. The heterogeneity in patients' preferences was largest for target group to vaccinate as shown by the relatively flat bell curves for the WTP for both vaccination policies. Conversely, the heterogeneity in patients' preferences was the least for requiring a pap smear every 5 years. Patients were WTP more for a vaccination that could be used by girls, compared to the other vaccination characteristics shown. Nearly 70% of patients were WTP for a vaccination that vaccinated girls only and nearly 60% were WTP for a vaccination which never needed a booster. Almost three quarters of patients wanted to be compensated a monetary amount if a policy were adopted in which neither girls nor boys were vaccinated.
Conclusions: By plotting the WTP densities we are able to convey easily the mean WTP, the spread and thereby the heterogeneity in patients' preferences and information of the share of the patients that places a positive or negative value on the attribute. 
Candidate for the Lee B. Lusted Student Prize Competition