Fig 1.
Note: The French version of the choice tasks, as seen by the respondents, is in SM B. The version shown here is a U.S.-English translation for illustration, which is why the currency symbol appears as “$” rather than “€”. The upper part of the picture reminds the respondent of the choice frame. The middle part represents the first stage choice between two competing nursing homes. The bottom part represents the second stage choice between the selected nursing home or home.
Table 1.
Attributes and levels.
Table 2.
Sample descriptive statistics.
Fig 2.
Results of random-intercept logit model (N = 2651 individuals). The detailed coefficients are showed in SM G. In the dummy coding model (panel A), the reference categories for each attributes were poor rating (equipment-atmosphere and professional care attribute), nursing home not in closest town, and €500 out-of-pocket (OOP) cost.
Table 3.
Latent class estimation.
Table 4.
Description of respondents’ profiles, by latent classes.