Abstract
Background: Most applications of choice-based conjoint analysis in health use choice tasks with onlytwo profiles, while those in marketing routinely use three or more. This study reports on a randomizedtrial comparing paired with triplet profile choice formats focused on measuring patient preferencefor hearing aids.
Methods: Respondents with hearing loss were drawn from a nationally representative cohort, completedidentical surveys incorporating a conjoint analysis, but were randomized to choice tasks with two orthree profiles. Baseline differences between the two groups were explored using ANOVA and chi-squaretests. The primary outcomes of differences in estimated preferences were explored using t-tests, likelihoodratio tests, and analysis of individual-level models estimated with ordinary least squares.
Results: 500 respondents were recruited. 127 had no hearing loss, 28 had profound loss and 22 declinedto participate and were not analyzed. Of the remaining 323 participants, 146 individuals were randomizedto the pairs and 177 to triplets. The only significant difference between the groups was time to completethe survey (11.5 and 21 minutes respectively). Pairs and triplets produced identical rankings of attributeimportance but homogeneity was rejected (P<0.0001). Pairs led to more variation, and were systematicallybiased toward the null because a third (32.2%) of respondents focused on only one attribute. This isin contrast to respondents in the triplet design who traded across all attributes.
Discussion: The number of profiles in choice tasks affects the results of conjoint analysis studies. Heretriplets are preferred to pairs as they avoid non-trading and allow for more accurate estimation of preferencesmodels.
Methods: Respondents with hearing loss were drawn from a nationally representative cohort, completedidentical surveys incorporating a conjoint analysis, but were randomized to choice tasks with two orthree profiles. Baseline differences between the two groups were explored using ANOVA and chi-squaretests. The primary outcomes of differences in estimated preferences were explored using t-tests, likelihoodratio tests, and analysis of individual-level models estimated with ordinary least squares.
Results: 500 respondents were recruited. 127 had no hearing loss, 28 had profound loss and 22 declinedto participate and were not analyzed. Of the remaining 323 participants, 146 individuals were randomizedto the pairs and 177 to triplets. The only significant difference between the groups was time to completethe survey (11.5 and 21 minutes respectively). Pairs and triplets produced identical rankings of attributeimportance but homogeneity was rejected (P<0.0001). Pairs led to more variation, and were systematicallybiased toward the null because a third (32.2%) of respondents focused on only one attribute. This isin contrast to respondents in the triplet design who traded across all attributes.
Discussion: The number of profiles in choice tasks affects the results of conjoint analysis studies. Heretriplets are preferred to pairs as they avoid non-trading and allow for more accurate estimation of preferencesmodels.
| Original language | English |
|---|---|
| Publisher | National Bureau of Economic Research |
| Number of pages | 33 |
| DOIs | |
| Publication status | Published - 2011 |
Publication series
| Name | NBER Working Paper |
|---|---|
| Publisher | National Bureau of Economic Research |
| No. | 17482 |