Conjoint Measurement: The Effect of
Visual Representation on Choice Sets,
Accuracy, and Precision
Thesis Defense
Introduction
› The field of preference measurement remains very active, important, and growing
› Level of craft à Visualization of attributes (Eggers, Hauser & Selove, 2016)
› Many research about verbal description and pictorial
representation (or actual products) (Holbrook & Moore, 1981; Smead et al., 1981; Domzal & Unger, 1985; Anderson, 1987; Louviere et al., 1987; Vriens et al., 1998)
Research Questions
1. "To what extent could static visual images increase the number of choice sets being answered (i.e., willingness-to-respond)?" and;
Methodology
› Choice-Based Conjoint analysis (CBC) with a Fractional Factorial Design
› Multinomial Logit Model › Between-Subjects Design › Attributes & Levels
Empirical Study (1)
› Sample:
§ Excluding vegetarian
§ Visual: 112 respondents (49%) vs. Text: 115 respondents (51%) § Males: 96 (42%) vs. Females: 131 (58%)
§ Age: mean 26.07 years (min. = 18; max. = 61)
› CBC:
§ Price Variable
Empirical Study (2)
› T-test: significant difference in number of choice sets between the two representations
› Visual: 13.01 choice sets vs. Text: 9.77 choice sets
› T-test: significant difference in processing choice sets
Empirical Study (3)
› Accuracy of Image Representation:
§ Respondents who were shown images
were more likely to value differences in cheese, lettuce, and tomatoes
› Precision of Image Representation:
§ The total range for image
representation (5.61125) exceeded the total range for the text
Empirical Study (4)
› Accuracy of More Choice Sets:
§ Respondents who answered more choice sets were relatively more likely to value
differences in the type of meat, tomatoes, onions, type of sauce (only in image representation)
› Precision of More Choice Sets:
§ The total range increased from 1-5 choice sets to 11-15 choice sets
§ In the second holdout set had less absolute difference, which indicate that in the
Limitations & Future Research
› Conjoint Analysis
§ Results different real market experience.
› Meat-eaters vs. Meat-reducers vs. Vegetarians › Generalizability