Abstract
An often-mentioned advantage over traditional survey experiments is
that conjoint experiments have the potential to mitigate social desirability
bias. Because respondents are presented with numerous features, the
research objective is concealed and respondents can find multiple justifications
for any given choice. However, if and to what extent this is
true may depend on design choices – a concern that has received little
empirical attention. In this paper, I randomly assign respondents to six
conjoint designs in order to induce respondents’ awareness to sensitive features
and their possibilities of acting strategically. The results show that
while design variations has a substantial effect on respondents’ awareness
to sensitive features, it has no detectable effect on their priorities. The
evidence suggests that the fear of social desirability bias should not force
researchers to use conjoint designs that are otherwise suboptimal.
that conjoint experiments have the potential to mitigate social desirability
bias. Because respondents are presented with numerous features, the
research objective is concealed and respondents can find multiple justifications
for any given choice. However, if and to what extent this is
true may depend on design choices – a concern that has received little
empirical attention. In this paper, I randomly assign respondents to six
conjoint designs in order to induce respondents’ awareness to sensitive features
and their possibilities of acting strategically. The results show that
while design variations has a substantial effect on respondents’ awareness
to sensitive features, it has no detectable effect on their priorities. The
evidence suggests that the fear of social desirability bias should not force
researchers to use conjoint designs that are otherwise suboptimal.
| Originalsprog | Engelsk |
|---|---|
| Udgiver | Department of Political Sceince, University of Copenhagen |
| Antal sider | 34 |
| Status | Udgivet - sep. 2018 |
Emneord
- Det Samfundsvidenskabelige Fakultet