Survey response bias and the ‘privacy paradox’: evidence from a discrete choice experiment

ABSTRACT

The discrepancy between the stated and revealed value of personal information is frequently referred to as the ‘privacy paradox.’ We test for evidence that survey response bias contributes to the privacy paradox through a discrete choice survey experiment based on hypothetical ride-hailing services, some of which track the rider’s location and share it with third parties. The value that respondents place on their personal location data is measured by the willingness to pay higher prices to avoid rides with location sharing. Our survey experiment compares a within-subjects design in which location sharing was presented as an attribute of the hypothetical rides within each choice scenario to a between-subjects design in which location sharing was not presented as an attribute – instead, the location-sharing practices of the hypothetical firms were presented to a treatment group before the respondents made their choices. On average, survey respondents placed a positive value on their personal location data under both survey designs. The difference in value between the survey designs was statistically insignificant, indicating there is no evidence that the within-subjects design led to the type of survey response bias that might contribute to the privacy paradox.

KEYWORDS:

• Privacy
• survey experiment
• mixed logit
• willingness to pay

JEL CLASSIFICATION:

• C25
• D12
• C90

Acknowledgement

We thank Elizabeth Allen for designing the logos of the hypothetical ride-hailing services and Kenneth Train for many helpful comments.

Disclosure statement

No potential conflict of interest was reported by the authors.