Internet versus mail: are stated preferences affected by the mode of sampling in a choice experiment?
Publikation: Konferencebidrag › Paper › Forskning
Standard
Internet versus mail : are stated preferences affected by the mode of sampling in a choice experiment? / Olsen, Søren Bøye.
2007. Paper præsenteret ved EAERE 2007 Annual Conference, Thessaloniki, Grækenland.Publikation: Konferencebidrag › Paper › Forskning
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - CONF
T1 - Internet versus mail
AU - Olsen, Søren Bøye
N1 - Conference code: 15
PY - 2007
Y1 - 2007
N2 - In a choice experiment setup, this paper suggests that internet surveys can be a viable alternative to traditional mail surveys when gathering feedback from a sample of respondents. In a study concerning preferences for avoiding encroachment of nature areas in future motorway planning in Denmark, two samples of respondents are surveyed – one by internet and one by mail. Results suggest that additional self-selection mechanisms in the internet survey mode lead to the internet sample being less representative of the survey population than is the case for the mail sample. Similar response rates are obtained in the two samples, but the internet sample is apparently less prone to protest bidding. A thorough parametric analysis cannot reject that stated preferences elicited by use of the two different modes of sampling are identical. This suggests that the fear of a potential survey mode effect is unfounded, and using the internet for stated preference surveys does not introduce any additional major biases that are not already present in the traditional paper-and-pencil surveys.
AB - In a choice experiment setup, this paper suggests that internet surveys can be a viable alternative to traditional mail surveys when gathering feedback from a sample of respondents. In a study concerning preferences for avoiding encroachment of nature areas in future motorway planning in Denmark, two samples of respondents are surveyed – one by internet and one by mail. Results suggest that additional self-selection mechanisms in the internet survey mode lead to the internet sample being less representative of the survey population than is the case for the mail sample. Similar response rates are obtained in the two samples, but the internet sample is apparently less prone to protest bidding. A thorough parametric analysis cannot reject that stated preferences elicited by use of the two different modes of sampling are identical. This suggests that the fear of a potential survey mode effect is unfounded, and using the internet for stated preference surveys does not introduce any additional major biases that are not already present in the traditional paper-and-pencil surveys.
KW - Former LIFE faculty
KW - survey mode effect
KW - representativity
KW - choice experiment
KW - willingness-to-pay
KW - random parameters
KW - mixed logit
M3 - Paper
Y2 - 27 June 2007 through 30 June 2007
ER -
ID: 8072244