Titre :
|
Effects of simplifying choice tasks on estimates of taste heterogeneity in stated-choice surveys. (2010)
|
Auteurs :
|
JOHNSON (Freed) : USA. Research Triangle Institute. Rti-Hs. Research Triangle Park. NC. ;
OZDEMIR (Semra) : USA. Gillings School of Global Public Health. University of North Carolina. Chapel Hill. NC. ;
PHILLIPS (Kathryn-A) : USA. University of California. San Francisco. CA.
|
Type de document :
|
Article
|
Dans :
|
Social science and medicine (vol. 70, n° 2, Janvier 2010)
|
Pagination :
|
183-190
|
Langues:
|
Anglais
|
Mots-clés :
|
Sérologie
;
Estimation
;
Enquête
;
Surveillance
;
Dépistage
;
Sida
;
Diagnostic
;
VIH
;
Homme
;
Amérique
;
Virose
;
Infection
;
Rétrovirus
;
Virus
;
Goût
;
Immunopathologie
;
Amérique du Nord
|
Résumé :
|
[BDSP. Notice produite par INIST-CNRS IpGR0xJA. Diffusion soumise à autorisation]. Researchers usually employ orthogonal arrays or D-optimal designs with little or no attribute overlap in stated-choice surveys. The challenge is to balance statistical efficiency and respondent burden to minimize the overall error in the survey responses. This study examined whether simplifying the choice task, by using a design with more overlap, provides advantages over standard minimum-overlap methods. We administered two designs for eliciting HIV test preferences to split samples. Surveys were undertaken at four HIV testing locations in San Francisco, California. Personal characteristics had different effects on willingness to pay for the two treatments, and gains in statistical efficiency in the minimal-overlap version more than compensated for possible imprecision from increased measurement error.
|