Bulletin of the World Health Organization

Rural practice preferences among medical students in Ghana: a discrete choice experiment

Margaret E Kruk, Jennifer C Johnson, Mawuli Gyakobo, Peter Agyei-Baffour, Kwesi Asabir, S Rani Kotha, Janet Kwansah, Emmanuel Nakua, Rachel C Snow & Mawuli Dzodzomenyo

Volume 88, Number 5, May 2010, 333-341

Table 2. Mixed logit modela results for discrete choice experiment on rural posting attributes influencing students’ stated job preferences

Attribute Parameter Model 1 Model 2
βb 95% CI βb 95% CI
Salaryc Mean 0.12 0.09 to 0.15 0.12 0.08 to 0.15
SD 0.09 0.04 to 0.14 0.10 0.05 to 0.15
Allowance for children’s educationd Mean 0.70 0.54 to 0.86 0.68 0.48 to 0.87
SD 0.48 0.23 to 0.73 0.49 0.25 to 0.74
Improved infrastructuree Mean 1.42 1.17 to 1.66 1.44 1.14 to 1.74
SD 1.19 0.94 to 1.44 1.22 0.96 to 1.48
Supportive managementf Mean 1.17 0.96 to 1.39 1.05 0.79 to 1.30
SD 1.08 0.84 to 1.33 1.10 0.85 to 1.35
2 years before study leaveg Mean 1.04 0.83 to 1.24 1.18 0.92 to 1.44
SD 1.01 0.78 to 1.25 1.04 0.79 to 1.29
No housingh Mean −1.59 −1.88 to −1.31 −1.56 −1.90 to −1.22
SD 1.18 0.87 to 1.50 1.23 0.91 to 1.56
Superior housingh Mean 0.65 0.44 to 0.86 0.83 0.56 to 1.11
SD 0.78 0.46 to 1.10 0.83 0.50 to 1.15
Utility cari Mean 0.78 0.61 to 0.94 0.73 0.53 to 0.94
SD 0.58 0.33 to 0.83 0.62 0.37 to 0.87
Female interaction terms
Salary 0.01 −0.04 to 0.06
Allowance for children’s education 0.09 −0.19 to 0.38
Improved infrastructure 0.02 −0.37 to 0.41
Supportive management 0.40 0.03 to 0.77
2 years before study leave −0.33 −0.68 to 0.03
Basic housing −0.18 −0.66 to 0.30
Superior housing −0.42 −0.84 to −0.01
Utility car 0.17 −0.13 to 0.47
No. of respondents 302 301
No. of observations 6644 6622
Log-likelihood −1541.7595 −1529.8463
Likelihood ratio χ2 209.28 208.02

CI, confidence interval; SD, standard deviation.

a Mixed logit model estimated using Stata’s mixlogit command and 500 Halton draws.

b The coefficient (β) represents the mean relative utility of each attribute conditional on other attributes in a choice set, while the standard deviation of the random coefficients reflects the degree of heterogeneity among respondents in the utility of a given attribute.

c Continuous variable; the coefficient represents the magnitude of increase in utility for every 10% increase in salary.

d Compared with no allowance for children’s education.

e Compared with basic infrastructure, equipment and supplies.

f Compared with unsupportive workplace and management.

g Compared with study leave after at least 5 years.

h Compared with basic housing.

i Compared with no utility car.

[an error occurred while processing this directive]