Skip to main content

Table 3 Multinomial logistic regression model for variables that explain participants' preference

From: Participants' preference for type of leaflet used to feed back the results of a randomised trial: a survey

Participant preference Coefficient, B Standard Error P-Value Exp(B) 95% CI for Exp(B)
No preference Vs Four-page      
Gender 0.656 0.518 0.205 1.927 0.698 to 5.317
Age 0.013 0.027 0.620 1.013 0.962 to 1.068
Education -0.111 0.086 0.199 0.895 0.756 to 1.060
Employed 0.090 0.700 0.897 1.094 0.277 to 4.320
Treatment 0.080 0.460 0.862 1.083 0.440 to 2.666
Constant 0.013 2.101 0.995 1.013 0.016 to 62.231
One-page Vs Four-page      
Gender -0.679 0.492 0.168 0.507 0.193 to 1.330
Age 0.027 0.027 0.317 1.027 0.974 to 1.084
Education 0.015 0.069 0.828 1.015 0.887 to 1.162
Employed -0.686 0.866 0.428 0.504 0.092 to 2.748
Treatment -0.396 0.490 0.419 0.673 0.257 to 1.759
Constant -1.833 1.853 0.323 0.160 0.004 to 6.040
One-page vs No preference      
Gender -1.335 0.616 0.030 0.263 0.079 to 0.880
Age 0.014 0.033 0.674 1.014 0.950 to 1.083
Education 0.126 0.098 0.199 1.134 0.516 to 1.373
Employed -0.776 0.977 0.427 0.460 0.068 to 3.126
Treatment -0.476 0.579 0.411 0.621 0.200 to 1.933
Constant -1.846 2.476 0.456 0.158 0.001 to 20.230
  1. Log likelihood = -115.5; number of observations = 121; likelihood ratio χ2-test statistic, 9.9 df = 10 (P = 0.445); pseudo R2 = 0.041