Skip to main content
. 2017 Oct 19;2:24. doi: 10.1186/s40834-017-0052-7

Table 3.

Estimated coefficients and standard errors for multinomial probit model on contraceptive method choice (EDHS 2011)

Method choice IUD/Implants Injectables Condoms
Fixed Effects Estimate Std. Err. Estimate Std. Err. Estimate Std. Err.
 Constant −1.597*** 0.192 1.799*** 0.150 −2.176*** 0.400
Education
 No education (Ref) 0 0 0 0 0 0
 Primary −0.331*** 0.0777 −0.171*** 0.063 0.163 0.128
 Secondary −0.403*** 0.0844 −0.749*** 0.068 0.434 0.116
Age at start
  < 25 years (Ref) 0 0 0 0 0 0
 25–34 years 0.217* 0.119 −0.014 0.088 −0.460*** 0.141
 35–49 years 0.560*** 0.119 −0.349*** 0.089 −0.891*** 0.147
Wealth
 Low (Ref) 0 0 0 0 0 0
 Medium 0.394*** 0.139 0.149 0.120 NA NA
 High 0.026 0.142 −0.097 0.117 0.897** 0.395
Area
 Urban(Ref) 0 0 0 0 0 0
 Rural 0.805*** 0.084 0.889*** 0.069 −0.447 *** 0.137
Religion
 Orthodox (Ref) 0 0 0 0 0 0
 Muslim −0.006 0.070 −0.183*** 0.058 −1.232*** 0.131
 Others 0.519 0.484 0.481 0.497 NA NA
Correlation between women-level random effects across equations
 IUD/Im. & Injectables −0.097 (0.030) ***
 IUD/Im. & Condoms −0.302 (0.188)
 Condoms & Inject −0.077 (0.052)

The reference category is oral contraceptives. Bold depicts significance: *P < 0.05, **P < 0.01, ***P < 0.001