Skip to main content

Table 3 Estimated coefficients and standard errors for multinomial probit model on contraceptive method choice (EDHS 2011)

From: Side effects and the need for secrecy: characterising discontinuation of modern contraception and its causes in Ethiopia using mixed methods

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)

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