Skip to main content

Table 5 measures of components of variance and heterogeneity on unscheduled contraceptive discontinuation, multilevel logistic regression analysis EDHS 2016, January 2018

From: Reasons and multilevel factors associated with unscheduled contraceptive use discontinuation in Ethiopia: evidence from Ethiopian demographic and health survey 2016

RANDOM EFFECTS

MODEL-0

MODEL-1

MODEL-2

MODEL-3

Measure of general Contextual effect

 ICC

11.12%

8.08%

4.52%

3.94%

 MOR

1.844

1.671

1.457

1.419

 Var

0.4114

0.2893

0.1555

0.1349

 PCV

 

29.68%

62.20%

67.21%

 AUC % (95%CI)

75.35 (0.73835, 0.76859)

72.91 (0.71338, 0.74448)

69.89 (0.68255, 0.71530)

70.29 (0.68659, 0.71917)

MODEL Selection

 AIC

5227.386

5189.581

5118.615

5106.208

 AUC% (Fixed effect only)

Reference

59.94

63.13

64.83

Log-likelihood (LR-test p-value)

− 2611.6929 (0.0000)

− 2567.79 (0.0000)

− 2541.3073 (0.0000)

− 2510.104 (0.0002)

  1. ICC intra-class correlation, MOR median odds ratio, Var random intercept variance, PCV proportional change in variance, AUC area under the curve, AIC akaki’s information criteria, CI confidence interval