Table 2

Results from hierarchical Bayes model for utility of attributes of health facilities for delivery, reported for 1006 women from Jimma Zone, Oromia, Ethiopia

Variable	Coefficient^‡‡	SE	95% CI
Travel time^†	0.171^*	0.013	0.145	0.198
Cost^‡	−0.291^*	0.018	−0.326	−0.255
Provider^§
Doctor	2.104^*	0.058	1.989	2.218
Nurse	1.654^*	0.049	1.556	1.752
Good provider attitude/performance	1.445^*	0.049	1.349	1.542
Available drugs and medical equipment	3.877^*	0.074	3.732	4.022
Available transportation	0.449^*	0.027	0.396	0.502
n (observations)	8045
Root likelihood^¶	0.777
Per cent certainty^**	77.1%
p(χ²)^††	<0.001

↵* p<0.01.
↵† In hours.
↵‡ In 10 Birr.
↵§ In reference to health extension worker.
↵¶ With three alternatives per choice, a chance model would have a root likelihood of 1/3. A perfect model fit would have a root likelihood of 1.
↵** Indicates how much better the solution is than chance, as compared to a perfect solution. 0% corresponds to a model fit at the chance level, and 100% corresponds to a perfect model fit. This corresponds to a log likelihood of −2029.87.
↵†† The probability that the fully specified model is not different from a null model (in which all attribute coefficients are zero).
↵‡‡ Mean of individual coefficients.