Decomposing gender inequalities in self-assessed health status in Liberia

Figures & data

Table 1. Overall and rural sample characteristics.

	Overall sample			Rural sample
Variable	Men	Women	Differences	Men	Women	Differences
SAH	42.3	37.0	−5.0***	39.6	29.9	−10.0***
Urban	34.5	46.5	12.0***	-	-	-
Age	46.5	45.7	−0.70	47.2	48.7	1.5**
Married	66.2	20.4	−46.0***	72.3	24.6	−48.0***
Household size	4.64	3.84	−0.80***	4.6	3.9	−0.7***
Education	66.6	34.5	−32.0***	59.6	21.1	−38***
Literate	61.9	31.6	−30.0***	53.9	18.2	−36***
Expenditure	53,968.2	48,416.0	−5552.20***	46,665.1	38,988.3	−7676.9***
Roof	64.9	72.4	8.0***	51.1	57.1	6.0***
Floor	41.6	47.1	6.0***	22.7	24.6	2.0
Toilet	26.5	27.5	1.0	17.6	16.4	−1.0
Lighting fuel	61.8	61.5	−0.3	54.5	52.5	−2.0
Cooking fuel	29.0	36.5	7.0***	8.9	10.3	1.0
Clean water rainy	61.9	71.8	10.0***	49.5	58.2	9.0***
Clean water dry	59.0	69.1	10.0***	45.7	55.2	9.0***
Radio	61.2	29.2	−32.0***	56.7	22.9	−34.0***
Mosquito net	51.5	48.0	−3.0**	54.5	51.5	−3.0
Illness	25.8	37.4	12.0***	25.4	36.4	11.0***
N	3017	1087		1976	582

Note: 1 USD = 133 Liberian dollar as of May 2018. *, **, ***. Statistically significant at the 10%, 5%, and 1% level, respectively based on two-tailed t-test. Mean and percentages are shown for continuous and dummy variables respectively.

Table 2. Logit decomposition of men-women gap in health status.

	Estimate	Standard error	Percent
Raw difference	0.054***	0.016	100
Explained: due to difference in characteristics	0.030***	0.012	54.4
Unexplained: due to difference in coefficients	0.025	0.020	45.6
Due to difference in characteristics
Urban	−0.001	0.003	−1.7
Age	−0.005***	0.000	−8.5
Married	0.003	0.009	5.6
Education	−0.009	0.013	−16.0
Literate	0.009	0.012	17.0
Expenditure	0.003***	0.001	5.1
Roof	−0.001	0.001	−1.1
Floor	−0.000	0.001	−0.2
Toilet	−0.001***	0.000	−1.3
Lighting fuel	−0.000	0.000	−0.1
Cooking fuel	−0.000	0.001	−0.1
Clean water rainy	−0.002	0.004	−3.2
Clean water dry	−0.001	0.004	−2.5
Radio	0.024***	0.006	44.0
Mosquito net	0.002***	0.001	4.0
Illness	0.008***	0.002	14.0
Due to difference in coefficients
Urban	−0.013	0.016	−23.2
Age	0.066	0.043	121.2
Married	0.001	0.006	0.6
Education	0.005	0.024	8.8
Literate	−0.010	0.023	−19.2
Expenditure	−0.046*	0.024	−85.0
Roof	0.048	0.001	88.2
Floor	−0.033	0.001	−60.5
Toilet	0.000**	0.000	−0.4
Lighting fuel	0.013	0.000	24.3
Cooking fuel	0.014	0.001	25.4
Clean water rainy	−0.010	0.050	−12.5
Clean water dry	0.002	0.042	4.2
Radio	0.002	0.008	3.7
Mosquito net	0.005	0.013	−8.5
Illness	0.003	0.010	5.3
Constant	−0.015	0.058	−27.7

Note: A detailed non-linear (logit) decomposition. The ‘Percent’ column gives the contribution of each variable to the overall difference in health status between men and women. This is computed by dividing the estimate by the overall difference (0.054). Total observations is 4104. *, **, ***. Statistically significant at the 10%, 5%, and 1% level, respectively.

Table 3. Logit decomposition of men and women in health status in rural areas.

	Estimate	Standard error	Percent
Raw difference	0.099***	0.021	100
Explained: due to difference in characteristics	0.044***	0.01	44.9
Unexplained: due to difference in coefficients	0.055	0.026	55.1
Due to difference in characteristics
Age	0.010***	0.001	9.8
Married	−0.001	0.011	−1.2
Education	−0.014	0.018	−14.3
Literate	0.020	0.016	20.0
Expenditure	−0.000	0.003	−0.2
Roof	0.001	0.001	−1.0
Floor	0.001	0.001	0.6
Toilet	0.001**	0.000	0.7
Lighting fuel	0.000	0.000	0.3
Cooking fuel	0.000	0.001	0.4
Clean water rainy	0.000	0.004	0.5
Clean water dry	−0.003	0.004	−2.8
Radio	0.028***	0.001	28.5
Mosquito net	0.001***	0.001	1.5
Illness	0.002	0.003	2.2
Due to difference in coefficients
Age	0.099	0.062	99.7
Married	0.005	0.011	5.4
Education	0.019	0.024	19.0
Literate	−0.024	0.021	−24.5
Expenditure	−0.014	0.033	−14.2
Roof	0.050**	0.023	50.4
Floor	−0.033***	0.012	−32.8
Toilet	−0.000	0.010	−0.6
Lighting fuel	−0.010	0.020	−8.6
Cooking fuel	0.10	0.010	6.9
Clean water rainy	0.023	0.050	23.0
Clean water dry	−0.021	0.047	−20.7
Radio	0.010	0.010	9.8
Mosquito net	0.010	0.020	8.1
Illness	−0.010	0.014	−7.4
Constant	−0.059	0.081	−58.6

Note: A detailed non-linear (logit) decomposition. The ‘Percent’ column gives the contribution of each variable to the overall difference in health status between men and women. This is computed by dividing the estimate by the overall difference (0.098). Total observations is 2558. *, **, ***. Statistically significant at the 10%, 5%, and 1% level, respectively.

Table

	Estimate	Standard error	Percent
Raw difference	0.054***	0.016	100
Explained: due to difference in characteristics	0.025***	0.012	46.4
Unexplained: due to difference in coefficients	0.029	0.020	53.6
Due to difference in characteristics
Urban	−0.001	0.003	−1.5
Age	−0.004***	0.000	−8.5
Married	0.001	0.009	1.7
Education	−0.009	0.013	−16.0
Literate	0.010	0.012	18.4
Roof	−0.000	0.002	−0.9
Floor	−0.000	0.000	−0.5
Toilet	−0.001***	0.002	−1.4
Lighting fuel	0.000	0.000	0.1
Cooking fuel	−0.001	0.002	−1.8
Clean water rainy	−0.001	0.004	−2.9
Clean water dry	−0.002	0.004	−2.6
Radio	0.024***	0.006	44.9
Mosquito net	0.002***	0.006	3.9
Illness	0.007***	0.002	13.5
Due to difference in coefficients
Urban	−0.014	0.016	−25.8
Age	0.060	0.045	111.3
Married	−0.000	0.006	−0.6
Education	0.007	0.026	12.8
Literate	−0.017	0.024	−31.1
Roof	0.046	0.026	85.0
Floor	−0.036	0.017	−65.8
Toilet	0.000	0.008	0.1
Lighting fuel	0.012	0.017	21.3
Cooking fuel	0.008	0.014	15.1
Clean water rainy	−0.004	0.047	−8.1
Clean water dry	0.001	0.044	−2.0
Radio	−0.001	0.008	−1.0
Mosquito net	−0.003	0.013	−6.1
Illness	0.003	0.011	5.1
Constant	−0.033	0.059	−60.7

Note: A detailed non-linear (logit) decomposition. The ‘Percent’ column gives the contribution of each variable to the overall difference in health status between men and women. This is computed by dividing the estimate by the overall difference (0.054). Total observations is 4104. *, **, ***. Statistically significant at the 10%, 5%, and 1% level, respectively.