In this study, a total of 4720 ever married women had been included in the analysis. Some variables were found with missing values and treatment of the missing values was done using missing completely at random technique. About 1 in 4 ever-married women in the age range of 25-34 have experienced intimate partner violence during their life time. More than a quarter of ever-married women living in rural part of Ethiopia have experienced an intimate partner violence during their life time. One thousand and fourteen (24.6%) of women who were married or living with partner have encountered violence from their intimate partner. Among ever-married women who have been participated in this study, 22.3 %, 7.6% and 6.1% of women have experienced emotional, sexual and physical violence respectively. The prevalence of intimate partner violence was found to be 25.6%(Table -1).
Table-1: Percentage of ever-married women age 15-49 who have experienced intimate partner violence during their life time, Ethiopia DHS 2016
Variables
|
Category
|
Intimate Partner Violence
|
Yes (%)
|
No (%)
|
Age group
|
15-24
|
243(21.3)
|
897(78.7)
|
25-34
|
506(25.1)
|
1506(74.9)
|
35-49
|
458(29.2)
|
1110(70.8)
|
Religion
|
Christian
|
750(28.2)
|
1914(71.8)
|
Muslim
|
431(21.6)
|
1562(78.4)
|
Other
|
26(41.3)
|
37(58.7)
|
Residence
|
Urban
|
288(23.8)
|
923(76.2)
|
Rural
|
919(26.2)
|
2590(73.8)
|
Region
|
Tigray
|
145(29.4)
|
348(70.6)
|
Afar
|
60(15.5)
|
327(84.5)
|
Amhara
|
159(27.8)
|
413(72.2)
|
Oromia
|
198(30.5)
|
451(69.5)
|
Somali
|
40(8.6)
|
424(91.4)
|
Benishangul Gumuz
|
125(32.6)
|
258(67.4)
|
SNNPR
|
137(24.3)
|
427(75.7)
|
Gambela
|
105(30.6)
|
238(69.4)
|
Harari
|
101(35.9)
|
180(64.1)
|
Addis Ababa
|
65(22.6)
|
223(77.4)
|
Dire Dawa
|
72(24.3)
|
224(75.7)
|
Marital Status
|
Married or living with partner
|
1014(24.6)
|
3109(75.4)
|
Widowed or Divorced or Separated
|
193(32.3)
|
404(67.7)
|
Occupation
|
Not working
|
541(22.8)
|
1837(77.2)
|
Employed (salary paid)
|
26(15.8)
|
139(84.2)
|
Merchant
|
198(28.1)
|
506(71.9)
|
Farmer
|
295(30.9)
|
661(69.1)
|
Other
|
147(28.4)
|
370(71.6)
|
Educational level of partner
|
No education
|
484(24.8)
|
1469(75.2)
|
Primary
|
363(28.2)
|
924(71.8)
|
Secondary and above
|
167(18.9)
|
716(81.1)
|
Wealth quintile
|
Poorest
|
340(24.1)
|
1072(75.9)
|
Poorer
|
203(27.8)
|
526(72.2)
|
Middle
|
198(30)
|
462(70)
|
Richer
|
172(28.2)
|
2438(71.8)
|
Richest
|
294(22.5)
|
1015(77.5)
|
Number of living children
|
0
|
101(19.2)
|
424(80.8)
|
1-2
|
420(25.6)
|
1221(74.4)
|
3-4
|
320(25.3)
|
943(74.7)
|
5+
|
366(28.4)
|
925(71.6)
|
Number of other wives
|
No other wives
|
843(23.7)
|
2715(76.3)
|
1-6 other wives
|
171(30.3)
|
394(69.7)
|
Partner drinks alcohol
|
No
|
723(20.6)
|
2779(79.4)
|
Yes
|
484(39.7)
|
734(60.3)
|
Woman smokes cigarettes
|
No
|
1181(25.3)
|
3484(74.7)
|
Yes
|
26(47.3)
|
29(52.7)
|
Woman chewed chat
|
No
|
1025(24.7)
|
3126(75.3)
|
Yes
|
182(32)
|
387(68)
|
Partner’s desire for children
|
Both want same
|
377(23)
|
1264(77)
|
Partner wants more
|
295(26.6)
|
815(73.4)
|
Partner wants fewer
|
101(36.2)
|
178(63.8)
|
Do not know
|
234(21.7)
|
846(78.3)
|
A multiple logistic model was fitted to the data to test the research hypothesis regarding the relationship between the likelihood that intimate partner violence is related with the predictor variables. A result displayed in Table-2 showed that intimate partner violence was significantly associated with geographical regions. The odds of intimate partner violence in Amhara, Oromia, SNNPR, Benishangul Gumuz, Gambela, Addis Ababa and Dire Dawa regions were not significantly different from intimate partner violence in Tigray region. Ever married women in Afar and Somali region have been 0.653 and 0.3084 times less likely to have intimate partner violence than in Tigray region respectively. Also, the finding in Harari region were 2.1759 times more likely to experience intimate partner violence than Tigray region.
Furthermore, findings show the difference in the area of residence related to intimate partner violence against women. It has been found that ever married women living in the rural areas were 1.262 times more likely to have intimate partner violence as compared to those living in the urban areas. Likewise, the result illustrated that ever married women employed (salaried) were 0.5899 times less likely to experience intimate partner violence compared to those not working (Table -2).
Additionally, the result shows that women who were smoking cigarettes were at increased risk of intimate partner violence. Ever married women with smoking cigarettes were 1.913 times more likely to have intimate partner violence compared to those who did not smoke cigarettes. In line with this, it has been found that ever married women who were not living together with their husbands/partners were 1.419 times more likely to have intimate partner violence compared to married or living with partner. In addition, women with their husbands having 1-6 other wives were1.612times more likely to experience violence than women with their husbands have no other wives. Ever married women with partner drinks alcohol were 2.692 times more likely to have intimate partner violence than their counterparts (Table -2).
Besides this, findings show that the log odds of the intimate partner violence for chewed chat. With this, the odds ratio (OR) of chewed chat as compared to that of the not chewed chat is given as OR =1.33, which implies that women who chewed chat were 33% more likely to have intimate partner violence compared to women who were not chewing chat. Women with number of living children 1-2, 5+were 1.366 and 1.692 times more likely to experience intimate partner violence compared to women who had no living children respectively. Husband's desire for children wants fewer were 1.496 times more likely to have intimate partner violence compared to husband's desire for children both husband and woman want same (Table-2).
Table -2: Multiple binary logistic regression models of factors associated with intimate partner violence among ever-married women in Ethiopia, Ethiopia DHS 2016
Variables
|
Estimate
|
S.E.
|
OR
|
Confidence Interval (eβ)
|
Z-value
|
p-value
|
Lower
|
Upper
|
Intercept
|
-2.05926
|
0.19061
|
0.127548
|
0.0874237
|
0.1846011
|
-10.804
|
< 2e-16 ***
|
Region (ref= Tigray)
|
Afar
|
-0.42618
|
0.1844
|
0.653001
|
0.4527895
|
0.9339428
|
-2.310
|
0.02088 *
|
Amhara
|
-0.20944
|
0.1428
|
0.811036
|
0.6130171
|
1.0730595
|
-1.467
|
0.14233
|
Oromia
|
0.23484
|
0.1419
|
1.264704
|
0.9583952
|
1.6717506
|
1.655
|
0.09786
|
Somali
|
-1.17633
|
0.2062
|
0.308409
|
0.2039295
|
0.4582929
|
-5.706
|
1.16e-08 ***
|
Benishangul Gumuz
|
0.25129
|
0.1582
|
1.285686
|
0.9427332
|
1.7529533
|
1.589
|
0.11206
|
SNNPR
|
-0.10631
|
0.14964
|
0.8991435
|
0.6704136
|
1.2056927
|
-0.710
|
0.47743
|
Gambela
|
0.15735
|
0.16289
|
1.1704018
|
0.8497303
|
1.6098032
|
0.966
|
0.33407
|
Harari
|
0.77744
|
0.17770
|
2.1758984
|
1.5352874
|
3.0824104
|
4.375
|
1.21e-05 ***
|
Addis Ababa
|
-0.12741
|
0.19557
|
0.8803744
|
0.5983241
|
1.2888041
|
-0.651
|
0.51475
|
Dire Dawa
|
0.06499
|
0.18598
|
1.0671515
|
0.7391351
|
1.5332466
|
0.349
|
0.72675
|
Type of place of residence(ref=Urban)
|
Rural
|
0.23243
|
0.10701
|
1.2616596
|
1.0242832
|
1.5583291
|
2.172
|
0.02986 *
|
Smokes cigarettes(ref=No)
|
Yes
|
0.64891
|
0.29215
|
1.91345
|
1.0740949
|
3.3937026
|
2.221
|
0.02634 *
|
Current marital status(ref= Married or Living with partner)
|
Widowed /divorced/separated
|
0.34988
|
0.10357
|
1.4188947
|
1.1567564
|
1.7363488
|
3.378
|
0.00073 ***
|
Number of other wives(ref = No other wives)
|
1-6 other wives
|
0.47752
|
0.10353
|
1.6120743
|
1.314742
|
1.9732178
|
4.612
|
3.98e-06 ***
|
Respondent's occupation (grouped) (ref = Not working)
|
Employed(salary paid)
|
-0.52772
|
0.23465
|
0.5899483
|
0.3657148
|
0.9204773
|
-2.249
|
0.02451 *
|
Merchant
|
0.12697
|
0.10724
|
1.1353823
|
0.9189643
|
1.3993882
|
1.184
|
0.23643
|
Farmer
|
0.06229
|
0.09673
|
1.0642744
|
0.8798846
|
1.2857112
|
0.644
|
0.51959
|
Others
|
0.07904
|
0.11872
|
1.0822445
|
0.855926
|
1.3634809
|
0.666
|
0.50557
|
Husband/partner drinks alcohol (ref= No)
|
Yes
|
0.99024
|
0.08445
|
2.6918727
|
2.2820390
|
3.1778415
|
11.725
|
< 2e-16 ***
|
Have you ever chewed Chat? (ref= No)
|
Yes
|
0.28518
|
0.11017
|
1.3299958
|
1.0701995
|
1.6485698
|
2.588
|
0.00964 **
|
Number of living children(ref= 0)
|
1-2
|
0.31196
|
0.12998
|
1.3661039
|
1.0622291
|
1.7687714
|
2.400
|
0.01639 *
|
3-4
|
0.25908
|
0.13802
|
1.2957318
|
0.9946261
|
1.6977970
|
1.901
|
0.05734.
|
5+
|
0.52568
|
0.13802
|
1.6916041
|
1.2943589
|
2.2242557
|
3.809
|
0.00014 ***
|
Husband's desire for children(ref= Both want same)
|
Husband wants more
|
0.16318
|
0.08877
|
1.1772497
|
0.9889775
|
1.4007256
|
1.838
|
0.06604.
|
Husband wants fewer
|
0.40272
|
0.13488
|
1.4958932
|
1.1460513
|
1.9453109
|
2.986
|
0.00283 **
|
Don't know
|
-0.15859
|
0.0902
|
0.853345
|
0.7146105
|
1.0178497
|
-1.758
|
0.07874
|
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
In Table-3, a chi-square test statistic was applied to assess heterogeneity in the proportion of ever married women who have intimate partner violence within regions. The test yield χ2 = 152.8934, d. f=10, P-value < 0.001. Therefore, regions are indeed significantly heterogeneous. Thus, the significance of this test gives us an evidence of heterogeneity of intimate partner violence across regions in Ethiopia. The variance of the random factor is significant which indicates that there is regional variation in intimate partner violence. The significance of the deviance-based chi-square value for the empty model shown implies that an empty model with random intercept is better than an empty model without random intercept. The significant deviance-based chi-square value and smallest AIC for random intercept model indicates that the random intercept and fixed slope model is a better fit as compared to the empty model. Similarly, the deviance-based chi-square test for significance of the random intercept model with the fixed coefficient and random coefficient model indicated that the random intercept model with the fixed coefficient were a better fit compared to the multiple logistic regression model. Therefore, random intercept model is better fitted this data.
Table-3: Multilevel logistic regression model comparison
|
Null model
|
Random intercept model
|
Random coefficient model
|
-2*loglikelihood
|
5255.201
|
5001.462
|
4999.45
|
Deviance based chi-square test
|
152.8934
|
118.3081
|
125.011
|
P-value
|
3.435549e-134***
|
2.148824e-97***
|
2.270592e-104***
|
Model Fit Diagnostics
|
AIC
|
5259.201
|
5037.462
|
5045.45
|
BIC
|
5272.12
|
5153.734
|
5194.02
|
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
In table-4, the intercept β0= -1.1160 interpreted as the odds of intimate partner violence in an average region. That is the intercept informs us that the average probability of intimate partner violence everywhere in Ethiopia is 0. 246754. The intra-region correlation obtained from the empty model with random effect was 0.0215. This value is called the intra class correlation coefficient (ICC) and interpreted as 2.15% of the variation in the intimate partner violence can be explained by grouping every married woman with respect to their region. The remaining 97.85% of the variation of intimate partner violence is explained within region-lower level units.
Table-4. Results for multilevel logistic regression model without explanatory variables
Fixed part
|
Coefficient
|
S.E.
|
Z-value
|
Pr(>|z|)
|
β0 –intercept
|
-1.1160
|
0.1429
|
-7.811
|
5.68e-15 ***
|
|
Random part
|
Estimate
|
S. E.
|
Wald test of 95% CI
|
Between –Region variance 𝛿20=𝑣𝑎𝑟(𝑈𝑜𝑗)
|
0.0724113
|
0.002433671
|
0.0459518
|
0.1141063
|
Deviance based chi-square 152.8934 p-value =3.435549e-134***
|
In Table-5 for It indicated that intimate partner violence variations among regional states of Ethiopia were non-zero. The random effects indicates that between region variance, the intimate partner violence is estimated as 𝛿20= 0.066 and the variance between each respondent within the region is estimated as =3.29. Then, the total variance equals to 3.356. This shows that 1.97% of the variance in the intimate partner violence can be attributed to difference between regions. The random intercept and fixed slope logistic regression model is a multilevel model which have random intercept and fixed coefficient of predictors. As can be seen from Table- 5, the analysis of multilevel logistic regression revealed that intimate partner violence in ever married women varied among regions. The result displayed that the region-wise difference in intimate partner violence was statistically significant. In addition, husband's desire for children, number of living children, women ever chewed chat, husband/partner drinks alcohol, respondent's occupation, number of other wives, current marital status, women smoke cigarettes, type of place of residence were also found to be significant determinants of variation in intimate partner violence among the regions. The estimated coefficients and odds ratio had similar interpretation with that of single level logistic regression analysis.
Table-5: Results of random intercept and fixed coefficient logistic regression model
Fixed effect
|
Β
|
S.E.
|
Z-value
|
P-value
|
Intercept
|
-2.10713
|
0.20029
|
-10.520
|
< 2e-16 ***
|
Type of place of residence(ref=Urban)
|
Rural
|
0.22372
|
0.10532
|
2.124
|
0.033649 *
|
Smokes cigarettes(ref=No)
|
Yes
|
0.65533
|
0.29175
|
2.246
|
0.024691 *
|
Current marital status (ref= Married or Living with partner)
|
Widowed /divorced/separated
|
0.34872
|
0.10335
|
3.374
|
0.000741 ***
|
Number of other wives (ref = No other wives)
|
1-6 other wives
|
0.46438
|
0.10301
|
4.508
|
6.54e-06 ***
|
Respondent's occupation (grouped) (ref = Not working)
|
Employed (salary paid)
|
-0.52444
|
0.23444
|
-2.237
|
0.025283 *
|
Merchant
|
0.12826
|
0.10696
|
1.199
|
0.230450
|
Farmer
|
0.07280
|
0.09637
|
0.755
|
0.450006
|
Others
|
0.08456
|
0.11860
|
0.713
|
0.475857
|
Husband/partner drinks alcohol (ref= No)
|
Yes
|
0.99561
|
0.08389
|
11.868
|
< 2e-16 ***
|
Have you ever chewed Chat? (ref= No)
|
Yes
|
0.31048
|
0.10993
|
2.824
|
0.004740 **
|
Number of living children (ref= 0)
|
1-2
|
0.31490
|
0.12981
|
2.426
|
0.015276 *
|
3-4
|
0.26051
|
0.13611
|
1.914
|
0.055621.
|
5+
|
0.52123
|
0.13775
|
3.784
|
0.000154 ***
|
Husband's desire for children (ref= Both want same)
|
Husband wants more
|
0.15934
|
0.08857
|
1.799
|
0.072007.
|
Husband wants fewer
|
0.40306
|
0.13474
|
2.991
|
0.002778 **
|
Don't know
|
-0.15387
|
0.09009
|
-1.708
|
0.087640.
|
Random part
|
Estimate
|
S.E.
|
Wald test of 95% CI
|
Random intercept
𝛿20=𝑣𝑎𝑟(𝑈𝑜𝑗)
|
0.06614167
|
0.002433671
|
0.04140978
|
0.1056446
|
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1