Status of women’s empowerment and its associated factors using multidimensional empowerment index in Tigray, Ethiopia: a community-based cross-sectional study

Abstract

Objective

To assess the status of women’s empowerment and its associated factors using multidimensional empowerment index in Tigray, Ethiopia.

Design

Community-based cross-sectional study.

Setting

Tigray regional state, Ethiopia.

Participants

A sample of married pregnant women (n=1477) whose gestational age was at least 8 weeks were the study participants.

Main outcome measures

The primary outcome of the study was women’s empowerment status, assessed using 24 indicators across five key domains: decision-making power, social independence, attitudes towards violence against women, social networking and household asset ownership. Equal weight was assigned to all domains and the weight assigned to each domain was distributed equally to the indicators within the corresponding domain. Women who scored at least 80% (ie, met 4 out of 5 domains) were considered as empowered.

Results

In total, only 8.2% (95% CI 6.9 to 9.78) of women were empowered. Intrafamilial factors: husband’s education with primary (adjusted OR (AOR): 2.66 (1.30 to 5.43), secondary AOR: 4.69 (2.23 to 9.84) and tertiary AOR: 3.60 (1.20 to 10.83)) levels, being from model households (AOR: 4.38 (1.33 to 14.47)), households with middle (AOR: 3.50 (1.13 to 8.37)) or high (AOR: 3.10 (1.25 to 7.67)) wealth index, enrolment in productive safety net programmes (AOR: 2.37 (1.25 to 4.50)) and age at first pregnancy (AOR: 1.16 (1.08 to 1.24)) were positively associated with women’s empowerment. From the community-level characteristics, dowry (AOR: 1.82 (1.10 to 5.30)) and perceived good availability of justice for women and girls (AOR: 3.00 (1.05 to 8.60)) were positively associated with women’s empowerment. Conversely, the history of an adverse pregnancy outcome was negatively associated with women’s empowerment (AOR: 0.51 (0.26 to 0.99)).

Conclusion

The overall status of women’s empowerment in Tigray was very low. Husband’s education, being model household, wealth index, enrolment in productive safety net programmes, dowry practice, perceived good availability of justice for women and girls and age at first pregnancy were found to be significant factors associated with women’s empowerment. In view of these empirical insights, several policy recommendations are proposed. First, interventions should focus on the identified associated factors, particularly in areas with lower empowerment scores such as partner’s education, improving access to justice for women, addressing harmful community practices such as dowry and strengthening household economic status through social protection and livelihood programmes. Second, the promotion of gender-balanced household dynamics through awareness campaigns and policy incentives could make a substantial contribution to women’s empowerment in the study area.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• One of the strengths is measuring women’s empowerment using multidimensional measurement indexes that consider individual, intrafamilial and community-level characteristics.

• Emphasising the dynamic nature of empowerment and collecting primary data within a specific context, allowing the inclusion of multiple dimensions of women’s empowerment.

• The survey’s design was cross-sectional, and a cause-and-effect relationship cannot be made.

• Tigrayan society is patriarchal, which leads to women’s responses being influenced by higher social desirability bias.

• The survey did not assess spouses’ perspectives on women’s empowerment and its determinants, and responses from both wives and their husbands to these questions were missing.

Background

Women’s empowerment is a complex and multidimensional concept that varies widely. Kabeer has defined it as a ‘process of change by which those who have been denied the ability to make choices acquire such an ability’’.^{1 2} This definition views empowerment as a dynamic process that changes over time and comprises three interrelated dimensions: resources, agency and achievements or outcomes.^{1 3} VeneKlasen and Miller also describe the concept as a process of transformation from a limited to an enhanced state of power and suggested three levels of change: personal, relational and environmental.⁴

Over the past decade, approaches to measuring women’s empowerment have advanced significantly, leading to the development of numerous indexes that capture women’s empowerment through key socioeconomic and health indicators.⁵ Based on prior studies, a numeric empowerment score is created from key domains: participation in household decisions^{6 7}; attitudes towards wife-beating⁶; social independence^{8 9} (education, media exposure, age at marriage, age/education gaps)⁶; economic power⁷ (asset ownership, income, employment) and personal agency^{8 10} (mobility, self-esteem, freedom from coercion and communication with partners on reproductive/sexual health).^{9 11} Studies measure and score women’s empowerment in different ways: as the share of indicators where a woman meets a minimum threshold of power,¹² as a binary status based on a cut-off score (eg, ≥75%–80%),^{13 14} or as a continuous index where higher scores reflect greater decision-making power and agency.^{15 16}

One of the most prominent indices of empowerment is the Survey-based Women’s emPowERment (SWPER) index, developed using Demographic and Health Surveys (DHS) data, which captures three empowerment domains: decision-making, social independence and attitudes to violence.⁶

Women’s empowerment is a core human right and a key objective of development policy.^{17 18} To that end, the United Nations (UN) adopted the fifth Sustainable Development Goal in 2015 to achieve gender equality and empower all women and girls by 2030 and to accelerate sustainable development.^{19 20} However, women’s empowerment remained very low, and its measurement continues to be a topic of debate.⁶ Despite their significant contributions to the development of civilisation, women have historically not been regarded as equal to men both by her family and the community.²⁰ As a result, they have often been denied equal privileges and opportunities in education, healthcare, economic resources, leadership and decision-making compared with men.²¹

Globally, women’s empowerment was about 61%, and 3.1 billion women and girls (more than 90%) of the world’s female population live in countries characterised by low women’s empowerment and low or middle performance in achieving gender parity.²² The proportion of women’s empowerment score ranges from Mali (20.4%),²³ Nigeria (48%)²⁴ and the Democratic Republic of Congo (65%)²⁰ for low human development index countries to 85.9% in high human development index countries. Sub-Saharan Africa (51.9%) and Oceania, excluding Australia and New Zealand (63%), on average, lag behind the global level by about 12%.²²

Despite the Ethiopian government’s commitment to advancing women’s empowerment and gender equality through a range of institutional and policy measures like explanation of primary healthcare services, women developmental army programmes, revitalising community health worker to health extension programme, community-based health insurances and charge-free maternity services,25,29 women’s empowerment remains an unmet objective across various aspects of development including educational attainment, employment opportunities, social and political leadership, decision-making and access to information.30,32 About 80% of the female population in the country grapples with persistent gender disparities, undermining efforts towards women’s empowerment.³³ Ethiopia also ranked 115 out of 144 countries in the gender equality index, calculated from women’s economic participation, educational attainment, health and survival and political participation.^{34 35} A recent study on women’s empowerment in Ethiopia revealed consistently low levels, with an overall empowerment score of only 23.7%.³⁶ Participation in key household decisions remained limited, as only 30.9% of women reported involvement in decision-making. Additionally, acceptance of wife-beating was alarmingly high, with 67.5% of women agreeing that a husband is justified in beating his wife under at least one circumstance. These findings highlight substantial gaps in women’s autonomy, agency and gender-equitable attitudes in the country.³⁶

The Tigray state, a patriarchal society within Ethiopia, is assumed to implement similar women’s empowerment measures to those of the whole of Ethiopia. For instance, Tigrayan women encounter substantial gender disparities in educational achievements,³⁷ economic prospects³⁸ access to information and tolerance of intimate partner violence,³⁷ as is true in Ethiopia. From 1990 to 2019, Tigray witnessed significant progress in health service utilisation, particularly increased maternal and child health services and reduced maternal mortality rates.³⁹ Tragically, the outbreak of war in Tigray severely impacted women empowerment; women’s representation in decision-making and leadership positions declined from 48% before the war to less than 5% after the war,⁴⁰ and 43.3% of women experienced at least one type of gender-based violence.⁴¹ This suggests that women’s and girls’ empowerment will remain elusive until gender gaps are eliminated. Gender equality and women’s empowerment in households, communities and societies are critical determinants of the health and well-being of women and their newborns.⁴²

Previous studies revealed a positive association between women’s empowerment and the use of health services.⁴³ For instance, empowered women are more likely to use reproductive health services like family planning, maternal, neonatal and child health service utilisations,^{31 44} health-seeking behaviour⁴⁵ and risk reduction of domestic violence.⁴⁶ In addition, few studies showed that individual factors such as women’s education, age at first birth²⁰; intra-familial factors such as income and family structure²⁰; and societal norms such as dowry practice and limiting gender equality by defining strict roles and expectations for women commonly influence women’s empowerment.^{20 24 36 47}

Moreover, the limited existing studies have primarily relied on secondary national survey data and incorporated only a narrow set of empowerment indicators, leading to notable methodological limitations.⁴⁸ Despite the growing attention to women’s empowerment globally, there is a lack of empirical studies assessing the status of women’s empowerment and its associated factors using multidimensional empowerment index in Tigray and Ethiopia. Therefore, without a clear understanding of the status and factors influencing women’s empowerment in Tigray, efforts to promote gender equality in the region are likely to be inefficient and lack a contextual evidence base. To address the existing gaps, it is essential to identify individual, intrafamilial and community-level covariates that influence women’s empowerment.

In this study, women’s empowerment was measured using a composite score based on the SWPER framework, which captures key dimensions of empowerment in low and middle-income countries and from the DHS, and alignment with the existing literature and context-specific tools and materials in Ethiopia^{6 9}, reflecting key dimensions including social independence, women’s decision-making power within the household, attitudes towards gender-based violence, participation in social networking groups or unions and ownership of household assets. We believe that understanding what constitutes these scores is essential for interpreting their association with maternal and child health outcomes and contribute to new knowledge and develop effective strategies for empowering women and narrowing gender gaps in postwar rehabilitation, recovery and reconstruction efforts of Tigray.

Therefore, this study aimed to assess women’s empowerment status and identify its associated factors using a multidimensional approach. We employed an adapted SWPER++index, along with two additional domains—social networking and household asset ownership—based on context-specific primary data collected in Tigray, Ethiopia.

Methods

Study settings and period

The research was conducted in the Tigray region of Ethiopia. Tigray is one of the 11 regional states within the Federal Democratic Republic of Ethiopia, and it shares its borders with Eritrea to the North, Afar regional state to the East, Amhara regional state to the South and Sudan to the West. The majority of the population, constituting 81.5%, resides in rural areas.⁴⁹ Tigray has seven administrative zones, including Mekelle, which is a special zone. The region consists of 93 districts (rural 57 and urban 36), also known as Woredas, and 865 smaller governmental administrative units called Tabias (rural 698 and urban 167).⁵⁰ The regional state has a total population of more than 7 million, with 50.8% females and 49.2% males.⁵¹

Data were collected between 20 May and 3 August 2024, from three randomly selected zones in the Southeastern, Central and Northwestern (out of the six) accessible zones of Tigray regional state in a postconflict setting. The study did not include the Western zone and some Northwest and Central zone districts for security reasons. Of the selected zones, we included nine districts or ‘woredas’, namely: Hintalo, Enderta, Rural Adwa, Laelay Maychew, Naeder, Aksum town, Laelay Koraro, Tahitay Koraro and Shire Endasilassie town (figure 1). In general, according to the population projection from the 2007 national census for 2024, the three zones were expected to have a total of 114 985 pregnant women; of these, the total number in the selected district ‘woredas’ was 31 144.⁵²

Figure 1. Location of study districts in Tigray state, Ethiopia 2024.

Study design and population

A community-based cross-sectional survey was conducted. The source population comprised all pregnant women in the community during the baseline survey and permanent residents (living for more than 6 months) in the region. The participants in the study were all married women aged 15–49 years, with a gestational age of 8 weeks or more. Since women’s empowerment is highly context-specific, temporary residents and internally displaced women affected by the war were excluded, as their situations may not accurately reflect the true empowerment status in the selected communities. Critically ill women were also excluded because of their limited ability to provide informed consent and to fully understand the empowerment indicators.

Sample size estimation

To determine the sample size, two population-proportion formulas were used. Then Epi-Info Statclc application V.7.2.5.0 was used to calculate the sample sizes. Thus, considering the following assumptions for double population formula: the proportion of an outcome among exposed group was 34% (p1=0.340), and among non-exposed group was 23.9% (p2=0.239)⁵³; pooled population proportion (p=0.193); r=1:2 ratio of exposure to non-exposure; 5% significant level at α level (two-sided); 80% power, design effect 1.8 and 10 non-response rate. Therefore, 1512 pregnant women were used as the final sample for this study.

Sampling techniques

A multistage cluster sampling technique was employed for the selection of the study units. Initially, three of the six accessible zones (namely, the central, Northwest and Southeastern zones) were selected using a simple random sampling method. Second, nine districts, or ‘woreda’ were selected randomly from the three zones: four from the central zone, three from the Northwest zone and two from the Southeastern zone. Third, 43 Tabias from 113 Tabias were selected using simple random sampling techniques. Finally, we randomly selected 86 ‘kushet’ (the smallest administrative unit in Ethiopia) from each subdistrict. Finally, all households in the selected kushets (clusters) with eligible pregnant women were included. In households with more than one eligible woman, all eligible participants were interviewed.

Measurement of variables

Outcome: the primary outcome of the study was the status of women’s empowerment, specifically whether a woman is classified as ‘empowered’ or not based on a cumulative empowerment index. It was measured using multidimensional, context-specific indexes constructed from five key domains, comprising 24 empowerment indicators/questions: decision-making power, social independence, attitude towards the justification of wife-beating, social networking membership and household asset ownership. These domains and items in our questionnaire were developed by adapting questions from the validated SWPER Global 2017 tool, the DHS and the relevant literature, while also incorporating context-specific instruments and guidelines to align with our study objectives.^{6 9}

To assess decision-making power, four questions were asked about women’s involvement in decisions regarding their healthcare, household purchases, use of money her husband earns and visiting family or relatives. Women’s responses were dichotomised to 1 if a woman had any decision (alone or jointly), and 0 if a woman had no say in one or more decisions. The weight assigned to each domain was shared equally by the indicators within the corresponding domain. Total scores ranged from 0 to 4. Accordingly, those women who scored the mean and above the mean were considered as empowered, and those who scored below the mean were considered as of disempowered.⁶

For the attitude towards the justification of wife-beating, respondents were asked five questions to determine whether they believe a man was justified in beating his wife under certain circumstances, such as going out without telling him, arguing with her husband, neglecting children, refusing sexual intercourse and burning food. The response ‘yes’ to the above questions means that she believes wife-beating is justified, and ‘no’ means she rejects wife-beating for that reason. As this index is negatively related to women’s empowerment, for analysis purposes, these questions were recorded oppositely from the first indicator (ie, 1 for ‘no’ and 0 for ‘yes’). If the woman disagreed with all these reasons, they are assumed to be empowered. Therefore, an overall binary variable was created with values 1 and 0, where 1 indicates that the woman disagrees with all the reasons, and 0 indicates that she agrees with 4 or fewer reasons.^{36 54}

The social independence domain was measured using relevant questions from SWEPER and the DHS. These include: (1) frequency of reading the newspapers or magazines; (2) frequency of listening radio or watching television; (3) a woman’s education in years of schooling completed; (4) education difference (woman’s minus husband’s education); (5) gap in ages (wife’s age minus husband’s age); (6) age at first marriage; (7) age at first cohabitation and (8) the woman’s age at first childbirth.³⁷ Recent studies in East Africa and similar settings use spousal age and education gaps as indicators of women’s social independence. Larger age and educational gaps weaken women’s autonomy, decision-making and bargaining power, reflecting patriarchal gender norms and reduced female agency.^{6 55} The domains were assigned equal weight; 1 point each. The weight assigned to each domain was shared equally by the indicators within the corresponding domain. Then, we calculated the total score for each domain. Using the mean score as a cut-off, those who scored above the mean were classified as empowered otherwise as disempowered (table 1).

Table 1. Dimensions, indicators, responses and cut-offs of women’s empowerment multi-dimensional Index (WEMI), Tigray, Ethiopia, 2024 (n=1477).

Domains	Indicators	Responses	Empowered if woman (cut-off)
Social independence	Women’s education in complete years of schooling	Years	Completed secondary or higher education
	Frequency of reading newspaper/magazines	Not at all=0 <once a week =1 >once a week =2	Read once a week or more
	Frequency of watching television/listening to radio	Not at all=0 <once a week =1 >once a week =2	Watch/listen once a week or more
	Age of women at first birth	Years	19 years and above
	Age of women at first marriage	Years	18 years and above
	Age of women at first cohabitation	Years	18 years and above
	Age difference	Woman’s minus husband’s age	Less than 5 years younger than her husband
	Educational differences	Woman’s minus husband’s years of school.	has equal or higher education than her husband
	Mean score of social independence (out of 0–8 total points)	1 if ‘≥mean score’, ‘0’ less than mean score	-Mean (3.48); % ≥ mean=53.4%
Household decision-making	Own healthcare	1 if ‘respondent alone and jointly’, ‘0’ otherwise’	Take decision alone/jointly with her husband
	Major household purchase	1 if ‘respondent alone and jointly’, ‘0’ otherwise’	Take decision alone/jointly with her husband
	Family/relatives visits	1 if ‘respondent alone and jointly’, ‘0’ otherwise’	Take decision alone/jointly with her husband
	Husband’s earning	1 if ‘respondent alone and jointly’, ‘0’ otherwise’	Take decision alone/jointly with her husband
	Mean score of decision-making (out of 0–4 total points)	1 if ‘≥mean score’, ‘0’ less than mean score	Mean (3.14); % ≥ mean = 54.4%
Attitude towards violence	Wife beating justified: goes out without telling her husband	1 if ‘no’ 0 ‘otherwise’	Does not agree
	Wife beating justified: neglects children	1 if ‘no’ 0 ‘otherwise’	Does not agree
	Wife beating justified: argues with him	1 if ‘no’ 0 ‘otherwise’	Does not agree
	Wife beating justified: refuses to have sex with husband	1 if ‘no’ 0 ‘otherwise’	Does not agree
	Wife beating justified: burns the food	1 if ‘no’ 0 ‘otherwise’	Does not agree
	Mean score of attitude towards violence (out of 0–5 total points)	1 if≥ if ‘no’ to all responses, ‘0’ otherwise’	Mean (3.6); % ≥ mean=12.8%
Social networking	Membership participation: in Health-related group	1 if ‘yes’, 0 ‘no’	Yes
	Membership participation: in Leadership and political organisations	1 if ‘yes’ 0 ‘otherwise’	Yes
	Membership participation: in economic unions like saving and credit	1 if ‘yes’ 0 ‘otherwise’	Yes
	Membership participation: cultural/faith-based organisations	1 if ‘yes’ 0 ‘otherwise’	Yes
	Mean score of social networking membership (out of 0–4 total points)	1 if’ ≥mean score’, ‘0’ less than mean score	Mean (0.48); % ≥ mean=31.1%
Asset ownership	Own land	1 if ‘respondent alone or jointly’, ‘0’ otherwise’	Owns alone/jointly with her husband
	Own house	1 if ‘respondent alone or jointly’, ‘0’ otherwise’	Owns alone/jointly with her husband
	Own mobile	1 if ‘yes’, ‘0’ if ‘no’	Yes
	Mean score of social networking membership (out of 0–3 total points)	1 if ‘≥mean score’, ‘0’ less than mean score	Mean (1.11); % ≥ mean=29.7%
Total women’s empowerment mean score (out of 0–5 total points)		1 if ‘≥80% score or/≥4 out of 5 points’, ‘0’ less than 80% scores/<4 out of 5 points	Mean (2.05)
%women scored≥80% or ≥ 4(out of 0–5 total points)			8.26%

Social networking domains were measured by women’s participation in (1) health issues group like Women Developmental Army (WDA), pregnancy forum and one to five network, (2) economic group membership (saving and credit associations like Adeday bank or other self-help group, etc, (3) leadership or political participation (from local to region level leadership, women’s affair, women’s association, etc, (4) cultural or faith-based associations like ‘Mahber’, Edir, equb, etc. Accordingly, a binary variable was created: 1 for women who participated in any social network group or association, and 0 for those who did not. Using an equal-weight assumption across the domains, women scoring above the mean were classified as empowered, while those scoring below were classified as disempowered.

Concerning the fifth domain, ownership of assets comprises three indicators: ownership of land, a house and mobile phones. Ownership of title deeds for houses, land, or phones was coded as 0 for participants with no ownership and one if the woman owned the asset alone or jointly with her husband/partner. In our analysis, an indicator for ownership of assets was generated by summing land ownership, house ownership and mobile phone ownership. By virtue of the coding of the parameters used to determine asset ownership, the overall asset ownership values ranged; using the mean score as a cut-off, those who scored above the mean were classified as ‘1’ otherwise as ‘0’⁵⁶ (table 1).

To construct the cumulative women’s empowerment index, equal weight was assigned to all domains, and the weight of domain was distributed equally to its subdomains. Therefore, based on the established cut-off value from previous studies^{13 14} and incorporating women’s own perspectives on empowerment, a respondent was classified as empowered if her empowerment score was ≥0.8 (ie, met 4 out of 5 domains), and as disempowered if her score was <0.8 (ie, ≤3 out of 5 domains).

Covariates of the study are grouped into four categories: (1) individual variables included women’s sociodemographic factors, access to information and behavioural factors. Women’s current age was examined as a continuous variable⁵⁷; maternal education was categorised into four categories: none, primary (1–8 grade), secondary (9–12 grade) and college and above; (ii) intrafamilial/household factors^{58 59} included wealth index,^{57 58} marital structure⁵⁷ and relationship, husband’s age, education, occupation, sex of the household head; (iii) community-based health insurance status, enrolment in the SafetyNet programme and family size; and (4) respondents’ reproductive and obstetric factors included age at first pregnancy (<19 years vs >19 years), number of pregnancies, number of births and previous obstetric complications. On the other hand, (iv) community-based factors consisted of place of residence (rural and urban), access to facility at local level, distances to nearby health facility by asking ‘How long does it take to go to the nearby health facility and back home on foot?’, access to transportation, existing culture and norms like dowry and other materials provided at marriage from women’s families.

The wealth index was assessed by asking about housing characteristics, access to improved drinking water and sanitation facilities and ownership of household assets, land and livestock. First, the dichotomised socioeconomic proxy indicator variables were standardised using principal component analysis, and factor coefficient scores were created. Then, the indicator values were multiplied by the factor scores and summed up to produce a standardised wealth index value. Finally, using the factor scores with the largest proportion of the variance, the wealth index was categorised into quintiles designating the lowest to the highest economic status.¹³ Likewise, time to fetch water was collected at inclusion by asking, ‘What is the time needed to fetch improved drinking water from the nearest source in minutes?’ Then, it was dichotomised at a cut-off point of 30, with time not exceeding 30 min showing better service.⁶⁰ Based on previous studies, women from model households and those enrolled in the Productive Safety Net Programme (PSNP) are associated with greater social and economic empowerment.^{61 62} Accordingly, we hypothesised that belonging to a model household and participating in social protection schemes would positively influence women’s empowerment in our community. Therefore, the model household was defined as a household that received short-term training and implemented the minimum package of the community health extension programme in Ethiopia after the training, and it was chosen as a model household.⁶³ Furthermore, the PSNP is a social protection programme by the Government of Ethiopia targeting food-insecure households, designed as an asset protection mechanism at the household level to create productive community assets.⁶⁴

Data quality control

Data collection was supervised by Midwives and Nurses (MSc). Data collectors and supervisors were trained in the methods, objectives, confidentiality and other technical aspects of the study for 3 days before data collection commenced. 5% of the tools were pretested to ensure clarity, wording, logical sequence and the relevance of the questionnaires. Furthermore, a smart mobile-based approach (KoboToolbox) that ensures a skip pattern: immediate scanning of the tool’s quality on the server avoids the cost of paper and is more friendly to data collectors. The questionnaires were prepared in English, translated into Tigrigna (the local language) and then retranslated into English by people proficient in both languages to maintain consistency. Data collectors were recruited from outside the study districts for the cross-sectional study. During actual data collection, the principal investigator and supervisors frequently supervised and checked the work of data collectors and provided clarification and directions to those in doubt.

Data processing and statistical analysis

The data collected with the Open Data Kit (KoboToolbox) were exported to SPSS V.27.0 for cleaning, and necessary corrections were made before the analysis. Meanwhile, analysis was conducted using both SPSS V.27 and Stata V.17.0. We summarised the characteristics of the study participants using descriptive statistics, including mean (±SD), median (IQR), frequency and proportions. Bivariate logistic regression analyses were done for the explanatory variables with the dependent variable to select contender variables for multiple logistic regression models. Variables found to be significant at p<0.05 were analysed for their relative effects using multiple logistic regression. Also, multicollinearity was assessed using the variance inflation factor. Model fitness was assessed using the Hosmer-Lemeshow test and the Receiver Operating Characteristic curve. Thus, the effect sizes of the sample were described using the adjusted OR (AOR), and those of the population were described using a 95% CI. A p value of less than or equal to 0.05 was set as a cut-off for statistical significance.

Results

Sociodemographic characteristics of respondents (1477)

A total of 1477 women were included in the study (response rate: 97.7%). The women’s mean age (±SD) was 29.3±5.9 years. Ninety-nine per cent (98.8%) of the participants were an orthodox Christian follower. Of the women examined, 590 (39.9%) had received primary education, and 44% were housewives. The majority of the women (91.4%) lived with their partner, and almost three-fourths of the women (74.5%) were rural residents. Only 6% of the women reported having access to the internet, and 4 of 10 respondents reported owning a mobile phone. At the same time, only 6% of them indicated that they had used their devices for financial transactions. In addition, the results showed that 1238 (83.8%) of the women were in a nuclear family and 85% were in a monogamous union. Also, the majority (82.67%) of the participants had access to an improved drinking water source, whereas only 294 (19.9%) had access to an improved sanitation facility. Close to one-third of the respondents’ spouses (31.5%) did not receive any formal education, and 41.3% were farmers (table 2).

Table 2. Sociodemographic characteristics of women and their households in Tigray, North of Ethiopia, 2024 (n=1477).

Variables	n(%)/mean (SD)/median (IQR)
Age at inclusion in years	29.3 (5.9)
Education
No formal education	257 (17.4%)
Primary (1–8)	590 (39.9%)
Secondary (9–12)	514 (34.8%)
College and above	116 (7.9%)
Religion
Orthodox Christian	1460 (98.8%)
Others	17 (1.2%)
Ethnicity
Tigrayan	1474 (99.8%)
Amhara	3 (0.2%)
Occupation (maternal)
Farmer	546 (37.0%)
Housewife	650 (44.0%)
Employed	94 (6.4%)
Daily labour	90 (6.0%)
Merchant	75 (5.1%)
Others	22 (1.5%)
Respondents living with her partner	1350 (91.4%)
Usual residence
Urban	377 (25.5%)
Rural	1100 (74.5%)
Own mobile phone	620 (42.0%)
Mobile phones for financial transactions (620)	59 (9.5%)
Ever used the internet	89 (6.0%)
Frequency of internet use (n=89)
No regular exposure	11 (12.36%)
Regular exposure	78 (87.64%)
Own bank/financial account	385 (26.1%)
Landownership	211 (14.3%)
Maternal self-esteem
Low	702 (47.53%)
Good	775 (52.47%)
Sex of the household
Male	1319 (89.3%)
Female	158 (10.7%)
Family structure
Nuclear	1238 (83.8%)
Extended	239 (16.2%)
Marital relationship
Monogamous union	1255 (85.0%)
Her second or more union	222 (15.0%)
Husband age (cont.	37.4 (7.9)
Husband educational status
No formal education	465 (31.5%)
Primary (1–8)	442 (29.9%)
Secondary (9–12)	420 (28.4%)
College and above	150 (10.2%)
Husband occupation
Farmer	610 (41.3%)
Daily labour	519 (35.1%)
Employed	215 (14.6%)
Others	133 (9.0%)
Is your household is model?	18 (1.2%)
wealth index
Poorest	276 (18.7%)
Poor	315 (21.3%)
Middle	285 (19.3%)
Richer	295 (20.0%)
Richest	306 (20.7%)
Community-based health insurance	102 (6.9%)
Safety net programme enrolment	147 (10%)
Source of drinking water
Unimproved access	256 (17.3%)
Improved access	1221 (82.7%)
Sanitary facility (toilet)
Unimproved sanitation	1183 (80.09%)
Improved sanitation	294 (19.91%)
Time needed to fetch water
30 min and below	1129 (76.4%)
30 min and above	348 (23.56%)
Income differences
Equal	405 (27.42%)
Wife less than husband	993 (67.13%)
Wife more than husband	79 (5.35%)

Reproductive and obstetric conditions of participants

In the study, two-thirds of respondents’ marriages (66.2%) were arranged by their parents without seeking the consent of their daughters. The respondents’ mean age at first marriage was 18.6 (SD: 3.26 years), and the mean age at first pregnancy was 20.5 years. The median parity of the women before the index pregnancy was 3, while 208 women (14.08%) had a history of an adverse birth outcome. Approximately one in three women has never used modern contraceptives, while 64.79% reported that their husbands are involved in their contraceptive use decisions (table 3).

Table 3. Obstetrics and reproductive health-related characteristics of the participants in Tigray, North of Ethiopia, 2024 (n=1477).

Variables	n (%)/mean (SD)/median
Choice of life partner
Respondent choice	523 (33.8%)
Family choice	1025 (66.2%)
Age at first marriage	18.9 (10–37)
Age at first cohabitation	20.5 (14–37)
Age at first pregnancy	20.6 (13–41)
Age at first birth (n=1243)	20.6 (14–38)
Gravidity before the index pregnancy (n=1295)	4 (2–11)
Parity before the index pregnancy (n=1295)	3(0–10)
Ever use modern contraceptive	1066 (72.17%)
Husband/partner support to use modern contraceptive
No	387 (26.2%)
Yes	957 (64.79%)
I do not know	133 (9.0%)
History of at least one adverse birth outcomes	208 (14.08%)
History of at least one abortion	257 (17.40%)
History of at least one known medical illness diagnosed by health professional	179 (12.12%)
Desired number of children
Both want same	1160 (78.54%
partner wants more	149 (10.09%)
partner wants fewer	81 (5.48%)
I do not know	87 (5.89%)
Total number of children alive
0	290 (19.63%)
1–2	560 (37.91%)
3–4	440 (29.79%)
5+	187 (12.66%)

Community-level characteristics of participants

The average walking distance to the nearest health facility was 34.3 min to a health post (the lowest tier in the three-tiered primary healthcare system typically serving people in the rural areas) followed by 58.3 min to health centre and the most prevalent modes of transportation used to go to primary healthcare unit were public cars and motor vehicles (table 4).

Table 4. Community characteristics of the study participants in Tigray, North of Ethiopia, 2024 (n=1477).

Variables	n (%)/mean (SD)/median (IQR)
Access of local facilities
Low facility at community level	848 (57.41%)
Good facility at community level	629 (42.59%)
Transport access to nearby health facilities	1006 (68.11%)
Means of transportation to nearby HFs (n=1006)
Ambulance	213 (21.17%)
Public car	497 (49.40%)
Motor vehicles	290 (28.83%)
Others	6 (0.60%)
Time to reach the nearest health post on foot (in minutes)	34.3 (2–180)
Time to reach the nearest health centre on foot (in minutes)	58.3 (5–605)
Time to reach the nearest hospital on foot (in hours)	3.23 (0.02–61)
Dowry received from parents at marriage	892 (60.39%)
Any type of materials received from parents at marriage	319 (21.60%)

HFs, Health Facilities.

Perception of respondents on the availability of facilities at the local level

In the study, respondents’ perceptions of eight key local facilities were assessed using a three-point rating scale: good, average, and bad. It was found that 45.09% of the respondents ranked educational facilities in their area as good. About 34.6% of respondents felt that women’s security in their area was good, while 58.36% described the security level as average.

Local leadership can contribute to the betterment of the community; however, only 5.89% of the respondents recorded good remarks for the local governors’ leadership. Better job opportunities in an area can help women attain economic independence. In addition, only 13.66% of the respondents thought that justice facilities for women and girls were good (figure 2).

Figure 2. Proportion distribution and percentages of respondents’ perceptions about the facilities at local level Tigray, Northern Ethiopia (n=1477).

Overall women’s empowerment status in Tigray

Among the women’s empowerment indicators, nearly 94% of women reported having decision-making power over their own healthcare. About 78.2% disagreed that wife beating is justified if a wife burns food. However, only 3.9% of participants reported regularly reading newspapers or magazines. In addition, 14.4% owned land and fewer than 8% reported involvement in leadership or political activities (figure 3A).

Figure 3. Women’s empowerment in Tigray, Northern Ethiopia (n=1477). (A) By indicator. (B) By domain and overall. HH, Household; SHG, Self Help Group.

The overall magnitude of women’s empowerment in Tigray was found to be 8.26% (95% CI: 6.96 to 9.78), leaving an empowerment deficit of approximately 91.74% (95% CI: 90.22% to 93.04%). When stratified by domain, the household decision-making domain obtained the highest empowerment (54.5% CI (51.95%–57.03%)), followed by social independence (53.4% CI (50.87%–55.95%)) and attitude towards violence obtained the lowest empowerment (12.8% CI (11.5%–14.96%). The majority of indicators in the decision-making domain suggest that respondents typically share decision-making responsibilities with their husband/partner, with nearly half reporting joint participation in household decision-making. Another domain of women’s empowerment was their attitude towards wife-beating, with only 12.8% of women disagree with all justifications for it (figure 3B).

To examine factors associated with women’s empowerment, we used the cumulative empowerment index across five domains. Results of the unadjusted and adjusted logit regression analysis are given in table 5. In the adjusted model, intrafamilial-level factors were positively associated with women’s empowerment; women who completed primary education had 2.66 times higher odds of being empowered compared with women with no education (AOR: 2.66 (1.30–5.43). Similarly, women who completed secondary school (AOR: 4.69 (2.23–9.84)) and college and above (AOR: 3.60 (1.20–10.83)) had 4.69 and 3.6 times higher odds of being empowered, respectively, compared with their counterparts. Furthermore, women from model households were 4.38 times more likely to be empowered than their counterparts (AOR: 4.38 (1.33–14.47). Likewise, a household with a middle wealth index (AOR: 3.50 (1.13–8.37)) and a high wealth index (AOR: 3.10 (1.25–7.67) were 3.5 and 3.10 times more likely to be empowered, respectively, compared with women who fall in the lowest category. On the other hand, women enrolled in the PSNP (AOR: 2.37 (95% CI: 1.25 to 4.50) had 2.37 times higher odds of being empowered compared with women who were not enrolled.

Table 5. Unadjusted and adjusted logit regression analysis of factors associated with cumulative multidimensional measurement of women’s empowerment in Tigray, northern Ethiopia (n=1477).

Characteristics	Unadjusted	P value	Adjusted	P value
Maternal education
None (ref)
Primary (1–8)	1.4 (053–2.02)	0.91	0.49 (0.38–1.78)	0.62
Secondary (9–12)	2.53 (1.36–0.4.69)	0.003	1.69 (0.90–4.32)	0.09
College and above	3.22 (1.51–6.87)	0.003	1.26 (0.67–6.56)	0.21
Self-esteem
Low (ref)
Good	1.55 (1.06–2.27)	0.024	0.93 (0.58–1.51)	0.77
Family structure
Nuclear (ref)
Extended	0.39 (0.19–0.78)	0.008	0.45 (0.19–1.03)	0.058
Husband education
None (ref)
Primary (1–8)	2.58 (1.39–4.79)	0.003	2.66 (1.30–5.43)	0.007
Secondary (9–12)	4.33 (2.40–7.81)	0.001	4.69 (2.23–9.84)	0.001
College and above	4.35 (2.15–8.80)	0.001	3.60 (1.20–10.83)	0.022
Husband occupation
Farmer (ref)
Daily labour	0.62 (0.39–0.97)	0.037	0.42 (0.23–0.76)	0.004
Employer	1.39 (0.86–2.27)	0.182	0.50 (0.22–1.11)	0.087
Others	0.54 (0.24–1.20)	0.13	0.22 (0.08–0.64)	0.006
Being from a model household
No (ref)
Yes	7.43 (2.83–19.60)	0.001	4.38 (1.33–14.47)	0.015
Wealth index
Lowest (ref)
Low	2.23 (1.01–4.92)	0.048	2.04 (0.82–5.10)	0.126
Middle	3.23 (1.50–6.98)	0.003	3.50 (1.43–8.37)	0.006
High	3.74 (1.75–7.96)	0.001	3.10 (1.25–7.67)	0.015
Highest	3.22 (1.50–8.92)	0.003	2.12 (0.97–5.67)	0.135
Community-based health insurance
No (ref)
Yes	1.87 (1.03–3.40)	0.041	1.33 (0.64–2.77)	0.44
SafetyNet enrolled HH
No (ref)
Yes	1.90 (1.14–3.17)	0.015	2.37 (1.25–4.50)	0.010
Age at first pregnancy	1.13 (1.70–1.18)	0.001	1.16 (1.08–1.24)	0.001
Gravidity	0.86 (0.77–0.97)	0.014	1.02 (0.87–1.20)	0.817
Hx of abortion	0.49 (0.27–0.91)	0.024	0.51 (0.26–0.99)	0.048
Desired number of children
Both want same (ref)
Parents want more	0.41 (0.18–0.95)	0.037	0.44 (0.18–1.08)	0.072
Partner wants fewer	0.51 (0.18–1.41)	0.19	0.62 (0.20–1.88)	0.396
I do not know	0.47 (0.17–1.31)	0.15	1.19 (0.38–3.69)	0.766
Dowry received at marriage
No (ref)
Yes	1.8 (1.18–2.67)	0.006	1.82 (1.10–3.00)	0.019
Distance to nearby HF (Hc)	0.99 (0.98–0.99)	0.030	0.99 (0.99–1.01)	0.878
Local facilities
Transportation facilities for women
Bad (ref)
Fair	1.05 (0.77–1.61)	0.77	1.10 (0.66–1.81)	0.740
Good	1.94 (1.01–3.74	0.047	2.13 (0.95–4.80)	0.067
Justice facility for women and girls
Bad (ref)
Fair	1.31 (1.16–2.56	0.008	1.34 (0.78–2.17)	0.305
Good	1.94 (0.78–4.81)	0.15	3.00 (1.05–8.60)	0.040

HFs, Health Facilities; HH, Household.

Age at first pregnancy showed a significant positive association with women’s empowerment (AOR: 1.16, 95% CI 1.08 to 1.24), indicating that a 1 year increase in age at first pregnancy is associated with 1.16-fold higher odds of being empowered. Conversely, women whose husbands were daily labourers (AOR=0.42; 95% CI 0.23 to 0.76) and those whose husbands were students (AOR=0.22; 95% CI 0.08 to 0.64) were 58% and 78% less likely to be empowered, respectively, compared with women whose husbands were farmers. Likewise, women with a history of at least one adverse pregnancy outcome (abortion) (AOR=0.51; 95% CI 0.26 to 0.99) were 49% less likely to be empowered compared with their counterparts. Among community-level factors, women who received a dowry from their maternal family at the time of marriage had 1.82 times higher odds of being empowered (AOR=1.82; 95% CI 1.10 to 5.30) than their counterparts. Similarly, respondents living in communities with better justice facilities for women and girls had threefold higher odds of being empowered (AOR=3.00; 95% CI 1.05 to 8.60) than those reporting the lowest level of justice services in their community (table 5).

Discussion

This study aimed to assess the status of the women’s empowerment index and the factors associated with it in Tigray, Ethiopia.

The overall magnitude of women’s empowerment in Tigray was found to be 8.26% (95% CI 6.96 to 9.78), leaving an empowerment deficit of approximately 91.74% (95% CI 90.22% to 93.04%). When stratified by domain, household decision-making was high at 54.3% but attitudes towards violence were low at 12.8%.

The level of women’s empowerment in this study was unexpectedly lower than in previous studies conducted in Mali (20.4%),²³ Ethiopia (27.7%),³⁶ Nigeria (48%),²⁴ the Democratic Republic of Congo (65%)²⁰ and the worldwide empowerment index of 60.7%.^{17 22} The possible reasons for the discrepancy might be due to the devastating war that erupted in Tigray, Ethiopia, on 4 November 2020, leading to a significant humanitarian crisis. As a result, the war in Tigray caused a massive internal displacement and over 70 000 refugees to neighbouring Sudan.⁶⁵ Additional possible reasons might be due to persistent dominance of patriarchal norms and traditional gender roles, leading to lower status of women in the study area,⁶⁶ tools of measurement and time and study setting differences.

In this study, women’s empowerment was measured by considering the composite values of decision-making, social independence, attitude toward gender norms, social networking and asset ownership, whereas empowerment in Ethiopia, Nepal and other African countries was measured in terms of the DHS empowerment indicators.⁶⁷ Due to their limited scope, such studies may not fully capture women’s participation in political, economic, social and cultural life, nor contextual factors. Therefore, building on recommendations from previous research, our approach advances DHS-based measures and definitions of empowerment by applying a multidimensional, context-specific framework informed by primary data, locally relevant covariates and a participatory design. It incorporates refined weighting of decision-making, asset ownership and access to technology (such as mobile phones), yielding a more accurate, meaningful and locally grounded assessment of women’s empowerment.⁹

According to this study, women whose husbands completed primary, secondary and higher education were associated factors of women’s empowerment. This finding was supported by studies from developing countries.¹⁹ The result shows that women whose husbands completed at least primary school had a greater impact on empowerment than those whose husbands were uneducated. This implies that education entails improved awareness and behavioural change between spouses. Accordingly, we hypothesise that an educated spouse is more likely to share family burdens, allow his wife to participate in decision-making and help her maintain gender-equitable norms. Similarly, studies across countries have found that an educated man may not be tied up to social and cultural norms as an uneducated man.⁶⁸

This study revealed a significant relationship between women’s empowerment and household wealth indexes. It suggests that women with middle and high-wealth household indexes were more likely to be empowered than those in the lowest wealth index. The result is corroborated by other findings from Ethiopia.^{36 69} The result suggests that women in the poorest households exhibit lower levels of empowerment than women in the richest households. Moreover, women and girls living in poverty are more vulnerable to sexual exploitation due to lack of income and resources. Poverty is a detrimental factor of well-being in women,⁷⁰ and a fundamental driver of gender inequality, in education, health and standard of living.⁶⁹

In the present study, mothers recognised as model households were more likely to be empowered than mothers who were not recognised as model households. A consistent finding was reported from previous studies in Ethiopia.^{71 72} This implied that model households may be agents of social transformation and mediators in solving the existing cultural and attitudinal bottlenecks. This finding is supported by previous studies, which indicate that model households benefited from improved access to information and short-term training on the health extension package.⁷³ The package includes components on maternal health, nutrition, women’s rights and gender equality. In addition, health extension workers educated women, one-to-one at home and in groups at health posts, on maternal health.^{71 72 74} Accordingly, we hypothesised that implementing health extension packages would be positively linked to increases in women’s empowerment.

The study found that women participating in PSNPs were more likely to be empowered than those not included in the programmes. Previous studies from Ethiopia,^{61 75} Egypt⁷⁵ and South Africa⁷⁶ corroborate the current finding. The conditional cash transfer (CCT) and the integrated nutrition social cash transfer (IN-SCT) mimic with the PSNP programme of the current study sites. Such support can enable financial independence of women, improve girls’ attendance in schools and enhance the bargaining power of women in decision-making.^{61 75 76}

The results confirm that an increase of 1 year in age at first pregnancy is an important contributor to women’s empowerment. The delay gives a chance to go to school and postpone early marriage. This finding was supported by studies from India,⁷⁷ Africa,⁶ and low and middle-income countries.⁹

Conversely, women who had a history of at least one adverse pregnancy outcome (abortion) were negatively associated with women’s empowerment. Previous literature found that women with strong religious beliefs had higher levels of self-judgement and greater perception of community condemnation than less religious women.⁷⁸ Tigrayan women are usually devout, and almost 98.8% are orthodox Christians, and the remaining 1.2% are Muslims or Catholics. We believe that spirituality may contribute to the disempowerment of women in this regard. In this context, previous study revealed that abortion among Tigray women can be viewed as an act of blasphemy or even as a source of social stigma, potentially relegating them to the status of ‘bad girl.⁷⁹

The current study also found that women who received a dowry from their family at marriage were more likely to be empowered. The finding is similar to other studies carried out in Ethiopia⁸⁰ and in several developing countries.⁸¹ A higher dowry enhanced women’s decision-making power and decreased the likelihood of women’s exposure to lethal domestic violence. On the contrary, dowry practices increased conflicts, leading to domestic violence, dowry deaths and marital instability in Bangladesh.⁸² The difference could be attributed to the cultural variation and the amount of dowry. For example, dowry in Ethiopia is a common and long-established cultural practice, particularly in the Tigray state.⁸³ A study done in Ethiopia showed that nearly 9 out of 10 married respondents received dowry payments. Dowry is considered a positive factor in how they are treated by their husbands, their families and the rest of their lives.⁸⁰ Tigray society has traditionally practiced dowry and there is a tendency to contract marriages between households of roughly commensurate wealth. On the other hand, dowry practice in Bangladesh seems to be based on competition for a groom of higher social, political, income or educational status, with huge amounts of money.⁸²

A positive relationship was observed between women’s empowerment and perceived readiness of good justice for girls and women in the study communities. This finding was corroborated by studies from Peru⁸⁴ and Nigeria.⁸⁵ Consistent with findings from previous studies,⁸⁶ we argue that effective implementation of the rule of law in society involves protecting women from recurrent violence and enabling them to exercise their agency within the community.

In this study, individual factors such as women’s education, occupation, age, family size and religion, previously identified as a key contributing factor to women’s empowerment,^{59 69} did not show a statistically significant association with women’s empowerment. Such differences highlight the importance of contextual factors between study populations and regional dynamics in shaping women’s empowerment trajectories. In addition, variations in the definition or measurement of women’s empowerment might account for the discrepancy.

Strengths and limitations

The study employed a multidimensional measurement indicator that considered individual, intrafamilial and community-level characteristics of empowerment. Emphasising the multidimensionality and dynamic nature of empowerment and collecting primary data in a specific context to incorporate additional dimensions of women’s empowerment (eg, decision-making, social independence, legal, political, leadership and gender norms), is advantageous for understanding changes in empowerment and its determinants.

Despite its strengths, this study has some limitations that must be acknowledged. The study’s limitations include the cross-sectional survey design, which prevents causal inference and the limited generalisability of the results due to the omission of non-pregnant women. Tigrayan society is patriarchal, which leads to women’s responses being influenced by higher social desirability bias. The survey did not assess spouses’ perspectives on women’s empowerment and its determinants, and responses to these questions from both wives and their husbands were missing.

Conclusions and recommendations

This study provides useful insights into women’s empowerment and its various determinants within the study area. The study found women’s empowerment to be very low but identified potential targets to enhance empowerment. For example, implementing community-based empowerment programmes, including enhancing justice for women and girls and gender-specific policies focused on increasing women’s involvement in decision-making could serve as effective interventions to improve women’s bargaining power, combat negative gender norms, narrow existing gender gaps and empower women.

The depiction of boundaries on this map does not imply the expression of any opinion whatsoever on the part of BMJ (or any member of its group) concerning the legal status of any country, territory, jurisdiction or area or of its authorities. This map is provided without any warranty of any kind, either express or implied.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.