Prevalence and determinants of ability to refuse sex among women of reproductive age in Tanzania: evidence from national representative surveys

Abstract

Background

Women’s ability to refuse unwanted sex is a fundamental dimension of gender equality and a key determinant of overall well-being. Despite its importance, there is limited evidence on the prevalence and factors influencing this ability in Tanzania. This study therefore aimed to investigate the prevalence and determinants of women’s ability to refuse sex among reproductive-age women in the country.

Method

This study analyzed secondary data from the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey (TDHS-MIS), using a weighted sample of 9,090 women aged 15–49. Descriptive statistics summarized socio-demographic characteristics, while binary logistic regression identified factors associated with women’s ability to refuse sex. Both unadjusted and adjusted odds ratios (ORs) with 95% confidence intervals were reported, with statistical significance set at p < 0.05.

Results

The study found that 69.6% of Tanzanian women of reproductive age had the ability to refuse sex (95% CI: 68.5%–70.4%). After adjusting for confounders, women’s ability to refuse sex was associated with their education, primary (aOR = 1.354, p < 0.001), secondary (aOR = 1.74, p < 0.001), higher (aOR = 2.253, p = 0.024), and their partners’ education, secondary (aOR = 1.269, p = 0.015), higher (aOR = 1.756, p = 0.009). Wealth (aOR = 1.241, p = 0.017), awareness of sexually transmitted infections (aOR = 1.368, p < 0.001), prior HIV testing (aOR = 1.714, p < 0.001), media access, low (aOR = 1.273, p < 0.001), medium (aOR = 1.612, p < 0.001), high (aOR = 1.954, p < 0.001), and contraceptive use, calendar-based (aOR = 1.311, p < 0.001) or other methods (aOR = 1.525, p < 0.001), were also positively associated. Women from Unguja (aOR = 0.425, p < 0.001) and Pemba (aOR = 0.376, p = 0.004) were less likely to refuse sex.

Conclusion

The study found that a majority of Tanzanian women have the ability to refuse sex and that the ability to refuse sex is significantly linked to education, economic status, partner’s education, access to media, STI awareness, HIV testing, and contraceptive use. However, women in Unguja and Pemba had lower ability to refuse sex compared to those in mainland urban areas. The findings highlight the need for targeted interventions to address regional disparities and promote equitable access to resources that support women’s ability to refuse sex.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12905-025-04203-2.

Background

Women’s ability to refuse sex is a critical aspect of sexual and reproductive health autonomy and a fundamental human right and cornerstone of gender equality [1]. The ability to refuse sex is a critical expression of bodily autonomy to safeguard women’s health and well-being [2, 3]. It is significantly associated with positive health outcomes, including reduced incidence of short birth intervals [4], reproductive tract infections [5], and sexually transmitted infections [6], as well as increased contraceptive use [7] and child survival rates [8]. Despite these benefits, many women globally continue to lack the power to refuse sex. According to the 2020 United Nations Population Fund (UNFPA) report, approximately 25% of women worldwide lack the ability to refuse sex, with the burden disproportionately affecting sub-Saharan Africa, where nearly 30% of women are unable to exercise this right [9].

In this study, sex refusal was used as a proxy measure for women’s ability to decline unwanted sexual activity. We defined unwanted sexual activity as instances where a woman does not desire sex but may still comply, often due to partner expectations, cultural norms, or relational obligations. This is distinct from non-consensual sex, which involves coercion, force, or lack of voluntary agreement.

While regional disparities are evident, for instance, Namibia reports 91.4% of women with ability to refuse sex compared to only 22.3% in Mali, the literature remains limited in explaining these differences with contextual rigor [10]. Previous studies have identified several socio-demographic, relational, and community-level determinants of women’s sex refusal. These include education levels [3], occupation [3], media exposure [11, 12], household wealth [12], marriage type [12], and community norms [12]. Some studies also link ability to refuse sex with the risk or experience of intimate partner violence (IPV), though findings are often inconsistent or contradictory. Some report that IPV limits sex refusal [12, 13], while others suggest that women with greater ability to refuse sex may also face increased IPV risk [14]. These inconsistencies highlight both the contextual complexity of the issue and the methodological limitations in previous research.

Many existing studies suffer from methodological flaws, such as a limited geographical focus [14, 15], inadequate sample sizes [15], and failure to account for contextual socio-cultural and legal frameworks [14]. Notably, research in Tanzania has largely focused on non-consensual sex (sexual violence)[16–19], rather than consensual but unwanted sex, a nuanced but crucial distinction. Furthermore, some studies do not differentiate between refusal of sex as a right and as a health-protective behavior [15], nor do they consider whether their findings are generalizable across diverse Tanzanian regions.

Moreover, the current situation in Tanzania is shaped by entrenched socio-cultural norms that discourage women, particularly married women, from exercising sexual autonomy [20, 21]. While Tanzania has made commendable progress in establishing legal and policy frameworks to protect women’s rights, enforcement remains weak due to prevailing patriarchal norms and limited institutional capacity. These challenges are compounded by economic dependencies, lack of awareness about sexual rights, and the fear of violence or abandonment, which further undermine women’s ability to refuse sex [22].

To better conceptualize the factors influencing women’s ability to refuse sex, this study is informed by the Theory of Gender and Power (TGP). This theory explains how gendered inequalities are structurally maintained through the sexual division of labor, the sexual division of power, and affective social norms [23]. In the Tanzanian context, these dimensions manifest in women’s limited economic independence, a male-dominated decision-making culture, and societal expectations that equate marital duty with sexual availability [24]. The TGP enables a multi-level understanding of women’s ability to refuse sex, positioning it not only as an individual behavior but also as an outcome shaped by broader socio-cultural, relational, and structural forces [23]. This theoretical lens supports the need for systemic interventions that go beyond individual-level empowerment to address much needed institutional and cultural transformations.

Despite growing global research on women’s sexual and reproductive health, limited evidence exists on the factors influencing women’s ability to refuse sex in Tanzania. Previous studies have largely focused on prevalence and individual-level determinants, with minimal attention to regional disparities and socio-cultural factors within the Tanzanian context. This study seeks to fill these knowledge gaps by examining the prevalence and determinants of women’s ability to refuse sex using data from the 2022 TDHS-MIS. By addressing methodological limitations of earlier studies and providing context-specific evidence, the findings aim to inform national policies and community-based programs that promote women’s reproductive rights, reduce gender-based coercion, and improve overall health and well-being. What distinguishes this study from prior research is its focus on national-level data, regional differences, and culturally relevant determinants, offering a more comprehensive understanding of women’s ability to refuse sex in Tanzania. The results are expected to benefit stakeholders in government, civil society, academia, and international development sectors committed to advancing gender equality and women’s empowerment in Tanzania and similar settings.

Methods

Study design

This study employed a cross-sectional analytical design using secondary data from the 2022 TDHS-MIS. The TDHS-MIS is part of the global DHS program and is designed to provide comprehensive data on health and population indicators. For this analysis, we focused on variables related to women’s ability to refuse sex, particularly the ability to refuse sex. The Tanzania context is characterized by persistent gender disparities, sociocultural norms that shape sexual decision-making, and a high burden of reproductive health challenges [24].

Study population

The study population consisted of women of reproductive age (15–49 years) who responded to the survey question on their ability to refuse sex. To ensure the estimates reflect current patterns, we restricted the analysis to data collected within the past five years. Women were included if they were aged 15–49 years, currently or previously partnered, and had responded to the relevant question on ability to refuse. We excluded respondents with missing or incomplete responses to this variable.

Study design, and sampling procedure

We used data from the 2022 TDHS-MIS, which employed a two-stage stratified sample design. A total of 629 clusters/enumeration areas were selected, in which 16,354 households were targeted (15,705 successfully interviewed). Among these, 15,699 women aged 15–49 were identified as eligible, and 15,254 completed the interviews with a response rate of 97%. Full details of the sampling design and methodology are available in the DHS report [25].

Data extraction

This study did not involve the development of a new data collection tool, as it relied on secondary data obtained from a nationally administered survey. We utilized the Individual Recode (TZIR82SV.ZIP) dataset to identify eligible participants. Among the 15,699 women interviewed, 15,254 were of reproductive age. Of these, 6,164 (40.4%) were excluded from the analysis due to missing information (6002) on the outcome variable or responses of “don’t know” (162). The final analytic sample, therefore, consisted of 9,090 women of reproductive age (Fig. 1).

Fig. 1.

Data extraction flow chart

Sample weighting

To account for the complex sampling design of the DHS and MIS surveys, all analyses were conducted using sampling weights provided in the datasets. The weights adjust for differences in probability of selection and response rates across clusters and strata, ensuring that the results are nationally representative. The individual women’s sample weight (v005) was applied after dividing by 1,000,000 as recommended by DHS guidelines.

Handling of missing data

We assessed patterns of missing data by generating binary indicators for missing versus observed responses and comparing them across key sociodemographic characteristics. Cross-tabulations and chi-square tests revealed that missing responses were significantly associated with residence, education, and wealth, indicating that the missingness was not completely at random. We further conducted logistic regression to identify predictors of missingness and applied multiple imputation using these variables to estimate missing values. Sensitivity analyses comparing complete-case and imputed datasets confirmed that the study findings were robust to the handling of missing data.

Data collection

The DHS–MIS data collection involved a multistage household-based survey using standardized questionnaires and biomarker assessments to obtain reliable demographic, health, and malaria indicators. Trained field teams conducted household listings and administered electronic questionnaires (CAPI) to collect data on fertility, maternal and child health, family planning, nutrition, and malaria prevention practices. Eligible women, men, and caregivers of children under five were interviewed after obtaining informed consent. Anthropometric measurements, hemoglobin testing using HemoCue, and malaria rapid diagnostic tests (RDTs) were performed on selected individuals, with proper biosafety measures in place, and referrals were made for positive cases. Data collection tools included structured questionnaires for household, women’s, men’s, and child health, calibrated anthropometric equipment, GPS devices, HemoCue analyzers, malaria RDT kits, and digital tablets, all supported by supervision checklists and daily data quality monitoring to ensure accuracy, confidentiality, and ethical compliance.

Study variables

We measured the response DHS-MIS item V850A (“ability to refuse sex”) in two ways: (i) cannot resist (V850A= “No”) and (ii) can resist (V850A= “Yes”). Women who responded “Yes” were assigned a score of one (1) and were considered to have ability to refuse sex, while those who responded “No” were given a score of zero (0) and were considered to lack this ability to refuse sex. The independent variables included age categories of women, the highest education level of both women and their husbands or partners, place of residence, zone of residence, wealth index, number of children ever born, access to media, history of pregnancy termination, current working status. Intermediate variables include awareness of sexually transmitted infections (STIs), history of HIV testing and use of any contraceptive method (Fig. 2).

Fig. 2.

Conceptual Framework

Harmonization and geographic units

Spatial analyses were carried out at the first administrative level (regions). The polygon shape layer (Arc 1960) was transformed and reprojected to WGS84/EPSG:4326. Region names were cleaned into a join key (region key) and reconciled with the regional labels of the survey based on V024. Since the shapefile represents pre-2016 boundaries (i.e., no Songwe), survey records with regional label Songwe were reassigned to Mbeya for spatial joining purposes since Songwe originated from Mbeya. Map region labels were located at point-on-surface coordinates calculated in UTM 36 S (EPSG:32736) and then plotted back in WGS84 with label-repulsion in order to steer clear of overlapping.

Regional prevalence

Weighted prevalence was calculated for each region r, as follows:

For each region , we computed weighted prevalence as

Where:

= weighted prevalence (proportion) in region .

= a specific region.

= an individual woman in region .

survey weight.

an indicator for the outcome mapped.

Spatial weights and autocorrelation

Adjacency between regions was defined by queen contiguity. Because Tanzania includes island regions and a few potential isolates, we used a minimal 1-nearest-neighbour (1-NN) bridge for isolates (based on region centroids) and then symmetrized the neighbor list. Global Moran’s I at 5% significance level was used to test overall spatial autocorrelation of regional prevalence. Local Moran’s I (LISA) and Getis Ord Gi* were used for the identification of hot and cold spots. This complementary approach enables capturing clear patterns of spatial clustering.

Software

We analyzed data in R software version 4.5.1 by applying the software sf, spdep, ggplot2.

Statistical analyses

Data were analyzed using SPSS version 25. Initially, descriptive statistics summarized the background characteristics of the women. Bivariate analyses employing chi-square tests examined associations between women’s ability to refuse sex and potential predictor variables. Subsequently, binary logistic regression was conducted to identify significant predictors of this ability to refuse sex. Variables with a p-value < 0.20 in the bivariate analysis were included in the multivariable logistic regression model. Statistical significance was set at p < 0.05.

Results

Socio-demographic characteristics of women of reproductive age in Tanzania

A total of 9,090 women of reproductive age participated in this study. The majority were between the ages of 20 and 34 (55.4%), resided in rural areas (68.8%), and were from Tanzania mainland rural regions (66.7%). Most had attained only a primary level of education (58.4%), had never used the internet (87.4%), and were sexually active (79.7%). A significant portion had heard about sexually transmitted infections (79.5%) and had been tested for HIV (92.7%). (Table 1)

Table 1.

Socio-demographic characteristics of women of reproductive age in Tanzania

Variables	Frequency (n)	Percent (%)
Age Categories
15–24 years	2135	23.5
25–34 years	3452	38
35–49 years	3504	38.5
Type of place of residence
Urban	2839	31.2
Rural	6252	68.8
Mainland/Zanzibar
Mainland urban	2752	30.3
Mainland rural	6067	66.7
Unguja (Zanzibar Island)	198	2.2
Pemba (Pemba Island)	74	0.8
Highest educational level
No education	1836	20.2
Primary	5312	58.4
Secondary	1836	20.2
Higher	106	1.2
Wealth index
Poorest	1672	18.4
Poorer	1686	18.5
Middle	1739	19.1
Richer	1940	21.4
Richest	2054	22.6
Access to Media
None	2990	32.9
Low Access	2920	32.1
Medium Access	2161	23.8
High Access	1019	11.2
Number of children ever born
No children	579	6.4
One to four	5931	65.2
Five and above	2580	28.4
Ever had a terminated pregnancy
No	7458	82
Yes	1632	18
Husband/partner’s education level
No education, preschool/early childhood education	1121	12.3
Primary	5701	62.7
Secondary	1932	21.3
Higher	287	3.2
Don’t know	48	0.5
Respondent currently working
No	3309	36.4
Yes	5781	63.6
Ever heard of a Sexually Transmitted Infection (STI)
No	1865	20.5
Yes	7226	79.5
Ever been tested for HIV
No	665	7.3
Yes	8425	92.7
Ever used anything or tried to delay or avoid getting pregnant
No	3595	39.6
Yes, used outside calendar	595	6.5
Yes, used in calendar	4900	53.9

Prevalence of women’s ability to refuse sex

A total of 2760 (30.4%) of interviewed women did not have the ability to refuse sex, while a total of 6330 (69.6%) of interviewed women had the ability to refuse sex. Weighted maps exhibited apparent regional variation in prevalence of all outcomes (autonomy and lack of autonomy). The prevalence of lack of autonomy to resist unwanted sex ranged from 16.31% in Singida to 57.98% in Simiyu as indicated in Fig. 3 and supplementary table (S1). Regarding autonomy to resist unwanted sex, the prevalence ranged from 42.02% in Simiyu to 83.69% in Singida as shown on Fig. 4 and Supplementary Table 2.

Fig. 3.

Map showing prevalence of lack of autonomy to refuse sex

Fig. 4.

Map showing prevalence of ability to refuse sex

Global Moran’s I did not provide a demonstration of positive country-wide clustering for both of the outcomes. It thus suggests spatial dependence is localized and not global. Mara region was a Gi*-hot-spot of lack of autonomy and also a cold-spot of autonomy (Fig. 5, Fig. 6, Fig. 7, Fig. 8). Some regions (e.g., Kaskazini Pemba and Singida) had localized deviations but didn’t reach Gi*-significance indicating that they are spatial outliers rather than reliable hotspots. The supplementary Table (S3) includes an exhaustive tabulation of regional findings, prevalences, LISA quadrant assignments, and Gi* classifications of both measured outcomes.

Fig. 5 .

Areal Getis-Ord Gi* for Regional Prevalence of Autonomy to Refuse Unwanted Sex

Fig. 6.

LISA for Regional Prevalence of Autonomy to Refuse Unwanted Sex

Fig. 7 .

Areal Getis-Ord Gi* For Regional Prevalence of Lack of Autonomy to Refuse Unwanted Sex

Fig. 8 .

.LISA For Regional Prevalence of Autonomy to Refuse Unwanted Sex

The relationship between women’s characteristics an ability to refuse sex

Variables demonstrating a significant relationship with women’s ability to refuse sex included age categories (χ²=12.629, p = 0.002), type of residence (χ²=204.008, p < 0.001), zone of residence (χ²=216.971, p < 0.001), highest level of education (χ²=332.854, p < 0.001), wealth index (χ²=373.294, p < 0.001), internet use (χ²=141.451,

p < 0.001), number of children ever born (χ²=46.538, p < 0.001), history of terminated pregnancy (χ²=8.766, p = 0.003), recent sexual activity (χ²=20.994, p < 0.001), husband’s education level (χ²=235.357, p < 0.001), current employment status (χ²=24.442, p < 0.001), awareness of sexually transmitted infections (χ²=180.066, p < 0.001), HIV testing history (χ²=157.102, p < 0.001), and the use of methods to delay or prevent pregnancy (χ²=176.504, p < 0.001). (Table 2)

Table 2.

The Relationship between women’s characteristics and ability to refuse sex

Variable	No ability to refuse sex n (%)	Have ability to refuse sex n (%)	χ2	p-value
Age Categories			12.629	0.002
15–24 years	704(33)	1431(67)
25–34 years	984(28.5)	2468(71.5)
35–49 years	1073(30.6)	2431(69.4)
Type of place of residence			204.008	<0.001
Urban	572(20.1)	2267(79.9)
Rural	2189(35)	4063(65)
Mainland/Zanzibar			216.971	<0.001
Mainland urban	542(19.7)	2210(80.3)
Mainland rural	2118(34.9)	3949(65.1)
Unguja (Zanzibar Island)	67(33)	132(66.3)
Pemba (Pemba Island)	34(45.9)	40(54.1)
Highest educational level			332.854	<0.001
No education	839(45.7)	997(54.3)
Primary	1555(29.3)	3756(70.7)
Secondary	355(19.3)	1480(80.7)
Higher	11(10.3)	96(89.7)
Wealth index			373.294	<0.001
Poorest	740(44.3)	932(55.7)
Poorer	638(37.8)	1048(62.2)
Middle	531(30.5)	1208(69.5)
Richer	480(24.7)	1460(75.3)
Richest	371(18.1)	1683(81.9)
Access to Media		396.981a	396.981a	<0.001
None	1265(42.3)	1725(57.7)
Low Access	883(30.2)	2037(69.8)
Medium Access	455(21.1)	1706(78.9)
High Access	158(15.5)	862(84.5)
Number of children ever born			46.538	<0.001
No children	201(34.7)	378(65.3)
One to four	1658(28)	4273(72)
Five and above	900(34.9)	1679()65.1
Ever had a terminated pregnancy			8.766	0.003
No	2314(31)	5143(69)
Yes	446(27.3)	1187(72.7)
Husband/partner's education level			235.357	<0.001
No education, preschool/early childhood education	512(45.7)	609(54.3)
Primary	1777(31.2)	3925(68.8)
Secondary	425(22)	1507(78)
Higher	35(12.2)	252(87.8)
Don't know	12(25)	36(75)
Respondent currently working			24.442	<0.001
No	1109(33.5)	2200(66.5)
Yes	1651(28.6)	4130(71.4)
Ever heard of a Sexually Transmitted Infection (STI)			180.066	<0.001
No	804(43.1)	1061(72.9)
Yes	1957(27.1)	5269(69.6)
Ever been tested for HIV			157.102	<0.001
No	345(51.9)	320(48.1)
Yes	2415(28.7)	6010(71.3)
Ever used anything or tried to delay or avoid getting pregnant			176.504	<0.001
No	1376(38.3)	2219(61.7)
Yes, used outside calendar	143(24)	452(76)
Yes, used in calendar	1241(25.3)	3659(74.7)

Factors associated with women’s ability to refuse sex

After controlling for confounders, several factors were significantly associated with women’s ability to refuse sex. Women with primary (aOR = 1.354, 95% CI: 1.198–1.530, p < 0.001), secondary (aOR = 1.74, 95% CI: 1.445–2.094, p < 0.001), and higher education (aOR = 2.253, 95% CI: 1.116–4.551, p = 0.024) were more likely to refuse sex compared to those with no formal education. Wealthier women (aOR = 1.241, 95% CI: 1.04–1.481, p = 0.017) and women whose partners had secondary (aOR = 1.269, 95% CI: 1.047–1.539, p = 0.015) or higher education (aOR = 1.756, 95% CI: 1.151–2.680, p = 0.009) also showed higher ability to refuse sex. Awareness of sexually transmitted infections (aOR = 1.368, 95% CI: 1.22–1.534, p < 0.001), prior HIV testing (aOR = 1.714, 95% CI: 1.437–2.044, p < 0.001), and access to media—low (aOR = 1.273, 95% CI: 1.153–1.430, p < 0.001), medium (aOR = 1.612, 95% CI: 1.386–1.876, p < 0.001), and high (aOR = 1.954, 95% CI: 1.564–2.442, p < 0.001)—were strongly associated with increased ability to refuse sex. Contraceptive use, both calendar-based (aOR = 1.311, 95% CI: 1.181–1.457, p < 0.001) and other methods (aOR = 1.525, 95% CI: 1.237–1.882, p < 0.001), further increased this ability. Conversely, women from Unguja (aOR = 0.425, 95% CI: 0.261–0.691, p < 0.001) and Pemba (aOR = 0.376, 95% CI: 0.194–0.729, p = 0.004) were less likely to refuse sex compared to those from Tanzania mainland urban areas (Table 3).

Table 3.

Factors associated with women’s ability to refuse sex

Variables	OR(95% CI)	p-value	aOR(95% CI)	p-value
Age Categories
15–24 years	Ref.		Ref.
24–34 years	1.234(1.098–1.386)	< 0.001	0.977(0.857–1.113)	0.723
35–49 years	1.114(0.993–1.250)	0.067	1.022(0.877–1.191)	0.777
Type of place of residence
Urban	Ref.		Ref.
Rural	0.468(0.421–0.520)	< 0.001	1.057(1.057–1.846)	0.847
Highest educational level
No education	Ref.		Ref.
Primary	2.031(1.821–2.266)	< 0.001	1.354(1.198–1.530)	< 0.001
Secondary	3.503(3.022–4.061)	< 0.001	1.74(1.445–2.094)	< 0.001
Higher	7.646(4.018–14.548)	< 0.001	2.253(1.116–4.551)	0.024
Wealth index
Poorest	Ref.		Ref.
Poorer	1.303(1.136–1.496)	< 0.001	1.025(0.886–1.185)	0.744
Middle	1.808(1.571–2.080)	< 0.001	1.166(0.998–1.363)	0.053
Richer	2.416(2.097–2.782)	< 0.001	1.241(1.04–1.481)	0.017
Richest	3.599(3.103–4.174)	< 0.001	1.229(0.977–1.545)	0.078
Access to Media
None	Ref.		Ref.
Low Access	1.691(1.519–1.882)	< 0.001	1.273(1.153–1.430)	< 0.001
Medium Access	2.748(2.748–3.118)	< 0.001	1.612(1.386–1.876)	< 0.001
High Access	4.008(3.333–4.821)	< 0.001	1.954(1.564–2.442)	< 0.001
Number of children ever born
No children	1.006(0.833–1.216)	0.949	1.067(0.842–1.352)	0.59
One to four	1.382(1.252–1.525)	< 0.001	1.023(0.902–1.161)	0.726
Five and above	Ref.		Ref.
Ever had a terminated pregnancy
No	Ref.		Ref.
Yes	1.198(1.063–1.350)	0.003	1.073(0.945–1.218)	0.277
Husband/partner’s education level
No education, preschool/early childhood education	Ref.		Ref.
Primary	1.855(1.629–2.113)	< 0.001	1.149(0.995–1.327)	0.059
Secondary	2.979(2.540–3.494)	< 0.001	1.269(1.047–1.539)	0.015
Higher	6.113(4.206–8.886)	< 0.001	1.756(1.151–2.680)	0.009
Don’t know	2.584(1.325–5.040)	0.005	1.696(0.837–3.435)	0.143
Respondent currently working
No	Ref.		Ref.
Yes	1.262(1.151–1.383)	< 0.001	1.097(0.993–1.213)	0.069
Ever heard of a Sexually Transmitted Infection (STI)
No	Ref.		Ref.
Yes	2.04(1.836–2.266)	< 0.001	1.368(1.22–1.534)	< 0.001
Ever been tested for HIV
No	Ref.		Ref.
Yes	2.682(2.287–3.145)	< 0.001	1.714(1.437–2.044)	< 0.001
Mainland/Zanzibar
Mainland urban	Ref.		Ref.
Mainland rural	0.457(0.410–0.509)	< 0.001	0.781(0.442–1.379)	0.394
Unguja (Zanzibar Island)	0.484(0.355–0.659)	< 0.001	0.425(0.261–0.691)	< 0.001
Pemba (Pemba Island)	0.292(0.183–0.467)	< 0.001	0.376(0.194–0.729)	0.004
Ever used anything or tried to delay or avoid getting pregnant
No	Ref.		Ref.
Yes, used outside calendar	1.963(1.607–2.397)	< 0.001	1.525(1.237–1.882)	< 0.001
Yes, used in calendar	1.829(1.667–2.008)	< 0.001	1.311(1.181–1.457)	< 0.001

Discussion

This study revealed that nearly one in three Tanzanian women of reproductive age lack the ability to refuse to refuse sex. Comparable studies from Ethiopia [26] and sub-Saharan Africa [10] reported similar levels of a diminished ability to refuse sex, highlighting that this problem extends beyond national borders. The persistence of such high levels across different contexts suggests deep-rooted socio-cultural norms that undermine women’s agency in sexual matters[27]. In Tanzania, these patterns may be influenced by traditional gender roles that prioritize male sexual entitlement and discourage women from asserting their preferences [28]. From a policy perspective, these findings call for stronger national strategies that integrate the ability to refuse sex into broader gender equality and sexual and reproductive health (SRH) programs.

Our analysis demonstrated that women with primary, secondary, and higher education were more likely to have the ability to refuse sex compared to women with no formal education. This finding is consistent with evidence from Ethiopia [26] and a multi-country study across 30 sub-Saharan African nations[10], which identified education as a powerful enabler of women’s ability to refuse sex. Education likely enhances knowledge of rights, decision-making capacity, and self-efficacy, thereby enabling women to more effectively negotiate intimate relationships. The implication is that ensuring equitable access to quality education for girls is a transformative long-term strategy for strengthening women’s ability to refuse sex in Tanzania.

The study also found that the education level of male partners was positively associated with women’s ability to refuse sex. This aligns with findings from a population-based study conducted in 30 sub-Saharan African countries [10] and research from India [29], which demonstrate that women living in more literate households are more likely to exercise a greater ability to refuse sex. Educated male partners may be more receptive to egalitarian norms and less likely to uphold traditional masculinity that pressures women into compliance. These findings point to the importance of engaging men and boys in educational and gender-transformative programs that challenge harmful norms and promote respect for women’s rights.

Women in higher wealth quintiles were more likely to report the ability to refuse sex in sexual decision-making compared to those in lower quintiles. Similar patterns have been observed in studies across Africa, where economic empowerment reduces dependency and expands women’s negotiating power within relationships [30, 31]. In the Tanzanian context, limited economic opportunities, especially in rural areas, may exacerbate women’s vulnerability to male dominance in sexual relations. This finding highlights the need for policies that strengthen women’s economic empowerment through access to credit, employment opportunities, and livelihood programs as a pathway to enhancing their ability to refuse sex.

Awareness of sexually transmitted infections, prior HIV testing, and contraceptive use were all significantly associated with greater ability to refuse sex. These findings are consistent with earlier studies that link health awareness and service utilization to improved agency in sexual and reproductive decision-making[32]. In Tanzania, women who access HIV testing or family planning services often receive counseling that increases knowledge, risk perception, and self-confidence. These results emphasize the importance of integrating ability to refuse sex into sexual and reproductive health services, particularly counseling and health education programs, to strengthen women’s capacity to make informed decisions.

Furthermore, the study revealed that women with media exposure were significantly more likely to demonstrate the ability to refuse sex compared to their counterparts without such exposure. Access to media platforms, including television, radio, and newspapers, appeared to enhance women’s awareness of their rights and sexual health, thereby equipping them with the confidence to assertively reject unwanted sexual advances. Media exposure likely served as a vital channel for disseminating information on gender equality, consent, and health-related decision-making, fostering both autonomy and self-efficacy. These findings highlight the transformative role of media in advancing women’s empowerment and underscore the need to strategically leverage media as a powerful tool for promoting sexual and reproductive health in Tanzania and globally.

A novel contribution of this study was the identification of regional disparities. Women residing in Unguja and Pemba (within the Zanzibar archipelago) were markedly less likely to report the ability to refuse sex compared to those in mainland urban areas. Previous research suggests that socio-cultural and religious traditions may reinforce male dominance and discourage women from refusing sex [33]. In contrast, while rural mainland women also showed reduced ability to refuse sex, the association was not statistically significant, suggesting that local belief systems may exert a stronger influence in Zanzibar. The policy implication is that national interventions must be sensitive to regional variations, with context-specific approaches tailored to Zanzibar’s socio-cultural setting.

This study is not without limitations. First, the cross-sectional design limits the ability to establish causal relationships; however, the use of regression modeling enhanced the robustness of the findings by adjusting for potential confounders. Second, underreporting of sensitive issues cannot be ruled out, given the cultural taboos surrounding sexuality. Third, although missed data were present, their significant association with observed variables indicates a potential source of bias. Fourth, the analysis included all women regardless of recent sexual activity, which may have introduced some bias, particularly from menopausal, breastfeeding, or postpartum abstinent participants. Finally, future qualitative research is warranted to provide deeper insights into the cultural, religious, and relational factors shaping women’s ability to refuse sex across different regions of Tanzania.

Conclusion

The study revealed that a substantial proportion of Tanzanian women can refuse sex. This ability to refuse sex is significantly associated with educational attainment, economic status, a partner’s [13] education level, awareness of sexually transmitted infections, HIV testing, and contraceptive use. However, regional disparities were noted, with women from Unguja and Pemba having lower ability to refuse sex compared to those from mainland urban areas. To enhance women’s ability to refuse sex, it is recommended to prioritize educational initiatives, improve economic opportunities, and promote awareness of sexual health. Additionally, targeted interventions should address regional disparities, ensuring that women in all areas, including Unguja and Pemba, have equal access to resources and support systems that empower them to exercise their ability to refuse sex.

Abbreviations

AOR

Adjusted Odds Ratio

CI

Confidence Interval

DHS

Demographic and Health Survey

HIV

Human Immunodeficiency Virus

IPV

Intimate Partner Violence

OR

Odds Ratio

STIs

Sexually Transmitted Infections

TDHS-MIS

Tanzania Demographic and Health Survey and Malaria Indicator Survey

UNFPA

United Nations Population Fund

Authors’ contributions

FVM led the conceptualization process, authored the methodology section, acquired the data, and conducted the data analysis, and initiated the manuscript drafting. RB wrote the introduction and critically reviewed the manuscript. All authors actively participated in the review process and ultimately approved the final version of the manuscript.

Funding

It was not a funded project; no specific funding to be declared.

Data availability

The dataset used is available upon request from DHS-Measure.

Declarations

Ethics approval and consent to participate

This study used secondary data without involving any human subjects. Hence, no formal ethical approval was required. However, the permission to use the data was obtained from the DHS program. The procedures for TDHS-MIS data collection were approved by the following organizations: Tanzania’s National Institute for Medical Research (NIMR), the Zanzibar Medical Ethics and Research Committee (ZAMREC), the Institutional Review Board of ICF International, and the Centers for Disease Control and Prevention in Atlanta, USA. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.