Statistical Modeling of Factors Affecting the Length of Time for Water Collection among Rural Women in Northern Uganda

Statistical Modeling of Factors Affecting the Length of Time for Water Collection among Rural Women in Northern Uganda

Nelson Kisubi., Babalola Bayowa Teniola and Nakafeero Doreen

Department of Mathematics and Statistics, School of Mathematics and Computing, Kampala International University, Uganda

DOI: https://dx.doi.org/10.47772/IJRISS.2025.9020337

Received: 12 February 2025; Accepted: 18 February 2025; Published: 22 March 2025

ABSTRACT

The study aimed to model the factors affecting the length of time for water collection among rural women in Northern Uganda. The specific objectives included determining the socio-demographic, economic, and community-level factors associated with water collection time, developing a binary logistic model to estimate these factors, and assessing the model’s goodness of fit. The research utilized secondary data from the 2016 Uganda Demographic and Health Survey (UDHS), administered by the Uganda Bureau of Statistics (UBOS). It focused on analyzing data from 3,483 rural women aged 15 to 49 years who participated in the survey. The analysis was conducted at three levels: univariate, bivariate, and multivariate. The study found that a woman’s age, sub-region, and household size were significant socio-demographic factors. The wealth quintile was the only economic factor associated with water collection time, while none of the community-level factors studied showed an impact. In conclusion, the risks of prolonged water collection time (exceeding 30 minutes) increased among older women compared to younger ones, among women from the Karamoja sub-region compared to those from the Lango, Acholi, and West Nile sub-regions, among households with larger family sizes, and among women from the poorest backgrounds. Therefore, targeted interventions should be implemented to assist older women, particularly those in the Karamoja sub-region and larger households, in reducing the time spent on water collection.

Keywords: Rural women; Time for water collection; Northern Uganda, Modeling.

INTRODUCTION

Globally, the duration spent on water collection remains considerable, particularly in regions where access to on-premises water supplies is limited. In 2015, approximately 26.3% of the global population lacked such access (Cassivi et al., 2018). It is estimated that approximately 135 million households globally are experiencing restricted access to water services and have to spend over 30 minutes on a round trip for water collection (Amankwaa et al., 2024; UNICEF/WHO, 2019). In India, the average time spent by women collecting and storing water ranges from 1 to 2 hours daily. In rural and peri-urban areas of India, where access to piped water is limited, adult women often bear the responsibility of water collection for their households. This task typically involves walking to nearby water sources, such as wells or community taps, and carrying containers filled with water back to their homes. The time spent on this daily chore can vary depending on factors such as distance to the water source, availability of water, and the number of trips required to meet the household’s water needs (Ferrant & Thim, 2019; Fletcher et al., 2017).

In developing countries, studies have suggested that extended time spent collecting water has adverse effects on women’s employment, health, and the well-being of both themselves and their children. In addition to these concerns, scholars have linked insufficient access to water in households with higher rates of short-term illnesses such as fever, cough, and diarrhea, as well as with challenges in reallocating domestic responsibilities, which may hinder children’s ability to attend school (Sedai, 2021). In sub-Saharan Africa, one roundtrip to collect water is 33 minutes on average in rural areas and 25 minutes in urban areas. In Asia, the numbers are 21 minutes and 19 minutes respectively. When water is not piped to the home the burden of fetching it falls disproportionately on women and children, especially girls. A study of 24 sub-Saharan countries revealed that when the collection time is more than 30 minutes, an estimated 3.36 million children and 13.54 million adult females were responsible for water collection. In Malawi, the UN estimated that women who collected water spent 54 minutes on average, while men spent only 6 minutes. For women, the opportunity costs of collecting water are high, with far-reaching effects. It considerably shortens the time they have available to spend with their families, on child care, other household tasks, or even in leisure activities (UNICEF, 2016).

In Uganda numerous households, particularly those residing in rural areas, devote over 30 minutes to walk and collect the water necessary for their families, thereby detracting from productive activities such as work and schooling. The national average time spent on water collection in Uganda stands at nearly 44 minutes (Cassivi et al., 2018). The arduous task of water collection predominantly falls upon women and children, particularly in rural communities where access to clean water is limited. Women, as primary caregivers, often shoulder the responsibility of ensuring their families have an adequate water supply for daily needs. Consequently, they spend considerable time each day trekking to and from distant water sources (Cassivi et al., 2018). In Northern Uganda, women typically spend over 30 minutes each day collecting water from distant, unsafe sources. This daily task not only consumes a significant portion of their time but also exposes them to various health risks associated with unsafe water. The burden of water collection often falls on women and girls, impacting their ability to engage in other productive activities such as education, employment, and family care. Spending excessive time on water collection reduces opportunities for income-generating activities, perpetuating the cycle of poverty (Mugumya et al., 2017; Mugumya et al., 2020).  However, there are limited studies in Uganda that have modeled the factors associated with extended length of time for water collection among rural women in Northern Uganda. Thus, need for this study.

RESEARCH DATA AND METHODS

Data sources

The study uses the cross-sectional secondary data from the Uganda Demographic and Health Survey (UDHS) 2016, administered by the Uganda Bureau of Statistics (UBOS). The UDHS survey divided the country into 112 districts, which were further grouped into 15 sub-regions for this study. A total of 20,880 households were selected for the survey, resulting in successful interviews with 18,506 women. Eligible participants included all women aged between 15 and 49 who were permanent residents of the selected households or visitors who had stayed in the household the night before the survey. Specifically, the study analyzed data from 3,483 rural females from Northern Uganda who participated in the 2016 UDHS survey and fell within the age range of 15 to 49 years.

Measures

Outcome variable

The outcome variable for this study was the length of time for water collection, which was measured in nominal form. A value of 1 denoted water collection time exceeding 30 minutes, while a value of 0 indicated water collection time of 30 minutes or less.

Explanatory variables

The explanatory variables comprised socio-demographic, economic, and community-level factors. Socio-demographic factors consisted of age (count), number of household members (count), religion (nominal), sub-region (nominal), marital status (nominal), household head (nominal), and education level (ordinal). The economic factors encompassed occupation (nominal) and wealth quintile (ordinal). The community-level factors included access to bicycles (nominal), access to motorcycles (nominal), access to cars/trucks, access to the radio (nominal), access to mobile phone (nominal), access to TV (nominal), and frequency of reading newspapers (ordinal).

Statistical analysis

The data analysis was conducted at three levels: univariate, bivariate, and multivariate. Univariate analysis provided descriptive statistics (frequencies and percentages) to summarize individual variables. Bivariate analysis explored relationships between pairs of variables using Chi-square tests identifying significant associations. Multivariate analysis employed techniques such as binary logistic regression to adjust for confounding factors and identify independent predictors of water collection time. This comprehensive approach ensured a thorough examination of the factors influencing the length of time required for water collection among rural women in Northern Uganda. The model underwent diagnosis at two stages: (i) examining for multicollinearity and (ii) evaluating the model’s goodness of fit. Through examining multicollinearity and evaluating the goodness of fit, the diagnostic procedures ensured that the logistic regression model was both reliable and robust providing valid insights into the determinants of water collection time.

RESULTS

Sample characteristics

Table 1 shows the characteristics of 3,483 women involved in this analysis. The majority of rural women from Northern Uganda were aged between 15 and 24 years (43.4%), while the minority were aged 45 years and above. Most were Catholic (61%), followed by Anglican (23.2%), and the least were Baptist (0.03%). A significant proportion of rural women in Northern Uganda possessed primary education (66.7%), while few had higher education (2.7%). Nearly 30% of the surveyed rural women were from Lango, followed by West Nile (29.7%), Acholi (22.8%), and Karamoja (17.5%). Most were married (41.4%), with the least being divorced (0.7%). A larger proportion of households were headed by males (67.3%), and 32.7% were headed by females. Most households consisted of 5 to 9 members (62.2%), while the least had 15 members and above (0.98%). The majority of the surveyed rural women were working (84.4%), and only 15.7% were not working. Most respondents were in the poorest wealth quintile (56.5%), while a few were in the richest wealth quintile (2.9%). Most households did not have a bicycle (55.7%), motorcycle (90.2%), or car (99.2%). In addition, the majority of homes did not have a radio (61.2%), television (97.9%), or mobile phone (81.9%). The findings also reveal that nearly 92.1% of respondents did not read newspapers, while a few read less than once a week (5.7%) or at least once a week (2.2%).

Table 1. Distribution of respondents by socio-demographic, economic, and community-level characteristics

Bivariate analysis of length of time for water collection by socio-demographic, economic, and community-level factors

Table 2 presents a bivariate analysis using the chi-square test to identify the factors that varied with the length of time for water collection. The length of time for water collection varied significantly with age (χ²=8.65, p=0.034), religion (χ²=36.57, p=0.000), education level (χ²=30.23, p=0.000), sub-regions (χ²=145.37, p=0.000), household size (χ²=10.75, p=0.013), wealth quintile (χ²=18.86, p=0.001), access to motorcycle (χ²=5.35, p=0.021), access to mobile phone (χ²=10.32, p=0.001), and frequency of reading newspapers (χ²=11.43, p=0.003) at 0.05 level of significance. However, the length of time for water collection was not associated with marital status, sex of household head, occupation, access to a bicycle, access to a car, access to a radio, and access to a television at a bivariate level.

Table 2. Bivariate analysis of length of time for water collection by socio-demographic, economic, and community-level factors

Multivariate analysis on factors associated with length of time for water collection among rural women in Northern Uganda

Table 3 presents the socio-demographic, economic, and community-level variables fitted in the multivariate binary logistic regression to determine the factors associated with the length of time for water collection among rural women in Northern Uganda. A binary logistic regression was constructed based on the significant predictors identified at the bivariate level to determine the effect of socio-demographic, economic, and community-level factors on the length of time required for water collection among rural women in Northern Uganda. All predictors/independent variables that were significant in the bivariate analysis were selected for multivariate analysis. These independent variables included age, religion, education level, sub-region, household size, wealth quintile, access to a motorcycle, access to a mobile phone, and frequency of reading newspapers. The model findings show that the odds of spending more time collecting water were higher among rural women aged 45 years and above (OR = 1.366, 95% CI: 1.016-1.836, p = 0.039) compared with those aged between 15 and 24 years. There were reduced chances of spending more time collecting water among rural women from Lango (OR=0.777, 95% CI: 0.612-0.986, P=0.038), Acholi (OR=0.368, 95% CI: 0.289-0.469, P=0.000), and West Nile (OR=0.386, 95% CI: 0.305-0.488, P=0.000) compared with women from the Karamoja sub-region. The odds of spending more time collecting water were higher among households with 10 to 14 members (OR = 1.630, 95% CI: 1.236-2.148, p = 0.001) compared with households with 1 to 4 members. Women from the richest wealth quintile (OR = 0.547, 95% CI: 0.331-0.905, p = 0.019) had a lower likelihood of spending more time collecting water compared with women from the poorest wealth quintile.

Table 3. Binary logistic regression findings showing the socio-demographic, economic, and community-level factors associated with the length of time for water collection

DISCUSSION

This study examined the question, “What socio-demographic, economic, and community-level factors are associated with the length of time for water collection among rural women in Northern Uganda?” Accordingly, the study tested the hypotheses: socio-demographic factors are significantly associated with the length of time for water collection among rural women in Northern Uganda; economic factors are significantly associated with the length of time for water collection among rural women in Northern Uganda; and community-level factors are significantly associated with the length of time for water collection among rural women in Northern Uganda.

The study found that socio-demographic factors associated with the length of time for water collection among rural women in Northern Uganda included age, sub-region, and household size. Specifically, older women (aged 45 years and above) were found to spend more time collecting water compared to younger women (aged 15 to 24 years) in rural areas of Northern Uganda. This may imply that older women in rural areas of northern Uganda may experience physical limitations such as reduced mobility or strength, making it more difficult for them to fetch water quickly. This can result in longer time spent walking to and from water sources or carrying heavy containers. The findings are consistent with those of Dongzagla et al. (2020), who found in Ghana that older individuals experienced longer water collection times (more than 30 minutes) compared to younger individuals. However, the findings contradict those of UBOS (2019), which reported that younger women spent more time collecting water compared to older women in Uganda.

The study revealed that rural women from Lango, Acholi, and West Nile had lower likelihoods of spending more time collecting water compared to women from the Karamoja sub-region. The findings may imply that women in Lango, Acholi, and West Nile regions may have easier access to water sources, such as boreholes or wells, located closer to their communities. This proximity reduces the time needed to collect water compared to women in Karamoja, where water sources may be more distant. The findings are consistent with those of UBOS (2021), which reported that women from the Karamoja sub-region typically walk longer distances to collect water, resulting in prolonged water collection times compared to women from other sub-regions in northern Uganda. The findings revealed that households with more members (10 to 14 members) had higher odds of spending more time collecting water compared to households with fewer members (1 to 4 members). This implies that larger households typically have higher water demands due to more individuals needing water for drinking, cooking, bathing, and other household chores. This increased demand necessitates more frequent trips to water sources and larger quantities of water to be collected, resulting in longer collection times. The findings are contrary to those of Boone et al. (2011), who found that having additional household members actually reduces an individual’s time spent on water collection.

The study identified the wealth quintile as the key economic factor influencing the time spent on water collection among rural women in Northern Uganda. The women from the richest wealth quintile were found to spend less time collecting water compared to those from the poorest wealth quintile. This may imply that women from wealthier households in rural areas in northern Uganda are more likely to have access to improved water sources, such as private wells or piped water, which are typically closer to their homes. This reduces the distance and time required to collect water. The findings corroborate those of Amankwaa et al. (2024) and Zozmann et al. (2022), who similarly observed that individuals hailing from households with affluent backgrounds, categorized as rich or richest, spent less time collecting water compared to their counterparts from economically disadvantaged backgrounds, classified as poor or poorest. However, the study revealed that certain community-level factors, including access to motorcycles, ownership of mobile phones, and frequency of reading newspapers, did not correlate with the duration of water collection among rural women in Northern Uganda.

Limitations

The research utilized secondary data extracted from the 2016 UDHS survey. However, it’s important to note that socio-demographic, economic, and community-level factors might have evolved since then. Consequently, the findings may not accurately reflect the current association between variables at the time of the study. Changes in demographics, social dynamics, or other factors could affect the generalizability of the 2016 UDHS survey data to the current population of interest.

CONCLUSIONS

The study concludes that women’s age, sub-regions, household size, and wealth quintile were the factors associated with the length of time for water collection among rural women in Northern Uganda. The risks of prolonged water collection time (exceeding 30 minutes) were found to increase among older women compared to younger ones, among women from the Karamoja sub-region compared to those from the Lango, Acholi, and West Nile sub-regions, among households with larger family sizes compared to smaller ones, and among women from the poorest backgrounds compared to those from the wealthiest ones. Therefore, there is a need to address these factors associated with prolonged water collection time among rural women in Northern Uganda.

ACKNOWLEDGMENTS

The authors express their gratitude to the DHS Program and ICF International for granting access to the data used in this analysis.

Disclosure statement

No potential conflict of interest was reported by the author(s).

FUNDING

This study received no specific grant from any funding agency, commercial entity, or not-for-profit organization.

REFERENCES

1. Amankwaa, G., Abrefa Busia, K., Agbadi, P., Duah, H. O., & Arthur-Holmes, F. (2024). Access heterogeneities and collection time inequalities of drinking water sources in Ghana: implications for water and development policy. Local Environment, 29(3), 263-278. https://doi.org/10.1080/13549839.2024.2306596
2. Boone, C., Glick, P., & Sahn, D. E. (2011). Household water supply choice and time allocated to water collection: Evidence from Madagascar. Journal of Development Studies, 47(12), 1826-1850.
3. Cassivi, A., Johnston, R., Waygood, E. O. D., & Dorea, C. C. (2018). Access to drinking water: time matters. Journal of Water and Health. https://www.pseau.org/outils/ouvrages/iwa oms access to drinking water time matters 2018.pdf
4. Dongzagla, A., Nunbogu, A. M., & Fielmua, N. (2020). Does self-reported water collection time differ from observed water collection time? Evidence from the Upper West Region of Ghana. Journal of Water, Sanitation and Hygiene for Development, 10(2), 357-365.
5. Ferrant, G., & Thim, A. (2019). Measuring Women’s Economic Empowerment: Time Use Data and Gender Inequality (Technical report, Issue. O. Publishing.
6. Fletcher, E., Pande, R., & Moore, C. M. T. (2017). Women and Work in India: Descriptive Evidence and a Review of Potential Policies. (HKS Faculty Research Working Paper Series RWP18-004. , Issue.
7. Mugumya, F., Asaba, R. B., Kamya, I. R., & Asingwire, N. (2017). Children and domestic water collection in Uganda: Exploring policy and intervention options that promote child protection. Child abuse and neglect in Uganda, 95-112.
8. Mugumya, T., Isunju, J. B., Ssekamatte, T., Wafula, S. T., & Mugambe, R. K. (2020). Factors associated with adherence to safe water chain practices among refugees in Pagirinya refugee settlement, Northern Uganda. J Water Health, 18(3), 398-408. https://doi.org/10.2166/wh.2020.230
9. Sedai, A. K. (2021). Who benefits from piped water in the house? Empirical evidence from a gendered analysis in India.
10. (2019). Uganda Time use Survey Report 2017-2018. UBOS.
11. (2021). 2021 National Service Deliverysurvey (NSDS): Water & sanitation. UBOS. https://www.ubos.org/wp-content/uploads/2022/12/NSDS-2021-h-Water-and-Sanitation.pdf
12. (2016). Collecting water is often a colossal waste of time for women and girls. UNICEF. https://www.unicef.org/press-releases/unicef-collecting-water-often-colossal-waste-time-women-and-girls
13. UNICEF/WHO. (2019). Progress on Household Drinking Water, Sanitation and Hygiene 2000-2017. Special Focus on Inequalities. https://data.unicef.org/topic/water-and-sanitation/drinking-water/.
14. Zozmann, H., Klassert, C., Klauer, B., & Gawel, E. (2022). Water Procurement Time and Its Implications for Household Water Demand—Insights from a Water Diary Study in Five Informal Settlements of Pune, India. Water, 14(7), 1009. https://www.mdpi.com/2073-4441/14/7/1009