📊 statistics

Psychometric Validation and Cultural Adaptation of the Household Water Insecurity Experiences (HWISE) Scale in Ethiopia using Horn’s Parallel Analysis Method

This study successfully culturally adapted and psychometrically validated the Household Water Insecurity Experiences (HWISE) scale for Sidaamu Afoo-speaking women in Ethiopia, demonstrating its unidimensional structure, high reliability, and strong validity through Horn's Parallel Analysis and other statistical methods.

Original authors: Adane Ermias, Sirak Robele Gari, Bizatu Mengiste, Argaw Ambelu

Published 2026-08-12

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Adane Ermias, Sirak Robele Gari, Bizatu Mengiste, Argaw Ambelu

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the world as a giant, bustling city where every home needs a steady stream of water to cook, clean, and stay healthy. Now, imagine that for some families, this stream is a shaky, unpredictable trickle. They worry about running out, they have to walk far to find it, and sometimes they have to skip washing their hands or even go to bed thirsty. This isn't just about being dry; it's a heavy, invisible weight that sits on a family's shoulders, changing how they eat, how they feel, and how they get along. Scientists call this "water insecurity." To fix it, they need a way to measure exactly how bad the problem is for different families. But you can't just use a ruler designed for a city in one country and expect it to work perfectly in another. The words need to make sense, the questions need to feel familiar, and the "ruler" itself needs to be tested to make sure it's not bent or broken. This is the job of psychometrics: the science of building and checking the measuring tools we use to understand human experiences.

The Story of the Water Ruler

Think of the HWISE Scale as a special, 12-question "water ruler" invented by scientists to measure how much water stress a family feels. It asks things like, "Did you worry about running out of water?" or "Did you have to skip a meal because of water problems?" Originally, this ruler was built in English for a global audience. But the researchers in this study, led by Adane Ermias and his team, wanted to use this ruler on a very specific group: women who speak Sidaamu Afoo in the Sidama region of Ethiopia.

You can't just take a ruler made for one culture and slap it onto another without checking if the markings still make sense. So, the team went on a mission to cultural adaptation. They didn't just translate the words; they acted like linguistic detectives. They asked, "Does the phrase 'in your household' mean the same thing as 'in your home' here?" They found that "in your home" felt more natural to the locals. They even held "cognitive interviews," which are like practice runs where women read the questions out loud and explain what they think the question means. It turned out that one word, "shame," was tricky. Some women felt it didn't quite capture the feeling of being left out or stigmatized by water problems, so the team tweaked the wording to make sure everyone understood the question exactly the same way.

The Big Test: Is the Ruler Straight?

Once they had their new, culturally friendly version of the ruler, they had to prove it actually worked. They gathered 264 women to take the test. But here's the tricky part: the answers weren't simple "yes" or "no." They were on a scale of "never," "rarely," "sometimes," and "often." In the world of statistics, this is like trying to measure a wobbly jelly with a straight ruler. Most standard math tools assume the data is smooth and straight, but these answers are bumpy and categorical.

To handle this, the researchers used a fancy statistical trick called Horn's Parallel Analysis. Imagine you have a bag of marbles (the real answers from the women) and you want to know if they form a single, strong line or a messy pile. You also have a bag of marbles you made up randomly (simulated data). You compare the two. If your real marbles form a much stronger, clearer line than the random ones, you know you've found a real pattern.

What They Found

The results were excitingly clear. The math showed that all 12 questions on the Sidaamu Afoo ruler were actually measuring one single thing: the feeling of water insecurity. It wasn't splitting into different categories like "worry" and "thirst" as separate things; they all pointed to the same core experience. This single "water stress" factor explained 67.6% of the answers, which is a huge chunk of the picture.

The ruler was also incredibly reliable.

Internal Consistency: The team calculated a score called ordinal alpha, which came out to 0.96. Think of this as a "trust score." A score of 1.0 is perfect, and 0.96 is practically perfect. It means if you asked the same woman the same questions in a slightly different order, she would give almost the exact same pattern of answers.
Stability: They asked the same women again two weeks later. The answers matched up with a score of 0.93, proving the ruler doesn't wobble over time.
Real-World Check: They also checked if the ruler made sense against real life. Women who reported high water insecurity scores actually spent more time fetching water, paid more for water, and had children with more diarrhea. The ruler wasn't just making up numbers; it was tracking real-life struggles.

What This Means

The study didn't find that water insecurity is a new problem, nor did it claim to have fixed the water supply. Instead, it proved that they now have a valid, reliable, and culturally tuned tool to measure the problem among Sidaamu Afoo-speaking women in Ethiopia. The paper explicitly ruled out the idea that the scale might be measuring multiple different things at once; the data showed it is strictly a one-dimensional measure of water stress.

The researchers are confident that this tool is ready to be used to check if water projects are working. However, they also noted that their study focused on women with young children in one specific district. They suggest that future explorers should test this ruler with different groups of people and in different seasons to make sure it holds up everywhere. For now, though, the team has successfully built a sturdy, accurate ruler for a community that needed one, ensuring that their voices about water insecurity can finally be measured with precision.

Technical Summary: Psychometric Validation and Cultural Adaptation of the HWISE Scale in Ethiopia

Problem Statement
Water insecurity is a critical global challenge, particularly in resource-constrained settings like Ethiopia, where infrastructure gaps and supply interruptions disproportionately affect vulnerable populations, especially women. While the Household Water Insecurity Experiences (HWISE) scale is a standard metric for measuring these experiences, its application requires rigorous local adaptation to ensure conceptual equivalence and psychometric robustness. Previous adaptations in Ethiopia and other low- and middle-income countries (LMICs) have utilized varying analytic approaches, with some studies employing Exploratory Factor Analysis (EFA) on Likert-type ordinal data without robustly addressing the non-normal distribution inherent in such response categories. This methodological limitation can lead to biased parameter estimations. Furthermore, there is a need for validated tools specifically tailored to Sidaamu Afoo-speaking populations in southern Ethiopia to accurately assess water insecurity among women, who bear the primary burden of water collection and management.

Methodology
This study employed a two-stage, community-based cross-sectional design to culturally adapt and psychometrically validate the HWISE scale for Sidaamu Afoo-speaking women in the Hawassa Zuria district of the Sidama National Regional State.

Cultural Adaptation (Stage I): The process followed a multi-step systemic procedure including forward translation by native speakers, back-translation by independent bilingual experts, and content validity analysis by a panel of 10 experts (covering environmental health, epidemiology, biostatistics, and local folklore). A Content Validity Index (CVI) was calculated. Cognitive interviews with 15 women were conducted to assess linguistic accuracy and cultural adaptability, utilizing "think-aloud" techniques to refine item comprehension.
Psychometric Evaluation (Stage II): Data were collected from 264 randomly selected adult women fluent in Sidaamu Afoo. A subset of 106 women was re-interviewed after a two-week interval to assess test-retest reliability.
Statistical Analysis: The study utilized R software and Classical Test Theory (CTT). Crucially, the analysis accounted for the ordinal nature of the 4-point Likert response categories (0–3).
- Construct Validity: Assessed via Exploratory Factor Analysis (EFA) using a polychoric correlation matrix. Horn's Parallel Analysis (PA) was employed alongside the Kaiser criterion and Cattell Scree Plot to determine the number of underlying factors, a method preferred for ordinal data with non-normal distributions.
- Reliability: Internal consistency was measured using the ordinal coefficient $\alpha$ (rather than standard Cronbach's $\alpha$ ) and Guttman Lambda 6. Temporal stability was assessed using the Intraclass Correlation Coefficient (ICC).
- Convergent and Discriminant Validity: Convergent validity was tested via Spearman's correlation against external constructs (water volume, collection time, cost). Discriminant validity was evaluated using the Mann-Whitney U test to compare scores across known groups (e.g., injury history, water source type, child diarrhea history).

Key Results

Cultural Adaptation: The translation process identified specific linguistic nuances, such as replacing "in your household" with "in your home" for better cultural fit. One item regarding "shame" was rephrased to "feeling insecure" after cognitive interviews revealed ambiguity. The final scale achieved an Item-level CVI (I-CVI) of $\ge$ 0.8 and a Scale-level CVI (S-CVI) of 0.9, indicating excellent content validity.
Construct Validity: The Kaiser-Meyer-Olkin (KMO) value was 0.924, and Bartlett's test was significant ( $p < 0.05$ ), confirming data suitability for factorization. Horn's Parallel Analysis revealed a unidimensional structure, with the first factor explaining 67.6% of the total variance. The real-data eigenvalue for the first factor (8.420) significantly exceeded the simulated eigenvalue (1.366), while subsequent factors were not retained. All 12 items demonstrated strong factor loadings (ranging from 0.674 to 0.888).
Reliability: The scale demonstrated excellent internal consistency with an ordinal $\alpha$ of 0.96 (95% CI [0.92, 0.99]) and a G6 estimate between 0.96 and 0.97. Test-retest reliability was also high, with an ICC of 0.93 (95% CI [0.90, 0.95]), indicating strong temporal stability.
Validity: Convergent validity was confirmed by significant correlations with external constructs: an inverse correlation with water volume consumed ( $\rho = -0.480$ ) and direct correlations with time spent fetching water ( $\rho = 0.549$ ) and cost ( $\rho = 0.549$ ). Discriminant validity was established as the scale significantly differentiated between groups with and without water-fetching injuries, improved vs. unimproved water sources, and households with vs. without recent child diarrhea.

Significance and Claims
The authors claim that the adapted HWISE scale is a valid and reliable instrument for measuring retrospective household water insecurity experiences among Sidaamu Afoo-speaking adult women in Ethiopia. By utilizing Horn's Parallel Analysis and polychoric correlations, the study addresses methodological gaps in previous adaptations, providing a more robust unidimensional measure that accounts for the ordinal nature of the data. The scale is presented as a rigorous tool for evaluating local water supply intervention programs. The authors note that while the scale is effective for the studied population, future research should aim to diversify samples and address seasonal variations in water insecurity.

More like this