Theoretical framework

Sustainable housing is grounded in the broader concept of sustainable development, commonly defined as development that meets present needs without compromising the ability of future generations to meet their own needs [19]. This concept has since evolved into a guiding principle for planning and policy across multiple sectors, including housing.

The theoretical foundation of this study is based on the triadic model of sustainability, which comprises economic, social, and environmental dimensions. This framework provides a holistic lens for evaluating long-term housing viability and has been widely adopted in sustainability research and policy [10]. The economic dimension focuses on affordability and financial accessibility; the social dimension emphasizes inclusivity, safety, and community well-being; and the environmental dimension addresses resource efficiency, resilience, and ecological impact.

This tripartite framework has become foundational in housing sustainability assessments globally. For instance, according to [20], sustainability in housing should integrate affordability (economic), inclusivity and safety (social), and environmental resilience. Similarly, [21] emphasizes that these pillars are interdependent and interact dynamically, requiring integrated, context-specific strategies for effective implementation. Recent studies further highlight that sustainable housing must respond to climate change, urbanization pressures, and socio-economic inequalities, particularly in developing countries [17,22].

Conceptual framework

This study conceptualizes sustainable housing as a multidimensional construct comprising economic, social, and environmental sustainability dimensions. These three pillars constitute the core explanatory variables that influence overall housing sustainability outcomes, consistent with established sustainability frameworks [10].

Within this framework, sustainability outcomes are assessed through indicators such as affordability, access to basic services, housing quality, environmental performance, and user satisfaction. The economic dimension captures housing affordability, cost efficiency, and financial accessibility; the social dimension reflects inclusivity, safety, and access to infrastructure; while the environmental dimension focuses on resource efficiency, energy use, and ecological impact [1,21].

Demographic characteristics, specifically age, gender, education level, and professional role, are treated as mediating variables that influence how individuals perceive and prioritize these sustainability dimensions. Empirical evidence shows that demographic factors shape housing preferences, environmental awareness, and affordability thresholds, thereby influencing sustainability evaluations [16,18].

The conceptual model, therefore, assumes that sustainability perceptions and housing outcomes are not uniform but are moderated by socio-demographic characteristics. This interaction provides a more nuanced understanding of housing sustainability by linking structural housing attributes with user-specific preferences and socio-economic realities.

Research gaps and emerging needs

While the academic and policy literature on affordable housing in Kenya has grown, several critical gaps remain. Little empirical research has been conducted on AHP in Meru and Embu counties [7,12]. There is a significant gap in studies evaluating the ecological sustainability of housing projects, particularly concerning green energy use and sustainable construction materials [1,13]. Limited research explores the influence of participatory planning and community engagement on project sustainability, despite its importance in enhancing ownership and long-term project success [1,23]. Most studies focus solely on economic sustainability, neglecting the interconnectedness of social and environmental factors that define true housing sustainability [1,13].

This study aims to fill these gaps by conducting a localized, multidimensional assessment that integrates demographic analysis and statistical modelling. It contributes to the growing call for data-driven, regionally informed, and socially inclusive housing policy frameworks in Kenya [24].

Despite the growing body of research, several gaps persist. First, most studies focus on Nairobi and Mombasa, neglecting semi-urban counties like Meru and Embu [7,12]. Second, there is limited exploration of how demographic variables influence sustainability perceptions, particularly across different stakeholder groups such as households, professionals, and local authorities [25]. Third, existing evaluations rarely triangulate economic, social, and environmental dimensions, often treating them in isolation rather than as an integrated sustainability framework [13]. Moreover, while sustainability policies exist on paper, few studies have measured their implementation effectiveness at the county level [12]. There is also minimal attention paid to the role of professional stakeholders such as contractors and academics in shaping housing outcomes and policy translation processes [23].

This study addresses these gaps by combining empirical data with stakeholder perspectives across two counties. It evaluates sustainability holistically and provides evidence-based recommendations for regional policy design.

Research design

The study adopted a descriptive cross-sectional survey design. Descriptive research is widely applied when the objective is to systematically describe characteristics of a population or phenomenon at a specific point in time [19]. This design was appropriate as it enabled the collection of data from a relatively large sample within a limited timeframe while examining relationships between demographic characteristics and perceptions of economic, social, and environmental sustainability within the AHP.

A cross-sectional approach was further justified because the study focused on capturing current perceptions, experiences, and evaluations without manipulating any variables [26]. Such designs are commonly used in housing and urban studies to obtain real-time stakeholder perspectives and contextual insights [27].

Study area and justification

The study was conducted in Meru and Embu counties, which are among the regions actively implementing the Affordable Housing Program. The two counties were selected due to their differing socio-economic, spatial, and infrastructural characteristics, which provide a suitable basis for comparative analysis.

Meru County exhibits a mixed urban–rural structure with moderate urban growth and ongoing housing developments. In contrast, Embu County is more semi-urban and has experienced relatively rapid urban expansion driven by population growth and infrastructure development [28]. These differences reflect distinct stages of urban development, governance structures, and service delivery capacities, making the two counties suitable for comparative sustainability analysis.

Target population and sampling technique

The target population comprised stakeholders directly or indirectly involved in the Affordable Housing Program (AHP), including public sector officials such as housing officers and urban planners, academic experts in housing and urban studies, private sector actors including architects, engineers, and contractors, as well as community representatives and prospective housing beneficiaries within Embu and Meru counties. The inclusion of respondents from different stakeholder groups was intended to capture diverse perspectives on the social, economic, and environmental dimensions of housing sustainability.

Purposive sampling was employed to select respondents with relevant knowledge, experience, or engagement in housing-related activities. This non-probability sampling technique was considered appropriate because the study sought information-rich participants capable of providing informed assessments of sustainability indicators within the AHP framework [29]. The approach also enabled the inclusion of respondents from policy, technical, and community contexts, thereby broadening the perspectives represented in the study.

A total of 215 respondents participated in the survey, comprising 132 respondents from Embu County and 83 respondents from Meru County. The variation in sample distribution reflected differences in housing activity levels, urban development patterns, and population concentration between the two counties. The final sample included respondents from public institutions (46.3%), academia (36.4%), and the construction sector (17.3%), as well as community-level perspectives from individuals familiar with local housing conditions and residential challenges. This composition strengthened the study by integrating policy, technical, and user-oriented viewpoints in the assessment of housing sustainability.

Instrumentation and data collection tools

Data were collected using a semi-structured questionnaire developed from established housing sustainability instruments [30]. The questionnaire consisted of five sections: demographic information; economic sustainability indicators (including affordability, household income, and construction costs); social sustainability indicators (including safety, tenure security, and access to services); environmental sustainability indicators (including use of green materials, energy efficiency, and waste management); and overall perceptions of program performance and satisfaction.

The instrument included Likert-scale items (1–5) and open-ended questions to capture qualitative insights. Respondents were also asked to rank their priorities across the three sustainability dimensions to support comparative analysis.

To enhance validity and clarity, the instrument was pilot-tested with 20 respondents in Tharaka Nithi County. Based on feedback, ambiguous items were refined, and the questionnaire layout was improved for clarity and usability.

Secondary data sources

Secondary data were obtained through document review and desk research. Sources included national housing policy documents, Kenya Vision 2030, the Big Four Agenda reports, County Integrated Development Plans, peer-reviewed journal articles, books, and credible institutional reports published between 2019 and 2024. Additional data were drawn from UN-Habitat and World Bank housing program assessments. These sources provided contextual grounding, supported questionnaire development, and informed comparative interpretation of findings.

Data collection procedure

Data collection was conducted over six weeks (February–March 2021). Questionnaires were administered face-to-face and online. Face-to-face administration was primarily used for public officials and community representatives, while online questionnaires (Google Forms) were distributed to academics and construction professionals due to logistical constraints.

Research assistants were trained to ensure consistency in data collection procedures and adherence to ethical standards. Regular supervision and follow-up checks were conducted to ensure completeness and accuracy of responses. The study involved non-sensitive survey data on housing perceptions. It did not include clinical, biomedical, or other high-risk human-subject research requiring invasive procedures or the collection of personal health information. Participation was entirely voluntary, and all respondents provided informed consent. Confidentiality and anonymity of respondents were maintained throughout the study, and no personally identifiable information was collected or disclosed.

Reliability and validity of the instrument

The questionnaire’s reliability was assessed using Cronbach’s alpha, in accordance with established methodological standards [31,32]. The reliability coefficients for the constructs were as follows: economic sustainability (α = 0.86), social sustainability (α = 0.81), and environmental sustainability (α = 0.79). All values exceeded the recommended threshold of 0.70, indicating satisfactory internal consistency [33].

Content validity was established through expert review by three housing specialists, a sociologist, and an environmental planner. According to [34,35], construct validity is further supported by an exploratory factor analysis (EFA), which, in this study, confirmed that the sustainability dimensions loaded distinctly with minimal cross-loadings, thereby validating the measurement structure.

Data analysis techniques

Quantitative data were analyzed using Stata version 18. Descriptive statistics, including means, frequencies, and standard deviations, were computed to summarise respondent characteristics and responses across the economic, social, and environmental sustainability dimensions.

An Exploratory Factor Analysis (EFA) was conducted to identify the latent constructs underlying participants’ perceptions of sustainability indicators. The EFA model is expressed as:

(1)

where represents observed variables, are latent factors, are factor loadings, and is the error term. The suitability of the data for factor analysis was assessed using the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and Bartlett’s Test of Sphericity. Factors were extracted using principal axis factoring and rotated using varimax rotation to enhance interpretability. EFA was appropriate for this study as it enables the identification of underlying structures among multiple correlated variables, making it widely used in social science research [36,37].

Principal Component Analysis (PCA) was subsequently applied to reduce the identified variables into a smaller set of uncorrelated components within each sustainability dimension. Each principal component is a linear combination of the original variables:

(2)

where are component loadings, components with eigenvalues greater than 1 were retained, following Kaiser’s criterion. PCA was selected to simplify the dataset and support the construction of composite indices, which is particularly useful in studies involving multidimensional constructs such as sustainability [38].

Generalized Linear Models (GLMs) were employed to examine the influence of demographic variables (age, gender, profession, and education level) on sustainability ratings. The general GLM specification is:

(3)

where is the link function, is the response variable, and are predictor variables. A Gaussian family with an identity link function was specified for continuous outcomes. GLMs were chosen for their flexibility in handling non-normal response distributions and for their suitability for modelling Likert-scale data as approximately continuous variables [39,40].

Independent samples t-tests and one-way Analysis of Variance (ANOVA) were used to assess statistically significant differences in sustainability perceptions across counties and stakeholder groups. These techniques are appropriate because the study involves comparing mean sustainability perception scores across independent categorical groups, with the dependent variables (economic, social, and environmental sustainability indices) measured on a continuous Likert-scale composite index.

The independent samples t-test is suitable for comparing the means of two independent groups, such as Meru and Embu counties, to determine whether observed differences in sustainability perceptions are statistically significant rather than due to random variation. For comparisons involving more than two groups (e.g., age categories, education levels, and professional roles), one-way ANOVA was employed to test whether at least one group mean differs significantly from the others. ANOVA is preferred in such cases because it controls for Type I error inflation that would arise from conducting multiple pairwise t-tests.

Post hoc tests (Tukey’s HSD) were conducted where applicable to identify specific group differences after a significant ANOVA result. These parametric tests are appropriate under the assumption that the data are approximately normally distributed, observations are independent, and variances are homogeneous across groups, conditions that are reasonably satisfied for aggregated Likert-scale indices in large samples [41].

Qualitative data from open-ended responses were analyzed using thematic analysis. Responses were coded inductively and deductively using NVivo version 12. Initial codes were generated, grouped into categories, and refined into overarching themes aligned with the sustainability framework. Thematic analysis was selected for its flexibility in identifying patterns across qualitative datasets and its suitability for mixed-methods research [42].

Results

This section presents findings derived from questionnaire data collected in Meru and Embu counties. Results are organized into three sustainability dimensions: social, economic, and environmental. In addition, multivariate analyses using Principal Component Analysis (PCA) and Generalised Linear Models (GLMs) are presented to examine latent structures in sustainability indicators and the influence of demographic variables on sustainability perceptions.

Overview of respondents

A total of 215 respondents participated in the study, of which 62.8% were male, and 37.2% were female, as shown in Table 1. In terms of educational attainment, 37.7% held a bachelor’s degree, while 21.9% had postgraduate qualifications (Master’s, postgraduate diploma, or PhD). Regarding professional background, 46.3% were from the public sector, 36.4% from academia, and 17.3% from the construction sector. The sample included respondents with varying levels of experience in housing-related activities.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Socio-economic characteristics of respondents.

https://doi.org/10.1371/journal.pone.0343915.t001

Social sustainability indicators

The analysis of social sustainability indicators (Table 2) shows that respondents assigned generally high importance to all variables, with mean scores ranging from moderate to very high across both Embu and Meru counties. Overall, safety and security emerged as the most important social criterion (pooled mean = 4.79), followed closely by access to electricity and health facilities (pooled mean = 4.74 for both). These findings indicate that basic service provision and safety considerations remain the primary determinants of perceived housing quality within the Affordable Housing Program (AHP). At the same time, comparatively lower importance was attached to infrastructure and amenity-based features such as lifts (2.95), daycare centres (3.11), and parking space (3.16).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Social Sustainability Indicators (Mean Scores by County).

https://doi.org/10.1371/journal.pone.0343915.t002

Marked spatial variations were observed between the two counties. As shown in Table 2, Meru respondents consistently assigned higher importance to accessibility-related indicators such as public transportation (4.81 vs 3.91), access roads (4.71 vs 4.00), and proximity to educational facilities (4.83 vs 4.25), with several of these differences being statistically significant (p < 0.001). In contrast, Embu respondents placed relatively greater emphasis on housing unit and neighbourhood quality attributes, including housing quality (4.55 vs 4.06), tenure security (4.58 vs 4.30), and privacy (4.27 vs 3.84). These patterns suggest that perceptions of social sustainability are strongly influenced by the level of urban development and the availability of infrastructure within each county.

Demographic differences further highlight heterogeneity in perceptions of social sustainability. Older respondents (ages 50+) consistently prioritized safety, infrastructure reliability, and service stability, reflecting concerns about long-term residential security. In contrast, younger respondents (below 35 years) placed relatively greater emphasis on digital readiness, particularly access to Wi-Fi and technology-enabled services. However, these features remain inconsistently integrated within AHP housing units. Overall, the findings presented in Table 2 demonstrate that social sustainability priorities are context-specific and shaped by both spatial development conditions and demographic characteristics, underscoring the need for differentiated housing design and planning approaches.

Economic sustainability indicators

The results presented in Table 3 indicate that affordability-related considerations primarily influence economic sustainability within the Affordable Housing Program (AHP). Housing affordability recorded the highest pooled mean score (4.80), followed by rental cost relative to income (4.73) and house price relative to income (4.71). These findings suggest that affordability pressures remain central in respondents’ evaluations of housing sustainability. The cost of construction also ranked highly (pooled mean = 4.48), highlighting concerns about rising building and material costs. In contrast, satisfaction of demand (4.25) and mortgage and interest rates (4.20) received comparatively lower ratings, reflecting continued concerns regarding housing finance accessibility and market responsiveness.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Economic Sustainability Indicators (Mean Scores by County).

https://doi.org/10.1371/journal.pone.0343915.t003

As shown in Table 3, statistically significant differences were observed between Embu and Meru counties for several economic indicators, including cost of construction (Embu = 4.24; Meru = 4.86; p < 0.001), employment opportunities (p < 0.001), transportation cost (p = 0.001), and satisfaction of demand (p < 0.001). Respondents in Meru generally expressed stronger concern regarding supply-side constraints and broader affordability pressures than respondents in Embu. However, housing affordability itself did not differ significantly between the two counties (p = 0.082), suggesting that affordability challenges are broadly shared across the study areas. Overall, the findings in Table 3 demonstrate that economic sustainability in the AHP is constrained not only by housing prices but also by financing accessibility, construction costs, and broader macroeconomic conditions.

Environmental sustainability indicators

The environmental sustainability results presented in Table 4 indicate that respondents placed high importance on environmental quality and resource management indicators within the Affordable Housing Program (AHP). Waste management emerged as the highest-rated environmental indicator, with a pooled mean score of 4.77, followed closely by materials with low environmental impact (4.71), air quality (4.70), use of appropriate materials (4.68), and water pollution control (4.67). These findings suggest that respondents are increasingly aware of the importance of environmentally responsible housing practices and their implications for health, livability, and long-term sustainability. In contrast, indicators associated with advanced sustainability practices, such as in situ energy production (3.27), use of recycled materials (3.26), cement substitutes (3.07), and thermo comfort (2.95), received comparatively lower ratings, indicating limited emphasis or lower implementation of green building technologies within existing housing projects.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Environmental Sustainability Indicators (Mean Scores by County).

https://doi.org/10.1371/journal.pone.0343915.t004

As shown in Table 4, statistically significant differences were observed between Embu and Meru counties across several environmental indicators (p < 0.05). Meru respondents assigned significantly higher importance to indicators related to pollution control, land use, biodiversity, urban density, and water management. In contrast, Embu respondents placed relatively greater emphasis on lighting quality, natural ventilation, and the use of regional or local materials. Particularly large disparities were observed for thermo comfort (Embu = 3.90; Meru = 1.45; p < 0.001), lighting quality (Embu = 4.55; Meru = 3.20; p < 0.001), and regional or local materials (Embu = 4.40; Meru = 3.42; p < 0.001). These variations suggest differences in environmental priorities, housing conditions, and infrastructure development between the two counties. Furthermore, the relatively low scores for recycled materials and renewable energy indicators suggest limited adoption of circular construction practices and clean energy technologies within the AHP.

Demographic characteristics also significantly influenced perceptions of environmental sustainability. Postgraduate respondents consistently assigned greater importance to environmentally sustainable housing features than respondents with lower levels of education. Similarly, female respondents demonstrated stronger prioritization of environmental quality indicators, particularly waste management, water conservation, and material sustainability. Subsequent GLM analysis confirmed education level as a statistically significant predictor of environmental awareness and sustainability prioritization (p < 0.01). Overall, the findings in Table 4 indicate that although environmental sustainability is increasingly recognized as an important component of housing development, implementation gaps remain in areas such as renewable energy integration, recycled material use, and climate-responsive construction practices.

Comparison of sustainability perceptions between experts and the general public

An independent-samples t-test was conducted to examine differences in sustainability perceptions between expert respondents (academics and construction professionals) and the general public. The analysis revealed statistically significant differences across selected social, economic, and environmental sustainability indicators, indicating that stakeholder groups prioritize different aspects of housing sustainability within the Affordable Housing Program (AHP).

For social sustainability indicators, expert respondents assigned significantly higher importance to technical and infrastructure-related factors such as effective property management (p = 0.007), design type (p = 0.013), unit size (p < 0.001), major and minor access roads (p = 0.011), and the number of bathrooms (p = 0.016) (see Table 5). Experts also placed greater emphasis on minimizing social segregation (p = 0.024). In contrast, members of the general public assigned greater importance to community cohesion (p = 0.037) and daycare centre availability (p = 0.091), reflecting stronger concern for neighbourhood livability and family-oriented residential environments. These findings suggest that while experts focus more on planning efficiency and housing functionality, the general public prioritizes social interaction and everyday residential convenience.

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Comparison of experts and General public.

https://doi.org/10.1371/journal.pone.0343915.t005

Economic sustainability comparisons showed fewer statistically significant differences between the two groups. However, experts expressed significantly greater concern about housing demand satisfaction (p = 0.001) and construction costs (p = 0.083), reflecting their closer engagement with housing supply systems, market constraints, and development feasibility. The general public, although also concerned with affordability, appeared more focused on practical housing accessibility and residential outcomes than broader market dynamics.

For environmental sustainability indicators, experts consistently reported higher ratings for technical environmental considerations, including noise pollution control (p = 0.013), water pollution management (p = 0.008), urban density (p = 0.012), water permeability (p = 0.030), waste management during operation (p = 0.017), and fresh water consumption (p = 0.006). Significant differences were also observed for land use and biodiversity considerations (p = 0.079). These results indicate that respondents with technical and professional backgrounds demonstrate stronger awareness of environmental planning and resource management issues. In contrast, the general public tends to prioritize environmental factors with immediate implications for residential comfort and livability.

Factor analysis and generalized linear model

To examine the underlying structure of sustainability indicators, the study employed Principal Component Analysis (PCA) as a data reduction technique. PCA was selected to identify latent dimensions within the social, economic, and environmental sustainability variables and to reduce redundancy among correlated indicators. Before extraction, the suitability of the dataset for factor analysis was assessed using the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and Bartlett’s Test of Sphericity. The overall KMO value of 0.812 exceeded the recommended threshold of 0.70, while Bartlett’s Test was statistically significant (p < 0.001), confirming that the correlation matrix was appropriate for component extraction. Components with eigenvalues greater than 1.0 were retained, and varimax rotation was applied to improve the interpretability of the extracted components. The PCA results identified seven social sustainability components, explaining 63.27% of the total variance; four economic sustainability components, explaining 60.50% of the variance; and six environmental sustainability components, explaining 65.88% of the variance. The extracted components represented meaningful sustainability dimensions: accessibility, infrastructure and services, affordability, housing quality, environmental performance, and community well-being.

The social sustainability PCA results indicated that accessibility to services, neighbourhood safety, housing quality, and community amenities were the dominant dimensions influencing respondents’ perceptions of sustainable housing. Economic sustainability components were primarily associated with housing affordability, construction and maintenance costs, income stability, employment opportunities, and mortgage accessibility. Environmental sustainability components addressed waste management, indoor environmental quality, water conservation, energy efficiency, sustainable material use, and pollution control. Collectively, these findings demonstrate that respondents’ sustainability perceptions were multidimensional, extending beyond physical housing attributes to encompass broader social, economic, and environmental considerations. The extracted factor scores from the PCA were subsequently used as dependent variables in the generalized linear models (GLMs) presented in Tables 6–8.

Download:

• PNG
  larger image
• TIFF
  original image

Table 6. GLM model for social factors.

https://doi.org/10.1371/journal.pone.0343915.t006

Download:

• PNG
  larger image
• TIFF
  original image

Table 7. GLM model for economic factors.

https://doi.org/10.1371/journal.pone.0343915.t007

Download:

• PNG
  larger image
• TIFF
  original image

Table 8. GLM model for Environmental factors.

https://doi.org/10.1371/journal.pone.0343915.t008

Generalized Linear Models (GLMs) were fitted to assess the influence of socio-demographic characteristics on the extracted sustainability dimensions. The PCA-derived factor scores served as continuous dependent variables, while education level, professional designation, gender, age category, and years of experience were included as categorical predictor variables. Dummy coding was applied for all categorical variables, and reference categories were specified as follows: no formal education for education level, private sector for professional designation, male for gender, 18–30 years for age category, and 0–9 years for years of experience. The reported regression coefficients therefore, represent the effect of each category relative to its corresponding reference group. Statistical significance was evaluated at p < 0.05, p < 0.01, and p < 0.001 levels.

The GLM results presented in Tables 6–8 indicate that education level was one of the strongest predictors of sustainability perceptions across all three domains. Respondents with postgraduate and PhD qualifications prioritized housing quality, environmental performance, and affordability-related dimensions significantly more than those with lower educational attainment. Age also significantly influenced sustainability perceptions, particularly in the social and environmental dimensions, where older respondents placed greater emphasis on accessibility, public services, and safety-related factors. Professional designations influenced several economic and social dimensions, with contractors emphasizing cost efficiency and public officers placing greater emphasis on infrastructure and regulatory considerations. Gender differences were comparatively smaller, although female respondents showed stronger concern for selected environmental sustainability indicators. Overall, the findings suggest that perceptions of housing sustainability are shaped not only by structural housing conditions but also by respondents’ socio-demographic and professional characteristics.

Social sustainability performance

The results show that social sustainability is the most developed dimension of the AHP in both counties. Safety and security were consistently rated highest among social indicators, consistent with prior research highlighting safety as a central determinant of residential satisfaction in urban housing systems [43]. Access to utilities such as electricity, water, and health facilities was also highly valued, reinforcing the importance of integrated infrastructure provision in housing delivery.

These patterns reflect broader structural inequalities in urban service provision typical of rapidly urbanizing contexts. According to [22], uneven urban infrastructure development remains a key constraint in achieving inclusive housing outcomes in developing economies, where housing delivery often outpaces service provision.

County-level differences: Urban development and planning context

The differences between Embu and Meru can be explained by their distinct urbanization trajectories and land-use governance structures. Embu County exhibits a more consolidated urban development pattern, influenced by stronger integration into regional transport networks and relatively more structured land use planning enforcement. This has contributed to improved infrastructure density and more stable service provision systems. As a result, residents place a higher value on environmental quality, neighbourhood aesthetics, and spatial livability.

In contrast, Meru County reflects a more transitional urbanization pathway characterized by dispersed peri-urban expansion, uneven infrastructure distribution, and slower alignment between housing development and service provision. Outlook [6] highlights that such transitional urban systems often prioritize basic service delivery needs such as water access, road infrastructure, and electricity reliability over environmental amenities. This explains why Meru respondents emphasized functional infrastructure more strongly than aesthetic or environmental attributes.

These findings align with [43], who argue that differences in planning systems and urban maturity significantly shape resilience priorities and housing expectations in developing urban contexts.

Demographic influences and behavioural drivers

Demographic characteristics significantly shaped sustainability preferences, reflecting deeper socio-economic and behavioural mechanisms documented in the literature. Recent studies underscore the importance of demographic factors in shaping sustainability perceptions, as individual preferences are influenced by differences in access to knowledge, social roles, and lived environmental risks [16,17].

Age differences reflect life-cycle housing priorities. Older respondents prioritized safety, stability, and long-term housing security, which can be explained by increased vulnerability to environmental and infrastructural risks as well as a stronger preference for reliability over innovation. In contrast, younger respondents showed a greater inclination toward smart housing features and shared amenities, reflecting increased exposure to digital technologies and evolving urban lifestyles [44].

Gender differences were particularly pronounced. Women placed greater emphasis on safety, sanitation, lighting, and access to water. Gendered roles and vulnerabilities in urban environments can explain this. Women are more likely to bear responsibility for household management and unpaid care work, which increases their dependence on reliable water, sanitation, and energy services [45,46]. In addition, women face higher exposure to safety risks in poorly serviced or inadequately lit environments, which explains their stronger prioritization of security-related housing features [47]. These structural inequalities make women more sensitive to deficiencies in basic infrastructure and neighbourhood safety.

Education emerged as a key determinant of preferences for environmental sustainability. Respondents with higher levels of education were significantly more likely to prioritize energy efficiency, waste management, and environmentally sustainable materials. This can be attributed to increased environmental awareness, greater access to information, and a better understanding of the long-term sustainability benefits. Empirical studies show that education enhances environmental competence across cognitive, behavioural, and attitudinal dimensions, leading to stronger pro-environmental decision-making [16,18]. Furthermore, more educated individuals are more likely to consider lifecycle costs and environmental externalities, rather than focusing solely on immediate affordability.

These findings demonstrate that sustainability preferences are not uniform but are shaped by socially embedded roles, access to knowledge, and differential exposure to environmental and economic risks.

Economic sustainability and structural constraints

Affordability was the dominant economic concern across both counties, yet satisfaction levels remained low, highlighting a persistent affordability gap in Kenya’s housing delivery system [2]. This challenge reflects broader macroeconomic and structural constraints. The African Development Bank [8] notes that housing affordability in African economies is heavily influenced by rising construction costs, reliance on imported materials, and constrained housing finance systems. Similarly, [4] highlights that financialization and financing structures in housing development often limit access for low-income households.

Differences across stakeholder groups further reinforce this interpretation. Public officials emphasized policy alignment, while contractors highlighted material costs and supply chain instability. This reflects a well-documented policy–implementation gap in infrastructure delivery systems in developing economies.

Environmental sustainability and the implementation gap

Despite strong policy-level support for sustainability principles [1], environmental integration remains the AHP’s weakest dimension. Respondents valued waste management and air quality but expressed dissatisfaction with renewable energy adoption and the use of local or green building materials. This gap reflects both institutional and structural constraints rather than the absence of materials or technologies.

Kibert [21] emphasizes that sustainable construction depends not only on material availability but also on contractor awareness, enforcement mechanisms, and lifecycle cost considerations. Similarly, [48] argues that decarbonization in developing countries is constrained by weak regulatory enforcement and limited green supply chains. Thus, low uptake of sustainable materials reflects systemic inefficiencies rather than purely technical limitations.

Integrated demographic effects and policy implications

A key contribution of this study is the demonstration that sustainability preferences are multidimensional and socially conditioned. Generalized Linear Model (GLM) results confirm that education is the strongest predictor of environmental and economic prioritization, while gender and age significantly shape social sustainability preferences.

These findings align with [22], which emphasizes that urban housing outcomes are increasingly shaped by demographic heterogeneity rather than uniform income categories. Similarly, [17] argues that sustainable development transitions require differentiated policy instruments rather than one-size-fits-all approaches.

County-based policy implications

Finally, the differences between Embu and Meru highlight the importance of spatially differentiated housing policy design. Meru’s emphasis on affordability and infrastructure reflects its semi-rural and transitional urban condition, while Embu’s focus on aesthetics and environmental quality reflects a more consolidated urban environment.

These findings support [43,44], who argue that urban planning systems must be context-sensitive and adapted to local governance capacity, spatial structure, and development stage. Decentralized housing governance, therefore, becomes essential for improving responsiveness, sustainability, and equity in housing delivery systems.