The initial item pool for the SFSI was not generated in isolation. It is the direct result of a synthesis of a 4 year-year longitudinal research program (2019-2023) in the study area. Prior phases of this program, employing a consistent sampling framework (n = 379 household heads), empirically identified the core sociological drivers of food insecurity in Torghar district—including institutional distrust (Khan et al., 2024c), poor governance (Khan et al., 2024c), military politics (Khan and Shah, 2024), socials safety net (Khan et al., 2023), social stratification (Khan et al., 2022), infrastructure lags and climate change (Khan et al., 2025), poverty (Khan et al., 2026), sociocultural dynamics and the erosion of traditional foodways (Khan et al., 2021). Qualitative and quantitively findings from these studies provided the substantive, context-specific content for the initial SFSI item pool, ensuring the index was grounded in documented community realities before participatory refinement. This study employed a constructivist quantitative design to develop and validate the SFSI. Departing from conventional metrics, the SFSI was constructed around four analytical lenses—historical, structural, cultural, and critical—to capture the socially produced nature of food insecurity among marginalized, postcolonial segments.

The study was conducted in Torghar district, Khyber Pakhtunkhwa, Pakistan—an agrarian region marked by severe food insecurity, extreme rurality (100%), and low human development index (HDI: 0.217) (Khan et al., 2025; Pakistan Bureau of Statistics, 2017). A stratified sampling technique based in selection of 379 households from Torghar's two tehsils (Judba and Khander) while using proportional allocation method (Bowley, 1926) from the total population of 26,464 as reflected in Pakistan Bureau of Statistics (2017), with Judba stood 14,972, and Khander's stood as 11,492 households respectively. By applying the proportional allocation method Judbha's had 214 households, and Khander had 165 households, constituting the whole sample size of 379, deemed appropriate for drawing the true inferences with Bougie and Sekaran (2025), sample size not less than then 300 is deemed appropriate to draw inferences in any survey study.

The participants met inclusion criteria: (1) ≥ 5 years of residency and (2) age ≥ 19 years. A significant methodological limitation must be noted: given the region's patriarchal norms, only male household-headed households were Interviewed/surveyed, aligning with local decision-making structures but unavoidably excluding women's perspectives—a limitation for future research. This approach balanced methodological rigor with cultural relevance for initial scale development, providing a representative snapshot of household-level food security dynamics as reported by male heads in a marginalized agrarian social structure.

Following the exploratory factor analysis, the original 4 theoretical dimensions of the SFSI were empirically restructured into a six-factor structure that better captured the empirical patterns. The revised factors include:

To operationalize the index, factor scores were computed using the Thurstone method, ensuring each dimension contributed proportionally to the composite SFSI score. Internal consistency was evaluated via Cronbach's alpha, with all subscales meeting reliability thresholds. Construct validity was assessed through correlation analyses testing hypotheses relationships between SFSI scores and key socioeconomic variables, ensuring alignment with the index's sociological foundations.

Descriptive statistics for all 20 SFSI items across four theoretical dimensions are presented in Supplementary Tables S1–S4 (corresponding to the original Historical, Cultural, Structural, and Critical dimensions). Overall patterns reveal distinct sociological contour of food (in) security in the study area. The composite SFSI score across households had a mean of 1.58 (SD = 0.304), indicating moderate levels of perceived sociological constraints on food access. Within the historical dimension, the strongest consensus emerged around intergenerational decline in food sovereignty, with item HIS-5 (“Our grandparents ate better from local farms”) receiving the highest mean score (1.83). Items related to land dispossession (HIS-1) and discriminatory land laws (HIS-3) also showed strong agreement, reflecting enduring grievances about colonial and post-colonial inequities (See Supplementary Table S1).

In addition, the cultural dimension revealed tension between preservation and erosion of traditional floodways. While concerns about media-driven dietary changes (CUL-4: “TV ads promote junk food”) scored highest (Mean = 1.85), items capturing gendered hierarchies (CUL-2: “Women eat last and least”) showed more variable responses, suggesting patriarchal norms are both recognized and potentially normalized (see Supplementary Table S2).

Moreover, structural barriers were prominently identified, particularly institutional distrust. Item STR-3 (“Ration shops cheat us”) elicited the strongest agreement in this dimension (Mean = 1.62), highlighting systemic failures in food distribution mechanisms. Economic pressures (STR-1, STR-5) and infrastructure gaps (STR-4) were also salient, though with slightly lower mean scores (see Supplementary Table S3).

Finally, the critical dimension demonstrated high awareness of systemic inequities. Item CRI-3 (“The system is rigged”) received the highest mean score across all items (1.98), followed closely by CRI-2 (“Cities waste food while we hunger”; Mean = 1.86). This indicates widespread perception of urban-rural disparities and systemic injustice in food systems (see Supplementary Table S4).

Given the exploratory nature of scale development, I employed the EFA to investigate the underlying structure of the SFSI items rather than to confirm a predetermined structure (See Table 3). The table revealed a proposed six-factor structure, which refined and refined the original theoretical dimensions of food security. The historical entitlement items (HIS-1 to HIS-5) loaded strongly on Factor 1 (loadings 0.673 to 0.700), suggesting this as a distinct dimension while showing intersecting secondary loadings—particularly HIST-4′s negative association with Factor 6 (−0.351), suggesting some respondents might have perceived lost traditions as separate from broader historical injustices. The structural deprivation items formed a coherent Factor 2, with STR-4 (“Market are too far”) showing the strongest loading (0.757), indicating infrastructure barriers were central to this dimension. However, STR-3′s primary loading on Factor 6 (0.591) rather than Factor 2 revealed that issues of ration shops corruption represented a distinct aspect of structural barriers in this context.

The critical consciousness items clustered effectively on Factor 3, with CRI-3 (“The system is rigged”) showing the strongest loading (0.744), supporting this dimension's focus on systemic critique. Interestingly, CRI-1's negative loading on Factor 6 (−0.493) suggested perceptions of corporate control may inversely relate to another hybrid factors in ways requiring further investigation. Cultural items split across two factors: cultural sovereignty (Factor 4, led by CUL-2's strong 0.857 loading on gendered eating practices) and cultural erosion (factor 5, with CUL-4's 0.607 loading on media-driven dietary changes). This bifurcation indicated that cultural dimensions of food security might operate through distinct mechanisms of preservation vs. transformation.

The analysis showed generally clean factor structure with most items loading strongly on their primary factors (≥0.40) and relatively low cross-loadings, supported by moderate uniqueness values (0.250–0.707). The oblique rotation revealed modest inter-factor correlations (all < 0.20), suggesting these dimensions were conceptually related but empirically distinct from each other. The emergence of a 6th hybrid factor (containing STR-3 and CRI-1) points toward potential intersections between structural barriers and critical consciousness that warranted further theoretical development. Overall, the EFA results supported the SFSI's multidimensional framework while revealing nuanced patterns that enhanced the understanding of how different aspects of food security interrelate in this context.

Table 4, the inter-factor correlation matrix reveals several important relationships between the six dimensions of SFSI. The strongest positive correlation emerges between Factor 4 (Cultural Sovereignty) and Factor 5 (Cultural Identity Erosion) (r = 0.163), suggesting that communities maintain traditional food knowledge were simultaneously experiencing some degree of dietary transformation—a finding that reflected the complex interplay between cultural preservation and change in food systems. On the other hand, moderate but meaningful positive correlations emerged between factor 1 (Historical Entitlement) and Factor 4 (r = 0.200), indicating that communities with stronger historical consciousness tend to maintain more promising cultural food practices. Similarly, Factor 3 (Critical Consciousness) explained a positive association with Factor 5 (r = 0.156), implying that awareness of systemic inequities displaying a recognition of cultural erosion. Notably, several factors demonstrated weak negative correlations i.e., factor 1 with Factor 3 (r = −0.062) and factor 6 (r = −0.062) and Factor 2 (Structural Deprivation) with Factor 3 (r = −0.074) and Factor 6 (r = −0.081). These inverse relationships suggested the historical grievances and structural barriers did not translate directly into critical consciousness about food systems, highlighting potential disconnects between lived experience and systemic understanding.

The strongest negative correlation occurs between Factor 4 and factor 6 (r = −0.142), indicating that communities with strong cultural food sovereignty tend to report fewer issues related to the hybrid dimension (containing ration shop corruption and corporate control). This pattern may reflect how cultural resilience buffers against specific systemic failures. Most correlations remain below [0.20], confirming that the 6 factors represent distinct albeit interrelated dimensions of food security. The generally low multicollinearity supports their treatment as separate components in subsequent analyses while acknowledging there nuanced connections in lived experience. These correlation patterns enrich our understanding of how historical, structural, cultural, and critical dimensions interact in shaping food security inferences.

Table 5, Presents the model fit statistics, which provide a nuanced assessment of the factor structure's validity in this exploratory context. The RMSEA value of 0.0918 (90% CI: 0.0822–0.102) suggests an adequate but imperfect fit, falling between conventional thresholds for acceptable (0.08) and poor (0.10) fit. The Tucker-Lewis Index (TLI) of 0.751 is below the preferred (0.90) benchmark for excellent fit but remains above the 0.70 threshold often considered acceptable for early scale development. These moderate values are not uncommon when measuring complex, interrelated sociological constructs where theoretical dimensions naturally overlap in lived experience. The Bayesian Information Criterion (BIC) value of −148 favors the six-factor solution, as the most parsimonious. While the significant chi-square statistic (χ² =357, df = 85, p < 0.001) indicates some discrepancy between model and data, this test is known to be overly sensitive with large samples like the present study (n = 379). Collectively, these indices support the exploratory use of the proposed 6-factor SFSI structure while clearly indicating opportunities for refinement in future proposed studies. The adequate-but-not-optimal fit reflects the challenge of quantifying multidimensional social phenomena and suggests that item refinement or modest model re-specification could enhance psychometric properties while maintaining theoretical coherence.

The assumption checks for our factor analysis confirm the appropriateness of the EFA approach for our dataset (Table 6). Bartlett's test of sphericity yielded a highly significant result (χ² = 2660, df = 190, p < 0.001), providing strong evidence against the null hypothesis that the variables are uncorrelated in the population. This significant result indicated that there were sufficient correlations among the study variables to proceed with factor analysis, as the items share enough common variance to potentially form meaningful factors. The extremely large chi-square value relative to the degrees of freedom suggests that items possessed substantial common variance, supporting the theoretical expectation that the various dimensions of food security are interrelated. This finding validates the decision to use factor analysis to uncover the underlying structures of our SFSI, as the test confirms that the correlation matrix is sufficiently different from an identity matrix to warrant dimensionality reduction. These results, combined with previously reported KMO measures of sampling adequacy (0.681), provided initial evidence that our data meet the key assumptions for EFA. The strong significance of Bartlett's test further reinforces the interrelated aspects of food security rather than completely independent constructs, aligning well with our multidimensional theoretical framework.

The Kaiser–Meyor–Olkin (KMO) measure provided important insights into the factorability of our individual items and the overall dataset (Table 7). The overall KMO statistics of 0.681 falls within the “mediocre” to “middling” range according to Kaiser's classification. While this suggests adequate factorability for an exploratory study of complex, multidimensional construct like food security, it appropriately indicated that the sociological dimensions measure were distinct yet interrelated in lived experience, not perfectly correlated constructs. This moderate value is conceptually appropriate for capturing the real-world complexity of food systems.

At the item level, a range of sampling adequacy values revealed intersecting patterns. The historical dimension items showed strong to very strong adequacy (0.651–0.777), with HIST-2 (“We grow fewer types of crops”) and HIST-4 (“We've forgotten old ways”) demonstrating particularly preliminary common variance (0.777 and 0.756 respectively). Most notably, item STR-3 (“Ration shops cheat us”) demonstrated the lowest sampling adequacy (MSA = 0.477). Despite this lower value, the item was retained based on three principled justifications that align with the participatory, theory-driven approach:

First, participatory validity: STR-3 emerged as a critical and recurrent theme during the deliberative workshops, capturing a context-specific form of institutional distrust and structural violence central to theoretical framework on power asymmetries.

Second, theoretical centrality: The item directly operationalizes concepts of structural deprivation through institutional failure, a core dimension of our sociological approach that would substantially weaken without it.

Third, empirical contribution: Despite its lower MSA, STR-3 loaded meaningfully onto the empirically derived hybrid factor (factor 6, loading = 0.591), confirming its substantive importance in capturing unique intersection of structural barriers and critical consciousness relevant to this agrarian context. The decision reflects the methodological commitment to balancing statistical criteria with participatory and theoretical validity in initial scale development. Future empirical studies could explore refining this item (e.g., splitting into behaviorally specific sub-items) while retaining its core meaning.

Conversely, critical consciousness items demonstrated good to superb adequacy (0.566–0.806), with CRI-5 (Droughts/floods destroy crops”) showing exceptional common variance (0.806). Collectively, these results provided confidence that the factor analysis can meaningfully uncover the underlying structure of the data. The decision-making process—prioritizing participatory validity and theoretical bases alongside statistical metrics—underscores the transformative epistemology of the SFSI, which values community-identified realities as paramount.

The scree plot (Figure 2) displays eigenvalues for 20 potential factors, where the steep decline in eigenvalues from Factor 1 to 6 suggests these dimensions captured substantial variance, while the subsequent flattening (“elbow”) after Factor 6 indicated diminishing returns. This supports retaining 6 factors—consistent with parallel analysis recommendations—as they represent the most meaningful, interpretable dimensions before the eigenvalues taper toward negligible contributions (resembling random noise). The plot thus validates the six-factor solution as optimally balancing explanatory power with parsimony.

The correlation matrix revealed significant relationship between the SFSI and key socioeconomic variables in Table 8. Education exhibited the strongest positive correlation with food security (SFSI: r = 0.364, p < 0.001), underscoring its pivotal role in enhancing food access and entitlements. Age further reflected a vigorous positive association (SFSI: r = 0.342, p < 0.001), suggesting older individuals may benefit from accumulated resources or traditional knowledge. Income correlates moderately with SFSI (r = 0.319, p < 0.001), though less strongly than education, highlighting that economic factor alone does not fully explain food security disparities. Notably, family type (e.g., patriarchal structures) have weak negative correlation with SFSI (r = −0.133), implying potential constraints on food access in certain household arrangements. Among covariates, education and age are strongly linked (r = 0.658, p < 0.001), while income and family type revealed a negative relationship (r = −0.253, p < 0.001), possibly reflecting economic pressures on traditional family units. These patterns align with SFSI theoretical grounding, emphasizing structural (education, income) and cultural (age, family type) dimensions of food (in)security.

The correlation heat map (Figure 3) visually reinforces key relationships between the SFSI and socioeconomic variables, with color intensity and direction (warm hues for positive, cool for negative) highlighting patterns. Education (darkest warm tone, r = 0.36 with SFSI) and age (r = 0.34) displayed the strongest positive associations with food security, confirming its central role. Income (r = 032 with SFSI) exhibited a moderate link, though weaker than education, aligning with SFSI emphasis on non-economic dimensions. Notably, family type had a faint cool tone (r = −0.13), suggesting mild negative effects, could be due to patriarchal constraints. The heat map also revealed a strong covariate relationship: the darkest warm cluster between education-age (r = 0.66) and income-age (r = 0.57) underscores their interdependence, while income-family type (r = −0.25) indicated economic pressures on traditional households. The gradient scale (from −1.0 to 1.0) clarifies magnitude, with SFSI correlations concentrated in the 0.3–0.4 range (moderate) and covariates reaching 0.5–0.7 (strong) respectively.

The presented GLM path analysis meets key statistical assumptions for valid interpretation. VIF for all predictors were below 2.0, confirming the absence of severe multicollinearity. Residual plots were also inspected and indicated no severe violations of the model's distributional assumptions. Crucially, it must be emphasized that due to the cross-sectional nature of the data, the identified pathways represent statistically significant associations and plausible mediated relationships consistent with sociological theory, but do not confirm causal effects. The model illustrates a coherent structure of how these variables are interrelated in this specific context.

The analysis reveals education's critical mediating role in the statistical associations with food security in Table 9. All three indirect effects via education were statistically significant (p < 0.01), with comparable effect sizes for age (β = 0.105) and income (β = 0.094), and a smaller but meaningful effect for family type (β = 0.041) respectively. The component paths converged with particularly strong relationships: age and income both strongly predicted education (β = 0.416 and β = 0.372 respectively), while education's effect on SFSI was moderate but vigorous (β = 0.253). To contextualize these effect sizes, the standardized coefficient for education (β = 0.253) could be considered a moderate-to-strong effect in social science research, and is notably larger than the non-significant direct effect of income (β = 0.002). This pattern challenges economic reductionism with a growing body of evidence from the Global South. For instance, studies in Pakistan and rural India have a symmetry that education and asset-based indicators often outperform income alone in predicting household food security, as they better capture long-term capabilities and access to institutions (e.g., Anwar et al., 2024).

The direct effects reflect a more complex picture. Family type showed a significant negative direct effect (β = −0.210, p < 0. 001), indicating an association where patriarchal household structures are linked to undermine food security. The magnitude of this negative effect is substantial, underscoring that gendered household dynamics exert a strong association with food security comparable to or greater than many economic factors found in other studies. Age maintains a positive direct effect (β = 0.214, p = 0.001), suggesting additional life experience benefits not captured by educational attainment. Notably, income's direct effect is negligible (β = 0.002, p = 0.981) implying its observed statistical association with the SFSI operates entirely through education—a striking inference that challenges conventional economic assumptions about food security determinants.

Lastly, total effects synthesized these pathways: age emerges as the strongest predictor overall (β = 0.319). The total effects of age highlighted the paramount importance of intergenerational knowledge and life-course assets, a finding that resonates with research in agrarian communities where older age confers access to traditional ecological knowledge and social networks that buffer against food shocks. Family type's total effect remains negative (β = −0.169), while income's total effect becomes marginal (β = 0.096, p = 0.133) These results initially support the SFSI's sociological framework, demonstrating that: (1) structural factors (education) mediate much of the observed statistical association from traditional predictors like income; (2) cultural/age-related factors retain independent importance; and (3) household dynamics (family type) exert unique negative pressures. The complete standardization of effects (β) facilitates comparison across variables, with education-mediated paths accounting for 33–100% of various total effects.

The path diagram (Figure 4) illustrates the relationships between key socioeconomic predictors, education, and the SFSI. Education acts as the central mediator, showing a modest direct effect on SFSI (β = 0.10). Age also has a small positive direct effect on SFSI (β = 0.09), while family structure exerts a negative direct influence (β = −0.12). Notably, income shows no direct effect on SFSI (β = 0.00); its association is entirely channeled through education. The model further reveals that education itself is strongly predicted by age (β = 0.42) and income (β = 0.37), with family type having a moderate positive influence (β = 0.27).

The inter-correlation among predictors provide additional context: age and income are strongly correlated (β = 0.51), as are age and family type (β = 0.38), while family type and income show a weak negative association (β = −0.12). Synthesizing these paths, the total effect of income on SFSI is fully mediated via education. Age influence SFSI through both a strong indirect pathway via education and a smaller direct effect. For family type, a small positive indirect effect through education is counteracted by its stronger negative direct effect on SFSI. This integrated model confirms that while education is a critical conduit for economic resource, non-economic factors—specifically age (reflecting intergenerational capital) and patriarchal family structures—retain significant independent associations with sociological food security.

The central fining of this study that education appears to mediate nearly all of income's statistical association with food security—provides initial empirical support that aligns closely with Sen's entitlement approach. Sen (1981) posited that hunger stems not from a lack of food, but from a failure of entitlements: the social, legal, and economic means by which an individual's command food. Our results operationalize this theory, suggesting that in Torghar, income alone is a meaningless abstraction. Its power to secure food is almost entirely contingent on being converted into educational attainment—a key institutional entitlement that enhances capabilities, access to information, and bargaining power within corrupt or inefficient systems (as evidenced by item like STR-3, “Ration shops cheat us”). These findings challenge neoliberal, income based models by highlighting the conversion factors (like education) that transform resources into genuine food access.

The significant negative direct effect of patriarchal family structure (β= −0.210, p < 0.001) on food security, independent of education or income, critically extends entitlement theory. It reveals a fundamental intra-household entitlement failure. Agarwal's (2021) feminist political ecology framework helps interpret this: patriarchal norms act as a social entitlement filter, systematically diverting food and resources away from women and other dependent members, irrespective of household-level aggregate entitlements. Our item CUL-2 (“Women eat last and least”) quantifies this gendered hierarchy. This finding aligns with broader analyses of how patriarchal control over resources is central to the functioning of rural political economies (Brysk and Mehta, 2017; Rao, 2024). It necessitates a shift in food security policy from the household as a black box to a site of gendered power analysis. However, the exclusive surveying of male heads, while pragmatic, means this measurement likely captures a perception or reported norm rather than direct, lived experience of deprivation by women themselves, underscoring a critical limitation in fully capturing this dynamic.

The SFSI's dimensions of critical consciousness and historical entitlement illuminate the Gramscian and Bourdieusian dynamics at play. High scores on items like CRI-3 (“The system is rigged”) indicate an awareness of systemic inequities—a crack in the hegemonic consent for a food system marked by corporate control and historical injustice (Brown, 2020; McMichael, 2009). Simultaneously, the erosion of cultural sovereignty (e.g., youth rejecting traditional foods) and the persistence of historical grievances demonstrate how symbolic violence (Bourdieu, 1984; Wang et al., 2025) operates: communities may internalize their deprivation as a natural order or cultural decline, rather than a political outcome. The SFSI thus captures the contradictory consciousness inherent in marginalized agrarian communities: a critique of the system co-exists with a loss of traditional, resilient food practices.

The finding that age is the strongest predictor of food security (β = 0.319) underscores the value of intergenerational knowledge as a non-market entitlement and a source of social capital. This aligns with political ecology perspectives that position traditional agroecological knowledge as a critical buffer against market and climate shocks (Watts and Bohle, 1993; Khan et al., 2025). However, the concurrent measurement of cultural identity erosion (e.g., CUL-4 on junk food advertising) signals a double vulnerability: the very reservoir of resilience is being depleted by globalized food cultures. This tension highlights a core crisis in the corporate food regime: it simultaneously makes communities dependent on volatile markets while undermining the local knowledge system that historically provided stability (Friedma and McMichael, 1989).

Collectively, these findings advance a sociology of food entitlements that integrates Sen's focus on legal-economic means with feminist, Gramscian, and Bourdieusian analyses of power. The SFSI demonstrates that food insecurity is associated with the intersection of structural barriers (mediated by education), patriarchal household filters, eroded cultural sovereignty, and a critical yet often disempowered consciousness of systemic injustice. This framework moves beyond diagnosing discrete “factors” to modeling the relational matrix of power that determines who eats and who does not. Theoretically, it bridges the macro-analysis of food regimes with the micro-politics on household allocation, providing a measurable link between global corporate structures and local experiences of hunger. The key implications is that achieving food security is not merely a technical or economic challenge, but a political project of transforming social relations—of strengthening entitlements, dismantling gendered hierarchies, and revitalizing the cultural and ecological knowledge that buffers communities against domination. Future research must build on this foundation by applying the SFSI longitudinally to test these relational dynamics over time and across diverse contexts, thereby refining our understanding of how power—in all its sociological complexity—continues to structure hunger in the post-colonial world.

Our findings offer specific, evidence-based insights to enhance the targeting and monitoring of existing national policies, such as National Food Security Policy (2018) and the Benazir Income Support Programme (BISP). First, the strong mediating role of education suggests that integrating educational conditions or supports into cash transfer programs could amplify their impact on long-term food security. For instance, piloting a module within BISP that verifies children's school attendance or subsidizes adult literacy—particularly for women—directly addresses the structural and critical pathways identified by the SFSI.

Second, the significant negative association between patriarchal family structures and food security underscores a critical gap in household level “utilization” strategies. This implies that gender-sensitive metrics, such as monitoring intra-household food allocation (e.g., through the SFSI item “women eat last”), should be incorporated into the monitoring frameworks of the National Zero Hunger Program to ensure interventions reduce gendered deprivation, not just caloric intake at the household level. Critically, such monitoring must find ways to directly and ethically capture women's perspectives to be effective.

Rather than prescribing broad institutional reforms, the primary application of the SFSI is as a diagnostic and monitoring tool to make existing policies more responsive. For government agencies, it can be used as a supplementary module in national surveys to map the sociological drivers of insecurity—identifying whether a region suffer primarily from structural deprivation (e.g., market failure) or cultural erosion (e.g., loss of traditional crops). This allows for spatially and socially targeted resource allocation. For NGOs, the SFSI provides a framework for participatory baselines to ensure projects address root causes like institutional distrust (Ration shop cheat us) rather than just symptoms. In this way, the SFSI does not call for policy overhaul but offers a method to bridge the gap between broad policy goals and the specific, power-layered realities of marginalized rural hinterland communities.

This study has several important methodological limitations that must be acknowledged to properly contextualize its contributions. First, regarding scale development and verification, it is crucial to frame this work as exploratory scale development rather than definitive confirmation. While the SFSI demonstrates promising psychometric properties, its endorsement remains incomplete. The moderate Kaiser-Meyer-Olkin measure (0.681) and model fit indices (RMSEA = 0.92, TLI = 0.751) indicate an acceptable but imperfect factor structure that would benefit from refinement in future confirmatory studies. Relatedly, item STR-3 (“Ration shops cheat us”) demonstrated lower sampling adequacy (MSA=0.477) but was retained due to its strong participatory validity and theoretical importance in capturing institutional distrust—a decision that highlights the tension between statistical criteria and contextual relevance in community-centered research.

Second, and most critically for a study focused on power and gendered deprivation, the exclusion of the women's perspective represents a significant constraint. Surveying only male household heads, while aligned with local patriarchal norms for feasibility, unavoidably limits our understanding of intra-household food allocation and gendered deprivation mechanisms central to the theoretical framework. This exclusion likely leads to systemic underestimation of food insecurity, particularly regarding items measuring gendered hierarchies (e.g., “women eat last and least”), and may bias the measurement of cultural and intra-household dynamics. Future applications of the SFSI must prioritize methodological innovations (e.g., separate interviews with women using female enumerators) to ethically include these vital perspectives.

Third, methodological constraints warrant consideration. The cross sectional design precludes definitive causal claims; while path analysis suggests plausible mediated relationships, these findings indicate association, not causation. The 3-point Likert scale, chosen to reduce cognitive load in a low-literacy setting, may limit detection of subtle attitudinal gradients compared to more granular scales. Additionally, the study is geographically confined to Torghar district, a marginalized agrarian region, limiting immediate generalizability to other socio-cultural contexts without adaptation. Finally, analytical consideration includes the need to interpret findings with appropriate caution. The identified pathways should be viewed as a theoretically coherent model of association rather than a causal map. The SFSI in its current form does not capture certain critical dynamics identified by feminist political ecology, such as unpaid care work or time allocation, which are the key to understanding women's food security.

Future research should address this limitation through: (1) ethical inclusion of women's perspectives using female enumerators and complementary instruments; (2) longitudinal or quasi-experimental design to test causal mechanisms; (3) cross-context proof and refinement of the SFSI in diverse agrarian settings; and (4) integration of mixed methods to enrich quantitative scores with narrative evidence of how power and culture shape food access in daily life.

Terminology Note: To ensure clarity for a multidisciplinary readership, key terms used throughout this manuscript are defined as follows: