Social Support and Work-Family Conflict Among Career Women: Implications for Sustainable Workforce Development Planning in West Sumatra, Indonesia

Sustainable workforce development planning has increasingly recognised that the long-term productive engagement of women in formal employment is not merely a matter of recruitment or entry-level access, but of retention, well-being, and equitable working conditions across the entire career lifecycle [1]. This recognition is institutionalised in the United Nations’ Sustainable Development Goals (SDGs), particularly SDG 5 (Gender Equality), SDG 8 (Decent Work and Economic Growth), and SDG 3 (Good Health and Well-being), each of which acknowledges the interconnection between women’s employment conditions and broader sustainability outcomes [2]. Despite this normative consensus, the planning mechanisms required to support sustainable female workforce participation remain incompletely specified in many regional development contexts, including the rapidly urbanising provinces of Indonesia.

In West Sumatra, female labour-force participation reached between 68.50% and 74.36% in 2022, reflecting a substantial and growing professional female population [3]. Yet participation statistics alone do not capture the conditions under which women sustain their careers without incurring disproportionate costs to their health, family relationships, and psychological well-being. A primary mechanism through which these costs are incurred is work-family conflict (WFC), the inter-role tension that arises when the time demands, behavior expectations, and psychological strains of professional and domestic roles are mutually incompatible [4]. Left unaddressed, WFC constrains sustainable career trajectories, contributes to voluntary workforce exit, and perpetuates gender-based occupational inequality [5, 6].

Social support comprising emotional, instrumental, informational, and appraisal assistance provided by members of an individual’s relational network is among the most frequently cited modifiable factors in WFC research [7, 8]. Theoretically grounded in Thoits’ [9] coping-resource model and Cohen et al.’s [10] multidimensional framework, social support is hypothesised to buffer the resource-depleting effects of competing role demands, thereby reducing conflict and sustaining occupational engagement. The Work-Home Resources (WHR) model of Ten Brummelhuis and Bakker [11] formalises this logic within a resource-conservation framework that has direct implications for workforce planning: if social support reliably attenuates WFC, then supporting the development of strong relational networks constitutes a viable workforce sustainability strategy.

However, the translation of this theoretical hypothesis into planning-relevant evidence is complicated by important contextual variables. The effectiveness of social support as a WFC buffer varies considerably across occupational sectors, cultural settings, and family structures [12, 13]. In West Sumatra, the dominant Minangkabau matrilineal tradition creates a distinctive socio-cultural context in which women occupy formal authority within extended-family systems while simultaneously bearing professional and Islamic domestic obligations [14]. Whether this context amplifies or attenuates the buffering function of social support and what this implies for sustainable workforce development planning has received no systematic empirical attention.

The present study addresses this gap by examining the correlational relationship between seven source-specific dimensions of social support from husbands (SSH), children (SSC), parents (SSP), extended family (SSEF), in-laws (SSI), work colleagues (SSCL), and direct leaders/supervisors (SSL) and WFC among 60 professional women in West Sumatra. Using PLS-SEM, the study evaluates both the psychometric quality of the measurement model and the structural significance of each hypothesised path. The findings are interpreted through the lens of sustainable workforce development planning, with direct implications for regional policy design, organisational human resource management, and professional counselling practice.

The study makes three contributions to the scope of sustainable development and planning research. First, it generates source-disaggregated empirical evidence on social support–WFC dynamics in an Indonesian regional context that is underrepresented in the international planning literature. Second, it interrogates the assumption that informal relational resources are sufficient to sustain female workforce participation, thereby reorienting planning attention toward structural and institutional interventions. Third, it integrates social science methodology with sustainable development planning frameworks, demonstrating how PLS-SEM-based workforce research can inform regional gender-equality planning aligned with SDG 3, SDG 5, and SDG 8.

3.1 Research design

A quantitative correlational design was adopted, consistent with standard approaches in social science research for examining relationships between variables without manipulation [33]. This design is well suited to the generation of planning-relevant baseline evidence: it permits estimation of the proportion of variance in a workforce outcome explained by a set of modifiable predictors, thereby informing decisions about where planning resources are most likely to yield sustainable returns.

3.2 Participants and sampling

Participants were recruited using purposive sampling from the professional female workforce of West Sumatra Province, Indonesia [34]. Inclusion criteria required participants to be: (a) female; (b) engaged in full-time formal professional employment; (c) married or in a committed cohabiting partnership; and (d) resident in West Sumatra Province. The final sample comprised 60 career women spanning education, healthcare, and financial services sectors, recruited over a one-month period.

Given the use of purposive and snowball sampling combined with online survey distribution, the sample should be considered non-probability and convenience-based, which may limit the generalisability of the findings beyond similar professional groups.

The sample size of 60 should be interpreted with caution given the relatively high model complexity (seven predictors and eight latent constructs). While PLS-SEM is known for its tolerance of smaller samples [35], recent methodological guidance emphasises that model complexity must be balanced with statistical power. Therefore, this study is positioned as exploratory rather than confirmatory, aiming to provide initial evidence on source-specific social support in the West Sumatran context.

Table 1 presents the demographic profile of the participants, providing contextual information for interpreting the findings within this exploratory sample.

Table 1. Demographic profile of research participants (n = 60)

The sample is dominated by participants from the education sector, which should be considered when interpreting the results.

To address potential overparameterisation, additional robustness checks were conducted using simplified specifications (see Section 4.3), including aggregated support models and sensitivity analyses.

3.3 Instrumentation

Two validated psychometric instruments were administered. The Social Support Scale was adapted from Cohen’s multidimensional framework to assess support received from seven relational sources: SSH, SSC, SSP, SSEF, SSI, SSCL, and SSL [10]. Each subscale measured emotional, instrumental, informational, and appraisal support dimensions across 31 items rated on a semantic differential scale. The WFC Scale comprised 27 Likert-format items spanning time-based and strain-based WFC dimensions, adapted from Greenhaus’s framework [4]. Table 2 summarises the instrument structure.

Instrument development followed a systematic process: literature review to identify construct indicators, expert panel review for content validity, and Bahasa Indonesia translation with back-translation to ensure semantic equivalence. A pilot test with 20 career women external to the main sample confirmed item clarity prior to full-scale data collection.

Table 2. Research instrument overview

3.4 Data collection

Data were collected via a structured online survey administered through Google Forms over a one-month period in West Sumatra Province. Participants were contacted through institutional networks in the education, healthcare, and banking sectors. Informed consent was obtained electronically; all data were pseudonymised before analysis to ensure participant confidentiality. Snowball referrals were used to extend reach within specialist occupational groups.

3.5 Analytical strategy: Partial Least Squares Structural Equation Modelling

Data were analysed using Partial Least Squares Structural Equation Modelling (PLS-SEM) in SmartPLS 4.0 [36]. PLS-SEM was selected over covariance-based SEM because it imposes no multivariate normality assumption, performs reliably with small-to-medium samples, and is suited to exploratory research contexts in which theoretical models are not yet fully established [34]. Analysis followed a two-stage protocol: outer model (measurement) evaluation preceded inner model (structural) testing.

Measurement model adequacy was assessed via: (1) individual indicator loadings ≥ 0.70 for item-level convergent validity; (2) Cronbach’s Alpha (α) and Composite Reliability (CR) ≥ 0.70 for internal consistency; (3) Average Variance Extracted (AVE) ≥ 0.50 for construct-level convergent validity; and (4) the Fornell–Larcker criterion for discriminant validity [37]. Structural model evaluation focused on R² and path coefficients estimated via bootstrapping with 5,000 subsamples (significance threshold: T > 1.96, p < 0.05) [35].

5.1 Why social support does not significantly predict work-family conflict: Structural explanations

The rejection of all seven hypotheses, despite a psychometrically sound measurement model, points to theoretically important boundary conditions on the social support-WFC relationship. Rather than indicating a failure of social support theory, these findings reveal the conditions under which informal relational resources are insufficient to address WFC, a conclusion with direct implications for sustainable workforce development planning.

The most plausible structural explanation is demand dominance: when objective occupational demands workload, shift hours, performance targets, administrative burden deplete shared resources at a rate that exceeds what social support can replenish, the WHR model [11] predicts that support will have no detectable effect on WFC. This interpretation is consistent with Hosseini et al.’s [19] qualitative evidence that rigid work schedules and absence of formal childcare are primary WFC drivers in Islamic-majority cultural contexts comparable to West Sumatra, and with Zhao et al.’s [38] finding that work stress in the teaching profession generates conflict patterns that are not easily ameliorated by relational support alone. Given that 65% of the present sample are teaching staff, this sector-specific dynamic is particularly salient for workforce planners in West Sumatra.

The following interpretation should be understood as contextual and exploratory, as the present study did not directly measure cultural norms, support expectations, or internalised gender roles. Therefore, this explanation should be interpreted cautiously and warrants further empirical investigation.

A complementary explanation involves cultural support saturation. The Minangkabau matrilineal system embeds women within dense extended-family networks that are normatively obligated to provide assistance. When social support functions as a cultural constant rather than a variable resource, statistical variance in support levels is suppressed across the sample, making associations with outcome variables statistically undetectable a ceiling effect analogous to that documented in collective cultural contexts by Lakey and Orehek [39]. For planners, this implies that additional investment in informal support networks yields diminishing returns in this context; structural interventions are the marginal route to WFC reduction.

A third explanation concerns measurement scope. The model’s R² of 0.161 unambiguously indicates that 83.9% of WFC variance is explained by factors not represented in the model. It is important to note that structural factors such as workload, scheduling inflexibility, and institutional childcare availability were not directly measured or tested in this study; their salience is therefore inferred from the residual unexplained variance in R² and from convergent evidence in the broader empirical literature, not from direct empirical testing within the present data. Other unmeasured relational factors, such as community support networks or peer solidarity among professional women, may similarly explain portions of the unexplained variance and warrant inclusion in future models. Variables such as job autonomy, commuting burden, workplace flexibility policies, institutional childcare access, individual self-efficacy and emotional regulation capacity, and organisational gender climate are each theoretically and empirically relevant WFC predictors that were not included. The omission of these variables reflects a study design that was necessarily scoped to test the social support hypothesis; future model extensions must incorporate these structural and individual-resource dimensions to produce planning-actionable explanatory power.

5.2 Comparison with international literature

The divergence between the present findings and comparable international studies is theoretically informative. Noor et al. [28] reported that spousal support alone explained 43% of WFC variance among Indonesian bank employees a result dramatically more powerful than the non- significant SSH path (β = −0.180, p = 0.482) observed here. This discrepancy likely reflects suppressor dynamics: in the present model, six additional support sources are simultaneously entered, and the high inter-construct correlations visible in Table 5 introduce multicollinearity that inflates standard errors and attenuates individual path significance. The planning implication is that source- aggregated models overestimate the actionable impact of any single support source, reinforcing the need for multi-source designs.

The contrast with Noor et al. [28], who found a strong and significant effect of spousal support, represents one of the most theoretically important findings of this study. Rather than contradicting prior research, this discrepancy highlights the importance of model specification. When spousal support is examined in isolation, its effect may appear substantial. However, when multiple highly correlated sources of support are included simultaneously, the unique contribution of each source becomes difficult to isolate.

This suggests that prior findings may partly reflect shared variance across support sources, rather than purely independent effects. From a methodological standpoint, this underscores the importance of testing both isolated and combined models in social support research.

Kossek et al.’s [23] meta-analytic finding that work-family specific supervisory support reduces WFC more effectively than general support is not contradicted by the non-significant SSL path in the present study; rather, it highlights a measurement precision issue. The SSL subscale in this study captured general leader supportiveness rather than work-family specific supervisory facilitation behaviours the dimension most relevant to WFC reduction in Kossek et al. [23]’s analysis. Future workforce development research in this context should incorporate validated work-family specific supervisory support instruments to test this distinction.

The comparison with Pluut et al.’s [8] Dutch dual-buffering study is equally instructive. The additive support effects observed in that study occurred within a highly institutionalised context of statutory flexible-work rights, generous parental leave, and egalitarian gender norms structural conditions that effectively convert relational support into actionable WFC-reduction resources. The absence of equivalent institutional infrastructure in the present context renders relational support less functionally effective, even when subjectively perceived as available. This cross-national comparison demonstrates that sustainable workforce development planning cannot rely on informal support norms in the absence of formal institutional complements.

5.3 Gender norms, childcare, and internalised role conflict

Gender role theory [26] provides a further explanatory layer. Aarntzen et al. [27] demonstrated that working mothers who internalise gender stereotypes experience guilt-mediated WFC that persists even when practical support is available. In West Sumatra’s Minangkabau context where Islamic domestic norms overlay the matrilineal expectation of female household centrality career women may experience a cognitively internalised form of role conflict that no amount of interpersonal support can fully address without accompanying shifts in cultural norms and organisational expectations. Yuan et al. [40] similarly found that work-related stress produces chained burnout effects among Chinese female nurses that are mediated by WFC, suggesting that gender-specific occupational stress mechanisms require targeted, sector-specific interventions.

The parental profile of the sample is also significant: 78.33% of participants had at least one child, and 15% had three or more. The concrete time demands of parenting impose structural constraints that instrumental support can only partially offset. The composite WFC score used as the outcome variable in this study may mask differential support effects across WFC dimensions: instrumental support might buffer time-based conflict while emotional support addresses strain-based conflict, but aggregation into a single score obscures these distinctions. Disaggregation of WFC into its component dimensions in future research would provide more precise planning-relevant evidence.

In addition to structural explanations, a measurement-related limitation should also be considered. A key limitation of the present model is the aggregation of WFC into a single composite construct. Prior research distinguishes between time-based and strain-based conflict, as well as bidirectional forms (WIF vs FIW). It is plausible that different types of social support operate selectively across these dimensions. For example, instrumental support may reduce time-based conflict, while emotional support may buffer strain-based conflict.

The use of a global WFC score may therefore obscure meaningful relationships. Future studies should model these dimensions separately to improve explanatory precision.

5.4 Guidance, counselling, and workforce sustainability

While the present findings indicate that informal social support does not significantly attenuate WFC, they do not diminish the importance of professional guidance and counselling as a workforce sustainability intervention. Counselling services embedded within workplace employee assistance programmes can target the individual cognitive and emotional mechanisms internalised gender guilt, coping self-efficacy deficits, role boundary management skills that perpetuate WFC at a level that relational support cannot reach [41]. The quality and specificity of support provided may be improved by family counseling interventions that consider spouses, parents, and in-laws as active participants in role negotiation rather than just as passive support providers [42]. Group guidance services allowing career women to share strategies, normalise their experiences, and develop collective coping repertoires represent a resource-efficient complement to structural workforce planning interventions [43].

This study examined the correlational relationship between seven source-specific social support dimensions and WFC among 60 professional women in West Sumatra, Indonesia, using PLS-SEM with SmartPLS 4.0. The measurement model demonstrated strong psychometric quality throughout: all 30 indicator loadings exceeded 0.85, Composite Reliability values ranged from 0.897 to 0.986, and AVE values ranged from 0.814 to 0.946. Despite this methodological rigour, bootstrapping revealed that none of the seven social support dimensions from SSH, SSC, SSP, SSEF, SSI, SSCL, or SSL significantly influenced WFC (all T < 1.96; all p > 0.05). The collective R² of 0.161 indicates that social support accounts for only 16.1% of WFC variance, with 83.9% attributable to factors not captured in the model.

The theoretical contribution of these findings is twofold. First, they challenge the universality of the social support buffering hypothesis by demonstrating that its effectiveness is contingent on cultural baseline support levels, the nature of dominant structural stressors, and the availability of institutional complements that convert relational resources into actionable conflict reduction. In the Minangkabau socio-cultural context of West Sumatra, where social support may function as a cultural constant and where demand-side structural factors appear to dominate WFC determination, interpersonal relational resources operate below their theoretical buffering potential. Second, the findings integrate social support theory with sustainable workforce development planning discourse, repositioning WFC as a structural planning problem rather than a relational management challenge.

The practical contribution is a set of evidence-based recommendations for sustainable workforce development planning in West Sumatra aligned with SDGs 3, 5, and 8: flexible scheduling and workload governance in public-sector education and healthcare; institutional childcare provision and employer incentive frameworks; supervisor training in work-family specific supportive management; integration of professional counselling services within workplace employee assistance programmes; and incorporation of gender-disaggregated WFC indicators into provincial human development monitoring systems.

Several limitations bound these conclusions. The cross-sectional design precludes causal inference. The sample of 60, while adequate for PLS-SEM, restricts statistical power and limits generalisability beyond the study context. Self-report measurement introduces the possibility of common method bias. The composite WFC outcome masks potentially differential effects across time-based and strain-based conflict dimensions. The model’s low R² signals the importance of unmeasured structural, organisational, and individual-resource variables.

The findings should therefore not be interpreted as evidence that social support is unimportant, but rather that its effects are context-dependent, method-sensitive, and potentially obscured in multi-source models with limited sample sizes.

Future research should address these limitations through larger, longitudinal, mixed-method designs that integrate quantitative WFC measurement with qualitative exploration of structural and cultural mechanisms. Comparative studies across West Sumatra’s diverse occupational sectors and ethnic subgroups would yield contextually richer planning evidence. Extended models incorporating workload intensity, job autonomy, organisational climate, individual self-efficacy, and institutional policy variables would substantially improve explanatory power. Such research is essential for constructing an evidence base adequate to support the gender-responsive, sustainability-oriented workforce development planning that West Sumatra’s professional women need and that the SDG framework demands.