Bridging SROI and Marketing Theory in CSR Evaluation: Beneficiary Perceived Social Return as a Psychometric Construct among Indonesian CSR Beneficiaries

Introduction

Over the past two decades, corporations have undergone a fundamental shift in how they conceptualize their social responsibilities. Corporate Social Responsibility (CSR) is no longer merely a philanthropic activity conducted at a company’s own discretion; it has become an inseparable component of modern corporate strategy and governance (Fatima & Elbanna, 2023). Indonesia reinforced this shift through two binding legal instruments: Law No. 19 of 2003 on State-Owned Enterprises, which mandates the Partnership and Community Development Program, and Law No. 40 of 2007 on Limited Liability Companies, which requires CSR for companies operating in the natural resources sector. As a result, Indonesian SOE CSR investment has reached a nationally significant scale, intensifying demands to demonstrate that such investment genuinely produces tangible change for beneficiaries (Barnett et al., 2020; Suryawati et al., 2023).

The question of how to measure the social impact of CSR programs reliably and meaningfully remains an incompletely resolved challenge. The Social Return on Investment (SROI) methodology, whose foundations were established by Olsen, (2003) and standardized by Nicholls et al., (2012) , has long been the most widely adopted approach. SROI offers a systematic and communicable framework, converting social changes into monetary values that can be reported to stakeholders and compared across programs. However, several recent in-depth reviews have revealed limitations that cannot be readily dismissed. Arvidson et al., (2013) identified the attribution problem as the most fundamental weakness in SROI the difficulty of isolating the specific impact of one program from the concurrent influence of other factors. Corvo et al., (2022), reviewing 187 SROI publications spanning two decades, found that SROI-generated figures are potentially inconsistent because they depend heavily on the selection of comparison benchmarks that are difficult to standardize. Giorbelidze, (2025) and Nielsen et al., (2021) documented that the financial cost and technical complexity of conducting SROI render the methodology difficult to implement for organizations lacking large-scale capacity. Damtoft et al., (2023) demonstrated that SROI’s principal limitation lies in its external monetization logic, which is not designed to integrate beneficiary perception measurement into a causal model. It should be noted that SROI can incorporate beneficiary inputs; however, the resulting figures are still converted through an external valuation logic that constrains the integration of beneficiaries’ psychological perspectives (Corvo et al., 2022; Siegal, 2022). This constitutes the first gap motivating the present study: the structural limitation of SROI in measuring social impact from beneficiaries’ internal perspective. Siegal, (2022) characterized this as SROI’s deepest gap its inability to measure how deeply and meaningfully beneficiaries perceive change in a format suitable for elucidating the psychological mechanisms through which value and satisfaction are formed. Gutiérrez-Nieto et al., (2025) uncovered a noteworthy paradox: the marketing domain is the field that least frequently employs SROI. This is surprising, given that the core objectives of CSR programs creating value, building satisfaction, and forging meaningful long-term relationships with stakeholders align closely with central concepts in marketing such as customer value, satisfaction, and relationship marketing. This paradox signals a second gap: the limited application of marketing theory frameworks particularly value and satisfaction constructs in evaluating CSR programs from the perspective of community beneficiaries. Meanwhile, the marketing research tradition on consumer value and customer satisfaction has built an exceptionally rich theoretical framework over more than four decades. Zeithaml, (1988) established foundational insights into how individuals evaluate the value they receive. Oliver, (1980) explained how satisfaction forms through the comparison between expectations and perceived reality. Cronin & Taylor, (1992) demonstrated the causal chain among quality, value, and satisfaction a finding that remains among the most influential in the services marketing literature. Sweeney & Soutar, (2001) validated the multidimensional nature of consumer perceived value. Yet this body of frameworks has not received adequate systematic application in the context of CSR program beneficiaries. Studies integrating value and satisfaction constructs into CSR evaluation have generally focused on consumers or corporate customers not on community beneficiaries who invest their time, energy, and personal commitment in anticipation of substantive transformation in return (Araújo et al., 2023; Kim et al., 2020). This constitutes the third gap: the absence of a psychometric scale specifically designed to measure community CSR beneficiaries’ perceived transformation in a format integrable into survey-based causal marketing models. The foregoing literature review identifies three interrelated gaps requiring simultaneous attention. First, the structural limitation of SROI in measuring social impact from beneficiaries’ internal perspective, as identified by Siegal, (2022), Corvo et al., (2022), and Damtoft et al., (2023). Second, the limited application of marketing theory particularly value and satisfaction frameworks in evaluating CSR programs from the community beneficiary perspective (Gutiérrez-Nieto et al., 2025). Third, the unavailability of a validated scale capable of measuring CSR beneficiaries’ perceived transformation in a psychometric format integrable into causal models (Siegal, 2022). Prior efforts to bridge stakeholder perspectives with social impact measurement have been made, yet each leaves unfilled gaps. Within the service-dominant logic (SDL) framework, S. L. Vargo & Lusch, (2004); S. Vargo & Lusch, (2008) assert that value is always co-created and resides in use, meaning that the beneficiary perspective should be central to measuring social program impact. Yet SDL does not yield psychometric instruments capable of operationalizing beneficiary transformation experiences in a structurally testable format. Within the stakeholder theory tradition, Freeman, (2010) and Donaldson & Preston, (1995) advocate recognition of stakeholder perspectives including those of beneficiaries, and several CSR scholars Thampi et al., (2024) and Yadava & Sinha, (2021) have proposed measuring program impact from their standpoint. Yet these efforts remain at the conceptual level without producing psychometric constructs integrable into causal value and satisfaction models. This is the critical gap: although both SDL and stakeholder theory normatively acknowledge the importance of beneficiaries’ internal perspective, neither has produced measurement instruments capable of linking beneficiaries’ transformation experiences to value and satisfaction formation mechanisms within an empirically testable framework. BPSR is explicitly designed to fill this void. It is from these gaps that the central research question of this study emerges: is it possible to develop a psychometric construct capable of measuring beneficiaries’ perceptions of the self-transformation they experience as a result of CSR program engagement, and how does such a construct function as a mechanism explaining why CSR program implementation produces varying levels of value and satisfaction among beneficiaries? This study addresses that question by developing a new construct named Beneficiary Perceived Social Return (BPSR). SROI and BPSR are two distinct yet complementary approaches. SROI measures CSR program impact from the outside through monetary value, whereas BPSR measures it from the inside through what beneficiaries genuinely experience, using survey instruments conventional in marketing research. BPSR does not seek to replace SROI; rather, it fills the gap that SROI has been unable to reach, so that two research traditions that have long operated in isolation can finally be connected. This study makes three original, mutually reinforcing contributions. First, the construct contribution: this study develops and validates BPSR as a novel psychometric construct with an original measurement scale. Its conceptual novelty lies in operationalizing beneficiary-perceived transformation as an internal psychological dimension an entity explicitly distinct from the value construct (evaluative in nature) and from satisfaction (affective-confirmatory in nature). The methodological novelty lies in the availability of a total of four items representing four dimensions of transformation, allowing BPSR to be integrated into a survey-based structural model without requiring the SROI monetary valuation procedure. Second, the model contribution: this study proposes and tests a four-variable relational model with sequential dual mediation

that elucidates the psychological mechanisms linking CSR implementation quality to beneficiary satisfaction. Third, the contextual and practical contribution: these findings emerge from the context of Indonesian SOE CSR programs an institutionally unique setting given their regulatory mandate and yield a perception-based diagnostic instrument that CSR program managers can deploy directly within ongoing evaluation cycles, without the full financial and technical complexity of SROI.

Theoretical Background and Hypothesis

Social Exchange Theory

Social Exchange Theory (SET) was selected as the overarching grand theory for this research model, while several middle-range theories serve as the operational foundation for each construct: Bandura, (1991) Social Cognitive Theory and Sen, (1999) Capability Approach for the BPSR transformation dimensions; Ajzen, (1991) Theory of Planned Behavior for the prosocial dimension; and Oliver, (1980) Expectation Disconfirmation Theory and Zeithaml, (1988) Means-End Model for BPV and BSAT. SET was chosen as the primary theory for its straightforward yet comprehensive capacity to explain the exchange relationship between a company as CSR program provider and the beneficiary community. Rooted in the works of Blau, (1964) and Emerson, (1976), SET posits that every social interaction is fundamentally an exchange process: each party offers something to the other in expectation of receiving something in return. Unlike economic exchanges, whose obligations are formal and contractual, social exchange is grounded in reciprocity norms that develop naturally as trust grows between the parties (Cropanzano & Mitchell, 2005). Cropanzano & Mitchell, (2005) identified two SET mechanisms most relevant to this research model. The first is the quality-of-exchange principle: the higher the quality of the program as perceived by beneficiaries in terms of relevance, transparency, consistency, and fairness the stronger and more positive the beneficiary response. The second is the generalized norm of reciprocity: recipients of high-quality programs tend not only to reciprocate directly to the provider but also feel compelled to contribute to the broader community as an expression of shared goodwill. Ahmad et al., (2023), in a recent systematic review of SET, confirmed that both mechanisms are increasingly relevant in the context of organizational community exchange in the contemporary era. In this research model, SET functions as the connective thread linking all four variables in a logically consistent manner. High-quality CSR program implementation (CSRI) represents a high-value social exchange offering from the company to beneficiaries. Beneficiary-perceived selftransformation (BPSR) is the response manifested as meaningful self-change resulting from that exchange. Positive value evaluation (BPV) is the response reflecting a judgment that the exchange is beneficial and meaningful to one’s life. Beneficiary satisfaction (BSAT) is the affective response the feeling of contentment regarding the overall exchange experience. Consequently, all ten hypothetical pathways in the model can be read and understood within a single, internally coherent SET narrative.

CSR Implementation (CSRI)

CSRI is defined in this study as the level of quality of CSR program implementation as perceived by beneficiaries, encompassing four core dimensions: program relevance to beneficiaries’ actual needs, process transparency, program consistency and sustainability, and fairness of implementation toward all participants (Fatima & Elbanna, 2023). These four dimensions rest on Carroll, (1991) CSR Pyramid, which identifies the hierarchy of corporate responsibilities across economic, legal, ethical, and philanthropic dimensions a conceptual foundation that has been broadly tested and recognized for more than three decades. Fatima & Elbanna, (2023), in their review of more than 200 CSR studies, found that program implementation quality is far more determinative of social impact than the magnitude of financial investment. Turker, (2009) established that transparency is the strongest goodwill signal distinguishing genuine from merely image-driven CSR programs. Islam et al., (2021) confirmed that fairness in benefit distribution is the most consistently robust factor in building beneficiary trust across diverse cultural contexts. Barnett et al., (2020) demonstrated that CSR programs genuinely designed to create social value produce measurably different impacts compared with those conducted solely to satisfy regulatory obligations.

Beneficiary Perceived Social Return (BPSR): A Novel Construct

BPSR is defined as beneficiaries’ perceptions of positive and meaningful self-transformation, experienced directly as a result of their engagement in a CSR program. The construct is built through the synthesis of four theoretical traditions, each explaining a distinct yet complementary dimension of transformation, yielding a measurement instrument comprising four original, theoretically grounded items. BPSR is specified as a reflective construct because its four indicators represent different manifestations of the same underlying latent construct namely, beneficiaries perceived self-transformation. The four items are expected to intercorrelate, given that an individual who experiences holistically meaningful selftransformation tends to undergo it simultaneously across behavioral, competence, social-attitude, and communityparticipation dimensions. The indicators are therefore interchangeable as reflections of the same latent construct. The first dimension is Behavioral Self-Regulation Return, grounded in Bandura, (1991) Social Cognitive Theory. Engagement in a high-quality CSR program encourages beneficiaries to observe changes in their own behavior, evaluate those changes against learned standards, and sustain them through intrinsically generated motivation. The second dimension is Applied Competence Return, grounded in Sen, (1999) Capability Approach. An effective program not only equips beneficiaries with new capabilities but also encourages them to actually apply those capabilities in real everyday life, as further operationalized by (Lent et al., 1994). The third dimension is Prosocial Awareness Return, grounded in Ajzen, (1991) Theory of Planned Behavior. A highquality CSR program not only alters specific beneficiary behaviors but also more profoundly transforms their perspective on and concern for broader social and environmental issues. Ahmad et al., (2023) demonstrated that attitudinal changes of this kind are the most enduring responses generated by a high-quality social exchange. The fourth dimension is Community Engagement Return, grounded in Cropanzano & Mitchell, (2005) norm of reciprocity and Lins et al., (2017) concept of social capital reinforcement. A meaningful CSR program motivates beneficiaries not merely to enjoy the benefits received but also to actively contribute to the surrounding community, thereby gradually building social capital whose benefits extend beyond the program itself. Conceptually, BPSR is positioned as a complement to SROI, not a replacement. SROI measures social impact from an external perspective through monetary value, while BPSR measures it from an internal perspective through what beneficiaries genuinely experience. Corvo et al., (2022) and Siegal, (2022) have explicitly identified the absence of such an approach as the largest gap in the social impact measurement literature. Within S. L. Vargo & Lusch, (2004) service-dominant logic framework, BPSR can be understood as an antecedent or prerequisite for the formation of value-in-use: BPSR measures

Beneficiary Perceived Value (BPV)

BPV is defined as beneficiaries’ holistic evaluation of how meaningful the CSR program they participate in is. This evaluation is formed through the trade-off between the benefits received and the non-financial sacrifices they have invested namely time, energy, and personal commitment (Sweeney & Soutar, 2001; Zeithaml, 1988). The benefits in question encompass three simultaneous dimensions: tangible and measurable functional benefits, social benefits in the form of a sense of meaningfulness for the community, and emotional benefits in the form of personally touching experiences. Zeithaml, (1988) explained that value evaluation forms through a sequential cognitive process: beneficiaries first assess program attributes, then experience the consequences, and ultimately arrive at a genuine value judgment. This sequential process is precisely what positions BPSR as an indispensable layer in the pathway from program quality to beneficiary value evaluation. Sweeney & Soutar, (2001) demonstrated that consumer perceived value is not unidimensional but encompasses three complementary dimensions: functional, social, and emotional. Meanwhile, McDougall & Levesque, (2000) established that the perceived fairness of the trade-off between benefits received and sacrifices made is the most consistently reliable predictor of an individual’s satisfaction with a service. Ghanbarpour et al., (2024) provided relevant findings in the context of transformative, socially responsible services: the social and emotional dimensions of perceived value proved to be stronger predictors than the functional dimension. Although their research context involved broader corporate stakeholders rather than CSR beneficiaries exclusively, these findings are consistent with the CSR program context, whose primary purpose is not merely to deliver practical benefits but to transform beneficiaries’ lives in a more comprehensive and meaningful manner. Accordingly, the same logic can be applied in this research model to understand how perceived value contributes to CSR beneficiary satisfaction.

Beneficiary Satisfaction (BSAT)

BSAT is defined as beneficiaries’ evaluative response to the overall CSR program experience, encompassing three simultaneous dimensions. The first is cognitive: whether the program meets or exceeds beneficiaries’ initial expectations. The second is affective: the extent to which beneficiaries emotionally enjoy the program implementation process. The third is volitional: beneficiaries’ conviction that their decision to participate was the right one (Oliver, 1997; Westbrook & Oliver, 1991). Its primary theoretical foundation is Oliver, (1980) Expectation-Disconfirmation Theory, which posits that satisfaction arises when program performance exceeds beneficiary expectations, while dissatisfaction arises when performance falls below previously formed expectations. Cronin & Taylor, (1992) established that satisfaction does not form directly from program quality alone but through intervening mechanisms namely, the value and transformation perceived by beneficiaries beforehand. Sirdeshmukh et al., (2002) confirmed that perceived value is the primary mechanism linking service quality to satisfaction and loyalty within an ongoing exchange relationship. The meta-analysis by Coelho do Vale et al., (2025) across more than one hundred CSR studies further reinforces this conclusion: perceived value is consistently demonstrated to mediate the causal chain between CSR programs and beneficiary satisfaction, across diverse industries and cultural contexts.

Conceptual Framework and Hypothesis Development

Building on the theoretical foundations elaborated above, this study develops a four-variable causal model with ten hypothetical pathways. Figure 1 illustrates the conceptual framework comprehensively, depicting the configuration of six direct effects (H1–H6) and four mediated effects (H7–H10).

Corporate Social Responsibility Implementation (CSRI) and Beneficiary Perceived Social Return (BPSR)

H1 is grounded in the quality-of-exchange principle within Social Exchange Theory. A CSR program satisfying all four implementation quality dimensions creates the most conducive conditions for the formation of beneficiary self-transformation as measured by BPSR. Program relevance to beneficiaries’ genuine needs fosters mastery experiences as explained by Bandura, (1991). Transparency builds trust, which in turn reinforces the norm of reciprocity between the company and beneficiaries. Consistency provides adequate time for beneficiaries to truly actualize the new capabilities they have acquired, as intended by Sen, (1999). Fair implementation strengthens a sense of ownership that motivates beneficiaries to contribute more actively to their community. Rudito et al., (2023) confirmed in the Indonesian SOE context that community empowerment programs implemented in earnest demonstrably produce real, measurable changes in beneficiary capacity. H1: CSRI has a positive and significant effect on BPSR

Corporate Social Responsibility Implementation (CSRI) and Beneficiary Perceived Value (BPV)

Each of the four CSR implementation quality dimensions makes a direct contribution to beneficiary value evaluation. Program relevance to beneficiaries’ actual needs directly augments perceived functional benefits while reinforcing the conviction that the time and energy invested are commensurate with the benefits obtained. Transparent and fair implementation builds a sense of social meaningfulness because beneficiaries feel treated with respect and equity. Consistent implementation cultivates confidence in the company’s genuine commitment, ultimately producing positive and meaningful emotional experiences. Araújo et al., (2023) and Kim et al., (2020) confirmed that beneficiaries’ perceptions of CSR program quality are consistently and directly related to their value evaluations, across diverse industries and contexts. H2: CSRI has a positive and significant effect on BPV

Beneficiary Perceived Social Return (BPSR) and Beneficiary Perceived Value (BPV)

H3 tests the role of BPSR as the previously invisible link in the chain between CSR program attributes and beneficiary value evaluation as explained by Zeithaml, (1988). Beneficiaries do not jump directly from assessing program quality to evaluating its value. There is one critical intervening stage: real, directly experienced self-transformation. Beneficiaries who perceive themselves as more disciplined, more competent, more attentive to their surroundings, and more active in their community have a far stronger basis for concluding that the program is genuinely of high value to their lives. Ghanbarpour et al., (2024) demonstrated that in transformative services, perceived transformation is the most powerful predictor of perceived value, directly supporting this hypothesis. H3: BPSR has a positive and significant effect on BPV

Corporate Social Responsibility Implementation (CSRI) and Beneficiary Satisfaction (BSAT)

A CSR program perceived by beneficiaries as relevant, transparent, consistent, and fair directly elicits feelings of satisfaction without requiring prior value/benefit evaluation. This occurs because the four CSR implementation quality dimensions naturally trigger conditions in which program performance exceeds beneficiaries’ initial expectations, as explained by Oliver, (1980). Beneficiaries who perceive that the program genuinely addresses their needs, is conducted honestly, is consistently present, and is implemented fairly will spontaneously feel satisfied even before evaluating the program’s overall value. Islam et al., (2021), Lee et al., (2020), and Rahmawati & Kustiawan, (2024) consistently confirmed that CSR implementation quality has a direct effect on beneficiary satisfaction across diverse service contexts. H4: CSRI has a positive and significant effect on BSAT

Beneficiary Perceived Social Return (BPSR) and Beneficiary Satisfaction (BSAT)

Satisfaction does not always form through a lengthy, deliberative cognitive process. Westbrook & Oliver, (1991) demonstrated that satisfaction can also arise spontaneously as a direct emotional response to a transformative experience. Beneficiaries who perceive real changes in themselves becoming more disciplined, more competent, more attentive, and more active in their community will experience satisfaction directly from that transformation without first needing to calculate and evaluate the program’s overall value. Satisfaction arising through this pathway is more personal, more profound, and more authentic because it originates from genuinely felt self-change. Muflih, (2021) and Glaveli, (2021) confirmed that this mechanism operates consistently in CSR and social service contexts. H5: BPSR has a positive and significant effect on BSAT.

Beneficiary Perceived Value (BPV) and Beneficiary Satisfaction (BSAT)

H6 is the hypothesis with the strongest empirical support in this model. The relationship between perceived value and satisfaction has been one of the most consistently and extensively confirmed findings in the marketing literature for more than three decades. Cronin & Taylor, (1992) were the first to establish definitively that consumer perceived value is the primary determinant of satisfaction even stronger than the direct effect of service quality on satisfaction itself. Sirdeshmukh et al., (2002) and McDougall & Levesque, (2000) reinforced this finding across diverse relational exchange contexts. Chi & Phan, (2026) confirmed it in the context most relevant to the present study: the relationship between CSR and customer loyalty. This long, consistent chain of confirmation makes H6 a robust empirical foundation within this research model. H6: BPV has a positive and significant effect on BSAT.

The Mediating Role of Beneficiary Perceived Social Return (BPSR)

H7 and H8 both test the role of BPSR as an indispensable mechanism in the pathway from CSR program quality to beneficiary value evaluation and satisfaction. Zeithaml, (1988) explained that value evaluation does not form directly from program attributes but must first pass through the stage of perceived consequences and this is precisely BPSR’s position in the model. In other words, high-quality CSR programs do not automatically yield optimal value judgments among beneficiaries without first producing real and meaningful selftransformation. Furthermore, Oliver, (1997) demonstrated that the deepest and most enduring satisfaction forms not from the assessment of program attributes alone but from the selfchange genuinely experienced by beneficiaries. Programs that fail to produce real self-transformation will only generate superficial and short-lived satisfaction. BPSR is the mechanism that transforms such satisfaction into something deeper, more authentic, and more stable. H7: BPSR significantly mediates the effect of CSRI on BPV. H8: BPSR significantly mediates the effect of CSRI on BSAT.

The Mediating Role of Beneficiary Perceived Value (BPV)

H9 and H10 both test the role of BPV as the mechanism linking antecedents to beneficiary satisfaction. Cronin & Taylor, (1992) long demonstrated that satisfaction does not form directly from program quality but must first pass through value evaluation, and Coelho do Vale et al., (2025), in their metaanalysis across more than one hundred CSR studies, confirmed that this pattern holds consistently across diverse contexts and industries. Moreover, the self-transformation perceived by beneficiaries through BPSR does not automatically and directly yield deep and stable satisfaction. Such transformation must first be processed into an explicit judgment about how meaningful and valuable the program is to beneficiaries’ lives, and it is from that evaluation that comprehensive and lasting satisfaction can form. McDougall & Levesque, (2000) and Lee et al., (2020) provide the empirical basis supporting this mechanism. Accordingly, beneficiary satisfaction forms through two complementary pathways: the direct emotional pathway from self-transformation tested in H5, and the elaborative pathway through the crystallization of transformation into value evaluation tested in H10. H9: BPV significantly mediates the effect of CSRI on BSAT. H10: BPV significantly mediates the effect of BPSR on BSAT.

Methods

Research Design

This study employs a quantitative approach with a causal associative design and cross-sectional data collection. It should be emphasized that the term “causal” here reflects the theoretically grounded direction of relationships among variables, not experimental evidence. Because data were collected at a single point in time, the chronological sequence of events between variables cannot be determined. Accordingly, follow-up studies measuring these variables at multiple time points are necessary to strengthen conclusions regarding the identified mediation mechanisms. This design was selected because the research model entails two methodological complexities requiring simultaneous attention: first, the development and validation of a new construct (BPSR) for which no previously validated scale exists in the literature; and second, the testing of dual mediation effects within an integrated structural model. This study was conducted in the context of the PLN Peduli CSR program managed by PT Perusahaan Listrik Negara

(Persero), one of Indonesia’s largest state-owned enterprises. Nationally, the PLN Peduli program spans several empowerment domains including economic, social, environmental, and governance dimensions (Persero, n.d.). It must be emphasized that the sample was drawn from PLN Peduli beneficiaries in West Kalimantan and is not intended to represent the program at the national level. Respondents participated in economic empowerment, skills training, and social-environmental activities organized by PLN’s West Kalimantan Main Distribution Unit in 2024 and 2025. Three primary reasons guided this contextual selection. First, PLN Peduli programs in West Kalimantan cover multiple districts and cities with diverse beneficiaries, enabling the sample to reflect adequate variation. Second, as an SOE operating under a strong regulatory CSR mandate, PLN constitutes a relevant and appropriate representation of the phenomenon under investigation. Third, the accessibility of data and beneficiary respondents in this region supported the required data collection quality.

Population and Sample

The research population consisted of beneficiaries of the PLN Peduli CSR program who had completed program participation. This criterion was established to ensure that beneficiaries had accumulated sufficient experience to meaningfully evaluate all four BPSR dimensions. Data were collected using questionnaires distributed to 320 PLN Peduli program beneficiary respondents in West Kalimantan. A total of 311 completed questionnaires were returned (response rate: 97.2%), of which 305 were suitable for analysis. Sample size was determined through two hierarchically applied criteria. First, statistical power analysis via G*Power 3 Faul et al., (2007) with medium effect size (f² = 0.15), α = 0.05, and power = 0.95 yielded a minimum of 119 respondents. Second, Hair et al., (2019) minimum recommendation of 200 respondents for PLS-SEM models. The number 305 meets and exceeds both criteria simultaneously, providing adequate confidence in the stability and sufficiency of the estimates for the PLS-SEM analysis in the context of this study. Non-probability purposive sampling was employed, with the following inclusion criteria: (1) respondents must be direct beneficiaries of the PLN Peduli program in West Kalimantan; (2) they must have participated in the program for at least one complete cycle; and (3) they must be willing to voluntarily complete the questionnaire fully and honestly. Respondents not meeting all three criteria were excluded. Recruitment was conducted by visiting PLN Peduli CSR program sites directly in each area, coordinating with field facilitators from PLN UID West Kalimantan who provided access to participant lists. Because non-probability purposive sampling was employed, the findings are not intended to be statistically generalized to all PLN Peduli beneficiaries PLN Peduli nationally, but rather to present preliminary evidence within the context of regulatory-mandated SOE CSR programs in Indonesia.

Research Location

This study was conducted across West Kalimantan, spanning Pontianak City, Sintang Regency, Singkawang City, Sekadau Regency, Sanggau Regency, and Ketapang Regency. These locations were selected because they constituted the target areas for PLN Peduli CSR program implementation by PT PLN’s West Kalimantan Main Distribution Unit (UID) in 2024 and 2025. It should be emphasised that nationally, the PLN Peduli program operates in various regions of Indonesia with a much broader scope, as stated in the publicly available corporate social responsibility report of PT PLN (Persero) (Persero, n.d.). This research is not intended to represent the national scope of the PLN Peduli program; the findings of this study are only applicable in the context of the implementation of the PLN Peduli program in the West Kalimantan region. The selected locations reflect: (1) a representative diversity of geographical and demographic characteristics enabling adequate capture of variation in beneficiary contexts; and (2) the fact that all beneficiaries across the six areas belong to the same program ecosystem PLN Peduli ensuring that program implementation quality, the antecedent variable in the research model, can be compared consistently and equitably across locations.

Measurement Instrument

All variables were measured using reflective constructs on a five-point Likert scale (1 = Strongly Disagree; 5 = Strongly Agree). The instrument comprised 16 items distributed across four constructs, each measured by four indicators. Table 1 presents the complete operationalization of all research variables. The five-point Likert scale assessed respondents’ attitudes, perceptions, and opinions. Respondents were asked to answer based on their actual experience. To minimize potential bias, the questionnaire was designed without respondent identification, with item ordering that does not steer respondents toward particular response patterns, and using neutral phrasing that does not lead toward any particular answer. The reflective specification for BPSR is grounded in the conceptual consideration that its four dimensions (Behavioral Self-Regulation Return, Applied Competence Return, Prosocial Awareness Return, and Community Engagement Return) are treated as manifestations of a single latent construct holistically perceived self-transformation. Although the four dimensions rest on different theoretical bases, they share the property of being internal beneficiary responses to engagement in a high-quality CSR program, such that change in one dimension tends to correlate with change in the others. This reflective specification is also empirically supported by the high correlation between BPSR items (all outer loadings ≥ 0.816; AVE = 0.751), which is consistent with the reflective and not formative specification. The formative composite alternative has been considered; however, given the theoretical unity of the underlying construct (one self-transformation perception) and the high observed convergence among items, the reflective model has been maintained as the most theoretically appropriate choice. As a robustness check, cross-loading analysis was conducted to ensure that all BPSR indicators loaded higher on their own construct than on any other in the model, consistent with the interchangeable nature of indicators in reflective specifications.

Data Collection Procedure

Data were collected through face-to-face surveys (offline questionnaires) and online surveys (Google Forms) using structured questionnaires distributed in 2025. Prior to completing the questionnaire, respondents received an explanation of the research purpose, data confidentiality assurances, a consent form, and instructions ensuring consistent comprehension. A cross-sectional survey approach targeted community members who were PLN Peduli CSR program beneficiaries in 2024 and 2025, residing in six districts/cities in West Kalimantan: Pontianak City, Singkawang City, Sanggau Regency, Ketapang Regency, Sekadau Regency, and Sintang Regency. Respondents answered all items and provided relevant demographic information. This process yielded 305 complete and valid questionnaires from PLN Peduli CSR program beneficiaries across the six locations, subsequently used to test the research model using PLS-SEM. To address potential mode effects (differences between face-to-face and online survey respondents), the researchers ensured uniformity of instructions and question format across both data collection modes. As a procedural control for common method variance (CMV), questionnaire item ordering was designed to spatially

Table 1. Research Variable Operationalization No. Variable Definition of Variables Dimension Measurement Item (Likert Scale 1–5) Main Reference 1. CSR Implem entatio n (CSRI)

CSR Implementation (CSRI)

is the level of quality of CSR program implementation as perceived by beneficiaries, encompassing need relevance, transparency, consistency, and fairness (Fatima & Elbanna, 2023). Philanthropic Responsibility: need alignment 1. This CSR program addresses my actual needs. Carroll, (1991); Fatima & Elbanna, (2023) Ethical Responsibility: process transparency 2. This CSR program is implemented transparently. Carroll, (1991; Turker, (2009) Economic/Legal Responsibility: program consistency 3. This CSR program is implemented consistently and sustainably. Carroll, (1979); Fatima & Elbanna, (2023) Ethical Responsibility: distributive fairness 4. This CSR program is implemented fairly for all participants. Carroll, (1991); Islam et al., (2021) 2. Benefici ary Perceiv ed Social Return (BPSR) Beneficiary Perceived Social Return (BPSR) refers to beneficiaries’ perceptions of positive self-transformation experienced as a result of engagement in a CSR program, encompassing behavioral regulation, competence actualization, prosocial orientation, and community participation (Novel construct). Behavioral SelfRegulation Return 1. Since participating in this CSR program, I have become more disciplined in my daily activities. Originally developed from a synthesis of Bandura, (1991); Cropanzano & Mitchell, (2005) Applied Competence Return 2. Since participating in this CSR program, I apply the new practices I have acquired from it. Originally developed from a synthesis of Lent et al., (1994); Sen, (1999) Prosocial Awareness Return 3. Since participating in this CSR program, I have become more attentive to social and environmental issues. Originally developed from a synthesis of Ahmad et al., (2023); Ajzen, (1991) Community Engagement Return 4. Since participating in this CSR program, I have become more actively involved in contributing to my community. Originally developed from a synthesis of Cropanzano & Mitchell, (2005); Lins et al., (2017) 3. Benefici ary Perceiv ed Value (BPV) Beneficiary Perceived Value (BPV) is beneficiaries’ holistic evaluation of the meaningfulness of a CSR program based on the tradeoff between benefits received (functional, social, emotional) and nonfinancial sacrifices invested (Zeithaml, 1988; (Sweeney & Soutar, 2001) Functional Value 1. This CSR program provides benefits that are genuinely useful for my life. Sen, (1999); Sweeney & Soutar, (2001) Social Value 2. My participation in this CSR program makes me feel part of something beneficial to many people. Sheth et al., (1991); Sweeney & Soutar, (2001) Emotional Value 3. Participating in this CSR program provides an emotionally meaningful experience for me. Sweeney & Soutar, (2001); Westbrook & Oliver, (1991) Value-for-Sacrifice 4. Overall, this CSR program delivers value commensurate with the time and effort I have invested. McDougall & Levesque, (2000); Zeithaml, (1988) 4 Benefici ary Satisfac tion (BSAT) Beneficiary Satisfaction (BSAT) is beneficiaries’ multidimensional evaluative response multidimensional beneficiaries of the overall CSR program experience, including cognitive (confirmation of expectations), affective (emotional response of the process), and volitional (affirmation of decision) dimensions (Oliver, 1980; (Westbrook & Oliver, 1991) Overall Satisfaction 1. Overall, I am satisfied with the outcomes I have obtained from this CSR program. Oliver, (1997) Expectation Confirmation 2. This CSR program has met my expectations regarding the benefits I would receive. Oliver, (1980) Affective Process Satisfaction 3. I enjoyed every stage of the implementation of this CSR program. Westbrook & Oliver, (1991) No-Regret Satisfaction 4. I feel that my decision to participate in this CSR program was the right one. Zeithaml et al., (1996) Source: Developed from various references. The BPSR construct was developed originally based on a synthesis of the literature by the researchers.

separate antecedent and consequent indicators, and respondents were assured that there were no right or wrong answers to reduce response bias. As a statistical check, the full collinearity VIF was calculated for all constructs following the CMV diagnostic procedure proposed by Kock, (2015). The Full Collinearity VIF values for all constructs are below 3.3, namely BPSR at 2.279, CSRI at 2.381, BSAT at 2.557, and BPV at 2.683, which provides initial assurance that CMV does not substantially threaten the interpretability of the model.

Data Analysis

Data analysis employed Partial Least Squares Structural Equation Modeling (PLS-SEM) using SmartPLS 4 software (Hair, Ringle, et al., 2022). The selection of PLS-SEM was based on two considerations. First, PLS-SEM is more appropriate for the development and testing of a new construct (BPSR) that does not yet have an established distribution. Second, PLS-SEM is optimal for complex causal models containing sequential dual mediation (Chin, 1998; Hair et al., 2017). Data analysis proceeded in two sequential stages. The first stage was measurement model evaluation, aimed at verifying that all indicators and constructs met the required measurement quality standards. This stage assessed each indicator’s outer loading, construct reliability through Composite Reliability and Cronbach’s alpha, convergent validity through AVE Fornell & Larcker, (1981), and discriminant validity through the HTMT criterion Henseler et al., (2014) and cross-loading analysis. The second stage was structural model evaluation, aimed at testing the strength and significance of inter-construct relationships. This stage assessed R² values, effect size f² Cohen, (1988), multicollinearity through VIF, pathway significance through bootstrapping with 5,000 subsamples, predictive relevance through PLSpredict Shmueli et al., (2019), and significance of mediation and serial mediation effects through specific indirect effects bootstrapping (Preacher & Hayes, 2008).

Ethical Approval

This study was conducted in accordance with applicable research ethics principles. All respondents participated voluntarily after receiving adequate explanation of the research objectives, benefits, and their right to withdraw at any time without consequence. Respondents’ identities were fully protected, and the collected data were used solely for academic purposes. Formal ethical approval is not required for this research because this questionnaire-based survey research has minimal risk characteristics for respondents, does not involve invasive procedures, sensitive personal data, or potential physical or psychological harm. All data collection procedures were carried out with due regard for the comfort and dignity of each individual involved.

Result and Discussion

Respondent Profile

A total of 305 respondents were included in the sample (see table 2). Respondents were individual beneficiaries of the PLN Peduli CSR Program, representing diverse age groups, genders, educational levels, and occupational backgrounds. This diversity reflects the variation of characteristics present within the program context studied, although it is not claimed to be a statistical representation of the entire national PLN Peduli beneficiary population, given the purposive sampling technique employed.

Outer Loadings and Convergent Validity

Measurement model evaluation (outer model) was conducted to examine the validity and reliability of all constructs, encompassing convergent validity, reliability, and discriminant validity assessments. This stage was completed prior to structural model evaluation, as measurement quality constitutes the primary foundation for interpreting analytical results. Convergent validity was assessed through two criteria: each indicator’s outer loading and each construct’s Average Variance Extracted (AVE). An indicator is deemed convergently valid when outer loading ≥ 0.70, and a construct meets convergent validity when AVE ≥ 0.50 (Hair, Hult, et al., 2022). Table 3 presents the comprehensive measurement model evaluation results. All indicators across the four constructs recorded outer loading values above 0.70. The highest outer loading was obtained for BPSR2 at 0.890, while the lowest was recorded for BSAT3 at 0.716. All indicators therefore satisfy the convergent validity requirement at the indicator level. AVE values for all constructs meet the convergent validity threshold, ranging from 0.654 (BPV) to 0.751 (BPSR). Notably, BPSR recorded the highest AVE among the four constructs (0.751), indicating that BPSR indicators measure their construct with the strongest convergence. Construct reliability was assessed using two measures: Cronbach’s Alpha and Composite Reliability (CR/rho_c). A construct is deemed reliable when Cronbach’s Alpha ≥ 0.70 and CR ≥ 0.70 (Hair, Ringle, et al., 2022). All constructs recorded Cronbach’s Alpha and CR values exceeding the minimum threshold of 0.70. Cronbach’s Alpha ranged from 0.822 (BPV) to 0.890 (BPSR), while CR ranged from 0.882 (BPV) to 0.923 (BPSR). These results confirm adequate internal consistency reliability: all constructs exhibit good internal consistency and are dependable as measurement instruments. Cut-off: Outer Loading ≥ 0.70; CR ≥ 0.70; AVE ≥ 0.50 (Hair, Hult, et al., 2022).

Discriminant Validity

Discriminant validity examines the extent to which each construct is genuinely empirically distinct from the others. Assessment was conducted using three criteria: cross loadings, the Fornell-Larcker Criterion, and the HeterotraitMonotrait Ratio of Correlations (HTMT). The cross loading criterion requires that each indicator’s loading on its own construct must exceed its loading on any other construct. Table 4 confirms that all 16 indicators consistently satisfied this criterion. All indicators recorded the highest loading values on their respective constructs, confirming that discriminant validity based on cross loadings is met. The second discriminant validity assessment applied the Fornell-Larcker Criterion (Henseler et al., 2014), which requires that the square root of AVE (√AVE) for each construct must exceed its correlation with any other construct. The √AVE values of all constructs exceed their correlations with other constructs. The √AVE of BPSR (0.867) is greater than BPSR’s correlations with BPV (0.706), BSAT (0.709), and CSR (0.617). The same condition holds for the other three constructs. Discriminant validity based on the Fornell-Larcker Criterion is thus confirmed (see Table 5). The third discriminant validity assessment employed the HTMT criterion (Henseler et al., 2014). All HTMT values fall below the conservative threshold of 0.85, confirming that the four constructs are genuinely empirically distinct entities (Table 6). This finding carries particular significance for BPSR: the confirmation that BPSR is distinct from BPV provides strong initial evidence that BPSR is not merely a relabeling of BPV but measures a genuinely different psychological entity self-transformation versus value evaluation.

Structural Model Evaluation (Inner Model)

Having established the validity and reliability of the measurement model, the next step involved evaluating the structural model (inner model). This evaluation encompassed: collinearity assessment, coefficient of determination (R²), effect size (f²), predictive relevance (Q²), and hypothesis testing through bootstrapping.

Multicollinearity, R Square, and Effect Size

Table 7 presents a comprehensive summary of structural model statistics. All structural VIF values confirm the absence of multicollinearity problems that could threaten the stability of path coefficient estimates. All inner VIF values fall below 3.3: CSRI as predictor of BPSR = 1.000; CSRI as predictor of BPV = 1.616; BPSR as predictor of BPV = 1.616; CSRI as predictor of BSAT = 2.191; BPSR as predictor of BSAT = 2.097; and BPV as predictor of BSAT = 2.708. R² values indicate meaningfully varying predictive power across constructs. BPSR is explained at 38.1% by CSRI, categorized as moderate per Hair et al. (2022) (0.25–0.50). BPV is explained at 63.1% by CSRI and BPSR, categorized as substantial. BSAT is explained at 63.2% by CSRI, BPSR, and BPV, also substantial. The high R² values for BPV and BSAT confirm that the proposed causal model possesses strong predictive power and that all included variables make meaningful explanatory contributions. The f² effect size profile reveals a highly informative influence hierarchy. Two pathways are categorized as large (f² ≥ 0.35): CSRI→BPSR (f² = 0.616), the pathway with the largest effect size in the model, followed by CSRI→BPV (f² = 0.356). Two pathways are categorized as medium (0.02 ≤ f² < 0.35): BPSR→BPV (f² = 0.298) and BPSR→BSAT (f² = 0.155). Two pathways are categorized as small (f² < 0.15): CSRI→BSAT (f² = 0.107) and BPV→BSAT (f² = 0.067). The medium-to-small effect size pattern on pathways toward BSAT indicates that direct effects on BSAT operate largely through mediation mechanisms, consistent with the causal chain of (Cronin & Taylor, 1992). VIF categorization: < 3.3 = no multicollinearity Categorization R²: < 0.25 = weak; 0.25–0.50 = moderate; 0.50–0.75 = substansial; > 0.75 = strong (Hair, Ringle, et al., 2022). Categorization f²: ≥ 0.35 = big; 0.15–0.35 = medium; 0.02– 0.15 = small (Cohen, 1988). Predictive Relevance (Q²predict) Table 4. Discriminant Validity, Cross Loadings (→ indicates loading on target construct) Indicator CSRI BPSR BPV BSAT CSR1 0.780 0.380 0.533 0.443 CSR2 0.870 0.599 0.666 0.608 CSR3 0.849 0.495 0.621 0.648 CSR4 0.829 0.555 0.573 0.584 BPSR1 0.563 0.871 0.658 0.597 BPSR2 0.532 0.890 0.616 0.640 BPSR3 0.628 0.887 0.670 0.682 BPSR4 0.378 0.816 0.475 0.518 BPV1 0.558 0.560 0.732 0.443 BPV2 0.423 0.457 0.732 0.491 BPV3 0.683 0.595 0.887 0.641 BPV4 0.637 0.655 0.871 0.695 BSAT1 0.660 0.677 0.690 0.887 BSAT2 0.556 0.711 0.620 0.867 BSAT3 0.410 0.398 0.355 0.716 BSAT4 0.613 0.481 0.614 0.805 Source: SmartPLS 4 Output. Note: Bold values are loadings on the target construct, which must exceed all other values in the same row. The PLSpredict procedure Shmueli et al., (2019) is used to

evaluate the out-of-sample predictive relevance of the model. All 12 indicators of the endogenous construct produced positive Q²predict values, ranging from 0.116 (BPSR4) to 0.450 (BPV3), confirming that the model has adequate outof-sample predictive relevance. The Q²predict value is generated through the out-of-sample holdout procedure and is conceptually different from the Q² Stone–Geisser produced through blindfolding. The comparison of RMSE and MAE values between PLS-SEM and the linear model (LM) shows that the PLS-SEM_RMSE values for most indicators are slightly higher than the LM_RMSE, indicating a marginal predictive advantage of the linear model at the indicator level (see Table 8). Nevertheless, referring to Shmueli et al., (2019), this condition does not negate the predictive relevance of the PLS-SEM model as long as all Q²predict values remain positive, which is met for all indicators in this model. The highest Q²predict values are found in BPV3 (0.450) and BSAT1 (0.425), while the lowest value is in BPSR4 (0.116), consistent with its relatively lower outer loading compared to other BPSR indicators.

Hypothesis Testing

The hypothesis testing results are presented in Figure 2 below. Table 9 presents the results of testing ten hypotheses using path coefficients through bootstrapping with 5,000 subsamples. All hypotheses are statistically supported (p < 0.05 for all paths; paths with p < 0.001 are reported as p < 0.001). The test results include six direct effects (H1–H6), four mediation effects (H7–H10), and one exploratory finding in the form of a three-stage serial mediation path (CSRI → BPSR → BPV → BSAT) reported separately and requiring replication before it can be generalised. The first main finding is that the CSRI → BPSR path (β = 0.617; t = 11.814) is the strongest path in the entire model. This finding is consistent with the core proposition of SET, that the quality of social exchange, rather than merely the existence of the program, is most strongly associated with the depth of transformation perceived by the recipients. In the context of Table 5. Discriminant Validity: Fornell-Larcker Criterion: FornellLarcker Criterion Construct BPSR BPV BSAT CSR BPSR 0.867 BPV 0.706 0.809 BSAT 0.709 0.714 0.822 CSR 0.617 0.722 0.693 0.833 Source: Primary Data, processed (2026) Note: Diagonal values (bold) represent the √AVE of each construct. Table 6. Discriminant Validity: Heterotrait-Monotrait Ratio of Correlations (HTMT) BPSR BPV BSAT CSR BPSR BPV 0.811 BSAT 0.791 0.822 CSRI 0.686 0.847 0.796 Source: SmartPLS 4 Output. All HTMT values < 0.85 (Henseler et al., 2014). Table 7. Statistical Model Structural: Structural VIF, R Square, Effect Size Collinearity Test (VIF) Coefficient of Determination Effect Size Path VIF Struktural Category VIF Construct Endogen R2 Category R2 Path f² Category f² CSRI -> BPSR 1.000 < 3.3 (no multicollinearity) BPSR 0.381 Moderate (0.25 – 0.50) CSRI - > BPSR 0.616 Big (≥0.35) CSRI -> BPV 1.616 < 3.3 (no multicollinearity) BPV 0.631 Substansial (0.50 – 0.75) CSRI - > BPV 0.356 Big (≥0.35) BPSR -> BPV 1.616 < 3.3 (no multicollinearity) BSAT 0.632 Substansial (0.50 – 0.75) BPSR - > BPV 0.298 Medium (0.02– 0.35) CSRI -> BSAT 2.191 < 3.3 (no multicollinearity) CSRI - > BSAT 0.107 Small (<0.15) BPSR -> BSAT 2.097 < 3.3 (no multicollinearity) BPSR - > BSAT 0.155 Medium (0.02– 0.35) BPV -> BSAT 2.708 < 3.3 (no multicollinearity) BPV - > BSAT 0.067 Small (<0.15) Source: SmartPLS 4 Output. Indonesian SOEs, this finding indicates that the relevance, transparency, consistency, and fairness of CSR program implementation are more strongly associated with the depth of transformation perceived by recipients compared to the program's budget size, although this relationship needs to be confirmed through longitudinal studies before it can be made a definitive strategic recommendation. H2 (β = 0.461; t = 6.717; p < 0.001) confirms the positive and significant influence of CSRI on BPV. This finding indicates that the quality of CSR program implementation is positively associated with the perceived value evaluation by the recipients. The second notable finding is that the BPSR → BSAT path coefficient (β = 0.346) exceeds that of BPV → BSAT (β = 0.258). This indicates that beneficiary-perceived selftransformation (BPSR) contributes more directly and more strongly to satisfaction than overall value evaluation a finding that challenges the conventional understanding in the service satisfaction literature, which has historically positioned perceived value as the dominant predictor. This finding introduces an important nuance, CSR beneficiary satisfaction forms not only through the cognitive-evaluative value pathway but more significantly through the affective-transformative pathway of experienced self-change. H7 (βindirect = 0.260; t = 5.210) confirms BPSR as a significant mediator in the CSRI → BPV path. This finding indicates that the quality of CSR programs is more strongly associated with higher value evaluations when meaningful selftransformation has been felt by the recipients, in line with (Zeithaml, 1988). Given that the research design is crosssectional, this relationship is indicative and requires confirmation through longitudinal. H8 (βindirect = 0.214; t = 4.473) confirms BPSR as a significant mediator in the CSRI → BSAT path. Together with H7, these findings indicate that BPSR plays an important mediating role in the pathway from CSR program quality to recipient value and satisfaction. Programs associated with more meaningful self-transformation for recipients tend to be associated with higher value and satisfaction as well, a proposition that requires confirmation through longitudinal studies before it can be recommended as a definitive program design. H9 (βindirect = 0.119; t = 2.537) confirms BPV mediation in the CSRI → BSAT pathway, applying Cronin & Taylor, (1992) causal chain to the Indonesian SOE CSR context. H10 (βindirect = 0.109; t = 2.212; p = 0.027) confirms that BPV significantly mediates the effect of BPSR on BSAT. Although the mediation coefficient of H10 is the smallest among the four mediation effects, it remains significant at p < 0.05, demonstrating that beneficiary-perceived self-transformation through BPSR does not automatically yield deep and stable satisfaction without first passing through the value evaluation stage. Such transformation must first be crystallized into an explicit judgment about how meaningful the program is to

beneficiaries’ lives, and it is from that judgment that satisfaction forms. As an exploratory finding beyond the ten established hypotheses, the three-stage serial mediation path (CSRI → BPSR → BPV → BSAT; β = 0,067; t = 2,125; p = 0,034) is also statistically supported. These findings indicate that the quality of CSR implementation is associated with recipient satisfaction through a sequential process: perceived selftransformation (BPSR) which then crystallises into value evaluation (BPV), and from that evaluation, satisfaction (BSAT) is formed. Given its exploratory nature and crosssectional research design, these findings require replication before they can be widely generalised. This discussion section interprets the research findings within the context of existing literature, explores their theoretical and practical implications, and identifies limitations and recommendations for future research.

Interpretation of Key Findings

This study yields four mutually reinforcing findings that collectively construct a coherent argument about how CSR programs operate as social exchange systems. Each finding does not merely confirm a hypothesis but reveals the psychological mechanism explaining why and under what conditions the relationship occurs. The first finding is the validation of BPSR as an empirically distinct construct from value and satisfaction. This provides strong preliminary evidence that BPSR measures a different psychological entity, namely the self-transformation of beneficiaries, which cannot be psychometrically reduced to existing constructs of value or satisfaction, although crosscontext replication is still needed to confirm this conclusion more definitively. Theoretically, this is not merely a matter of satisfactory psychometric figures. BPSR’s discriminant validity from BPV reveals a fundamental difference in psychological mechanisms: BPV operates through a comparative cognitive process in which beneficiaries weigh the benefits received against the sacrifices made, while BPSR operates through a transformative internalization process in which beneficiaries experience concrete changes in themselves. Both are psychological responses to the same program, yet they form through different pathways and serve different functions in the formation of satisfaction. This finding constitutes the logical prerequisite for all the mediation arguments constructed in this model. The second finding establishes implementation quality (CSRI) as a strong predictor of transformation. The operative mechanism is the reciprocity norm within Social Exchange Theory: beneficiaries who perceive the program as relevant, transparent, consistent, and fair respond by opening themselves more fully to the learning and change processes the program offers. Conversely, poor implementation activates psychological defense mechanisms that inhibit the internalization of change. A boundary condition worth noting is that this effect is likely stronger in non-transactional community-based program contexts such as SOE CSR, compared to the context of purely transactional services. The third finding reveals something theoretically unexpected: BPSR contributes more strongly to beneficiary satisfaction than BPV. This challenges the conventional view positioning perceived value as the dominant antecedent of satisfaction in relational exchange (Blut et al., 2023; Tedja et al., 2024). The mechanistic explanation lies in the psychological depth of the two pathways. Satisfaction mediated by value forms through situational cognitive calculation that is susceptible to change as comparison references shift. Satisfaction mediated by selftransformation, by contrast, is rooted in genuine identity and behavioral change, rendering it more stable and resistant to fluctuation. A boundary condition that must be emphasized is that this finding is likely specific to transformative service contexts where the program’s primary purpose is to alter beneficiaries’ capacities and behaviors, rather than merely transferring material benefits. In typical transactional contexts, the role of BPV likely remains dominant. The fourth finding concerns the overall mediation architecture of the model. The confirmed dual mediation roles of BPSR (toward both BPV and BSAT) and BPV (toward BSAT) demonstrate that CSR beneficiary satisfaction does not form through a single pathway but through two simultaneously operating and complementary ones: the transformativeaffective pathway through BPSR, and the evaluative-cognitive pathway through BPV. The fact that both pathways are statistically significant and both are necessary for comprehensively explaining satisfaction suggests that CSR program improvement interventions cannot be limited to addressing only one dimension. As an exploratory finding beyond the ten pre-specified hypotheses, the confirmation of the three-stage serial mediation pathway (CSRI → BPSR → BPV → BSAT; β = 0.067; t = 2.125; p = 0.034) carries important theoretical implications while reinforcing the overall coherence of the model. This finding reveals that CSR beneficiary satisfaction does not form through a single direct leap from implementation quality but through a stepwise, sequential psychological journey. Program implementation quality (CSRI) must first generate real and meaningful self-transformation in beneficiaries (BPSR); that transformation must then be crystallized into an explicit value evaluation (BPV); and only from that evaluation can comprehensive and enduring satisfaction (BSAT) form. Theoretically, this underscores that each stage in the causal chain is necessary and cannot be bypassed. High-quality CSR programs that fail to generate beneficiary-perceived selftransformation will not automatically produce positive value evaluations and will therefore fail to maximize beneficiary satisfaction. Conversely, self-transformation that has occurred does not directly culminate in satisfaction without first being processed into an explicit judgment about how meaningful the program is to beneficiaries’ lives. This sequential mechanism is consistent with the cognitive elaboration model in consumer psychology literature, which indicates that transformative experiences must be consciously internalized and evaluated before they can produce the affective response of satisfaction (Oliver, 1997; Zeithaml, 1988). Although exploratory and requiring replication before broad generalization, this finding opens a promising research agenda, particularly for understanding psychological mechanisms of satisfaction formation in community-based transformative service contexts.

Comparison with Previous Studies

The findings of this study are consistent with a number of prior studies while also providing original extensions. The confirmed effect of CSR implementation quality on beneficiary value and satisfaction aligns with the findings of Araújo et al., (2023), Kim et al., (2020), Islam et al., (2021), and Lee et al., (2020), who consistently confirmed this relationship across diverse industries and cultural contexts. The causal chain implementation quality → value → satisfaction, confirmed through H2, H6, and H9, is consistent with the foundational proposition of Cronin & Taylor, (1992) which remains among the most influential findings in the services marketing literature and is reinforced by the recent meta-analysis of Coelho do Vale et al., (2025) across more than one hundred CSR studies. This study, however, provides several significant extensions relative to prior research. First, it introduces BPSR as a construct previously nonexistent in the literature, filling the gap explicitly identified by Siegal, (2022), Corvo et al., (2022), and Gutiérrez-Nieto et al., (2025). It is important to explicitly clarify how BPSR differs from related constructs. BPSR differs from perceived value (BPV) in that BPV is comparatively evaluative

(benefits vs. sacrifices), whereas BPSR is transformatively internal (real changes within the beneficiary). BPSR differs from satisfaction (BSAT) in that BSAT is an affective confirmation of expectations, whereas BPSR measures substantive experienced change. BPSR also differs from empowerment or capability constructs in social psychology because BPSR explicitly links the transformation to engagement in a specific CSR program and operates as a psychometric construct integrable into SEM models. Unlike the SROI approach, which calculates the monetary value of impact from outside, BPSR measures the perception of transformation from within the beneficiary using standard survey instruments, making it complementary rather than competing. Second, the confirmed mediating role of BPSR in the CSRI → BPV pathway (H7) extends Zeithaml, (1988) means-end hierarchy to the CSR program context by adding a layer of consequences in the form of self-transformation previously unaccounted for. Theoretically, this explains why CSR programs with identical implementation quality can yield different value evaluations: the difference is mediated by the depth of transformation perceived by beneficiaries. Third, the finding that BPSR (β = 0.347) has a larger path coefficient than BPV (β = 0.264) in predicting BSAT can be explained through different psychological mechanisms: satisfaction mediated by self-transformation is rooted in genuine identity and behavioral change, rendering it more resistant to situational fluctuation than satisfaction mediated by cognitively calculative value evaluation. This finding challenges the general view of perceived value as the primary predictor of satisfaction (Blut et al., 2023; Tedja et al., 2024), subject to the boundary condition that this pattern is specific to community-based transformative service contexts where participation is non-transactional. It is important to note that these contribution claims do not imply that related constructs are entirely irrelevant or nonexistent. Several adjacent constructs have addressed aspects of psychological responses to social interventions: perceived impact in the program evaluation literature measures overall assessments of program effectiveness; psychological empowerment measures generalized feelings of control and self-efficacy (Llorente-Alonso et al., 2023; Maier & Baccarella, 2024; Spreitzer, 1995) and transformative service outcomes in the Transformative Service Research (TSR) tradition (Anderson & Ostrom, 2015; Bakrim & Fritzsche, 2025) conceptualize well-being improvement as a service outcome. BPSR differs from all three in terms of specificity: it measures self-transformation across four theoretically distinct dimensions, is causally connected to the exchange mechanism in SET, and is explicitly designed to bridge SROI’s social accountability logic with the value–satisfaction psychological model in marketing. BPSR is therefore not merely a relabeling of existing constructs but a conceptual reconfiguration that fills a specific void previously unoccupied by any construct in the literature.

Limitations and Cautions

While this study produces rich findings and makes original contributions, five principal limitations must be honestly acknowledged and should inform how readers interpret the strength of the claims advanced. First, the cross-sectional design limits the ability to capture the temporal dynamics of beneficiary-perceived transformation. All measurements were taken at a single point in time, making it impossible to determine how each BPSR dimension evolves, persists, or fades following program completion. This limitation also means that the mediation claims are cross-sectional in nature, consistent with the hypothesized causal model but not yet temporally demonstrable. Second, the research context, confined to the PLN Peduli CSR program in West Kalimantan, limits the external generalizability of the findings. Although the contextual selection was grounded in sound methodological considerations, the possibility that context-specific factors contributed to the magnitude of inter-variable relationships cannot be entirely ruled out. Third, BPSR remains at the initial validation stage within a single CSR program context; reliability and validity claims require strengthening through replication across more diverse contexts. Fourth, all data are self-reported and were collected from the same source at the same time (common method variance, CMV). Although procedural and statistical controls were applied, the potential for CMV cannot be fully eliminated and may influence the magnitude of reported correlations. Fifth, this study did not explore potential moderating variables, such as the type of CSR program, intensity and duration of beneficiary engagement, and respondents’ diverse demographic characteristics, that could affect the strength of inter-variable relationships.

Recommendations for Future Research

The limitations identified simultaneously open a rich research agenda. First, longitudinal studies measuring BPSR at multiple time points, before the program begins, during implementation, and several months after completion, are strongly recommended to understand how beneficiaryperceived transformation develops and persists over time, and to provide more definitive answers regarding whether the community engagement dimension requires a longer incubation period. Second, replication across CSR programs from other industries such as banking, mining, and manufacturing, across diverse Indonesian regions and other developing countries, is essential to validate the boundary conditions under which BPSR applies and to strengthen claims of generalizability. Third, exploration of potential moderating variables, including type of CSR program, intensity of beneficiary engagement, and demographic characteristics, would enrich understanding of the conditions under which BPSR operates most powerfully as a mediator. Fourth, exploring how the validated BPSR contributes to corporate social capital formation through downstream variables such as beneficiary affective engagement, institutional trust, corporate image, and beneficiary advocacy would constitute a valuable extension.

Conclusion

This study provides initial evidence for the development and testing of the Beneficiary Perceived Social Return (BPSR) construct as a psychometric bridge between SROI methodology and consumer valueand customer satisfaction-based marketing theory, within the specific context of the PLN Peduli CSR program in West Kalimantan. These findings are not intended as a claim of full cross-contextual validation but as an initial step requiring reinforcement through replication across CSR programs from more diverse sectors, regions, and cultures. Within the bounds of that context, BPSR demonstrates strong psychometric performance and fills the gap repeatedly identified in the literature: the absence of a construct capable of measuring social return from beneficiaries’ internal perspective in a format integrable into causal marketing models. BPSR’s role as a “bridge” between SROI and marketing operates through two distinct yet complementary mechanisms. First, BPSR provides initial evidence of strong psychometric performance and construct uniqueness in this sample, supported by genuinely distinct constructs from existing ones, demonstrated through strong outer loading profiles, superior AVE and Composite Reliability values, and confirmed discriminant validity against all other constructs in the model. Second, as a mediating construct, BPSR functions as a causal mechanism explaining why and