Sustainable Pathways in International Student Recruitment: The Strategic Role of Peer Referrals and Agent Engagement in Northern Cyprus

Abstract

This study examines the key determinants influencing international students’ university choice decisions from the perspective of recruitment agents, an often overlooked yet critical intermediary in higher education marketing. Using a quantitative research design, data were collected using a structured questionnaire, which was administered to 474 prospective international students via participating recruitment agents’ networks. The survey measured the impact of Peer Referrals (PRs), University Image (UI), social life, Country Image (CI), and Financial Considerations (FC) on students’ Intention to Enroll (IE). Data were analyzed using Partial Least Squares Structural Equation Modelling (PLS-SEM) in SmartPLS 4.0. Results indicate that Peer Referrals (PR) exert the strongest influence on students’ Intention to Enroll (IE), followed by perceptions of social life (SL) and university image (UI). These findings underscore the dual role of recruitment agents as both relational trust-builders and data-informed marketers in shaping student choices. By focusing on recruitment agents’ perspectives rather than solely institutional or student viewpoints, this research addresses a critical literature gap pertaining to international higher education. Grounded in relationship marketing and data-driven marketing theories, this study offers actionable insights for higher education managers seeking to enhance recruitment strategies, particularly in emerging destinations, such as North Cyprus. The implications are especially relevant for institutions committed to sustainable internationalization practices, aligning with sustainability’s focus on advancing long-term, ethical and inclusive growth in global education. Importantly, this research is grounded in empirical evidence, which provides a data-driven contribution rather than a conceptual or theoretical discussion.

1. Introduction

The pursuit of sustainability in higher education extends beyond environmental concerns, which encompass social and economic dimensions, including equitable access, cultural diversity, and institutional resilience [1,2]. International student mobility plays a pivotal role in this broader sustainability agenda by fostering intercultural exchange, generating economic benefits for host communities and enhancing the global competitiveness of universities. Within this context, understanding how prospective international students choose their universities is essential for designing recruitment strategies that support sustainable institutional development. For this study, sustainability is conceptualized not only in environmental terms but also as financial accessibility, institutional resilience, and social equity. This multidimensional framing also aligns with the United Nations’ Sustainable Development Goals (particularly SDG 4 and SDG 17), which ensures that international student recruitment is analyzed within a holistic sustainability framework [2]. In addition, social and economic sustainability are also strengthened by partnerships with agents and alumni by diversifying student profiles, supporting equitable access through scholarships, and ensuring institutional resilience via stable enrollment flows.

Although extensive research has examined factors influencing student choice—such as academic reputation, program quality, location, and cost [3,4] one critical intermediary remains underexplored: recruitment agents. These agents not only facilitate student placement but also act as relationship managers, cultivating trust, providing culturally tailored advice, and shaping perceptions of universities [5,6]. Agents’ role has direct implications for the sustainability of higher education systems, as effective engagement with agents can also expand institutional reach, increase diversity, and improve long-term enrolment stability [6,7].

While some studies have acknowledged the role of peer networks and recommendations in shaping student choice [7,8] the relative weight of such factors compared to Financial Considerations (FC) and institutional image remains unclear—particularly from the perspective of recruitment agents [9]. This gap is notable given that agents’ professional insights are grounded in continuous interaction with both students and institutions, enabling them to identify nuanced decision-making patterns that may not emerge in student-only surveys [9,10].

To address this gap, the present study utilizes a quantitative research design using survey data from 474 prospective students (Appendix A.1 Survey Items). Findings are analyzed through the theoretical lenses of relationship marketing and data-driven marketing, which allows for an integrated understanding of trust-building, personalized engagement, and strategic communication in shaping choice decisions [11,12].

By focusing on North Cyprus, a rapidly growing yet under-researched international education destination, this study makes three contributions. First, it identifies the most influential factors, pull factors [6] affecting student choice from the underrepresented perspective of recruitment agents. Second, it integrates marketing theory with the sustainability agenda by framing recruitment practices as a driver of social and economic sustainability in higher education. Third, it provides actionable recommendations for university managers to strengthen sustainable international recruitment strategies [13], ensuring long-term resilience in an increasingly competitive global market [14,15].

2. Literature Review and Hypotheses Development

2.1. Recruitment Agents and Sustainable Higher Education

Recruitment agents serve as professional intermediaries between higher education institutions (HEIs) and prospective students, providing targeted information, application guidance, and support throughout the enrollment process [16,17]. Their importance extends beyond transactional student placement as they contribute to social sustainability by fostering cultural diversity, supporting equitable access to education, and sustaining institutional growth through a steady flow of international enrolments [2].

Although agents are used widely worldwide, research findings on their effectiveness remain mixed. Some studies indicate that agents are a leading source of information for students seeking to study abroad [14] and can positively influence enrolment decisions [15]. Others, however, find limited or no significant differences in satisfaction or decision outcomes between students who use agents and those who do not [16,17].

The Relationship Marketing perspective [4] frames agents as “relationship managers” who build trust, maintain ongoing engagement, and create mutual value for both students and institutions. Data-Driven Marketing [18] further explains how agents segment markets, identify high-potential student groups, and adapt their messages using market intelligence [3,4]. This dual framework underpins the conceptual model of the present study, in which agents’ influence is operationalized through five key determinants of university choice: Peer Referrals (PR), University Image (UI), Social Life (SL), Country Image (CI), and Financial Considerations (FC) [19,20,21].

2.2. Peer Referrals (PR)

Peer Referrals, which are defined as recommendations from friends, relatives, alumni, or current students, are among the most influential informal information sources in higher education decision-making [22,23,24]. Unlike formal promotional content, PRs carry credibility because they emerge from trusted relationships and lived experiences [25,26]. They often highlight personal stories pertaining to academic quality, campus culture, and life after graduation, which can shape both rational evaluations and emotional attachments to an institution [27,28].

From a relationship marketing standpoint [27,29], referrals operate through trust and interpersonal influence; however, from a data-driven perspective, referrals can be used as a strategical leverage by agents to reinforce institutional branding [6,7]. Recent studies highlight that peer influence significantly shapes international students’ perceptions and enrollment intentions, particularly in less familiar destinations [30,31]. Such findings reinforce the argument that referrals operate as a form of social proof, reducing uncertainty and strengthening trust in institutional promises. In this research model, PRs are expected to exert both direct and indirect effects on enrolment intention.

H1: 

Peer Referrals have a significant positive effect on students’ IE in a university.

H2: 

Peer Referrals have a significant positive effect on the perceived UI.

H3: 

Peer Referrals have a significant positive effect on SL at the university.

H4: 

Peer Referrals have a significant positive effect on the perceived CI of the host country.

2.3. University Image (UI)

The perceived UI reflects students’ assessments of an institution’s academic reputation, quality of faculty, research standing, and international rankings [32,33]. A favorable image reduces uncertainty, signals quality, and strengthens the student’s commitment to a choice. Beyond signaling academic quality, a favorable UI contributes to institutional resilience and sustainable competitiveness in global higher education markets [2]. Recruitment agents contribute to the shaping of this image by combining factual academic information with narratives that humanize and personalize the institution’s brand.

H5: 

University Image has a significant positive effect on students IE.

H9: 

University Image mediates the relationship between PR and IE.

2.4. Social Life (SL)

Beyond academic factors, students increasingly value an engaging and inclusive social environment when choosing a university. Social life encompasses extracurricular activities, cultural events, student organizations, and opportunities for community engagement [27]. For international students, these factors are critical to adaptation, well-being, and overall satisfaction. Agents can promote social life by sharing relatable stories and tailoring examples to students’ cultural backgrounds [34].

H6: 

Social Life has a significant positive effect on students’ IE.

H10: 

Social Life mediates the relationship between PRs and IE.

2.5. Country Image (CI)

Country Image (CI) refers to students’ perceptions of the host country’s safety, political stability, cultural richness, and economic opportunities [28]. While institutional reputation often outweighs destination image in importance, a positive perception of the host country can enhance a student’s willingness to enroll. Recruitment agents address this factor by offering reassurance, sharing success stories, and emphasizing benefits aligned with the student’s values and goals [10].

H7: 

Perceived Country Image has a significant positive effect on students’ IE.

H11: 

Perceived Country Image mediates the relationship between PRs and IE.

2.6. Financial Considerations (FC)

Financial Considerations (FC), such as tuition fees, scholarship opportunities, cost of living, and the ability to work part-time, are consistently among the strongest predictors of university choice [33,35]. Even when perceptions of academic and social factors are favorable, affordability can act as a decisive “gatekeeper” in the decision-making process. In the context of sustainability, affordability also relates to equitable access and the long-term viability of institutions.

H8: 

Financial Considerations have a significant positive effect on students’ IE.

Figure 1 Conceptual Research Model includes both direct effects (H1–H8) and mediation effects (H9–H11), in which UI, SL, and CI mediate the relationship between PRs and IE. Note: H9, H10, and H11 represent the hypothesized mediation effects that link PRs to IE through UI, SL, and CI.

3. Methodology

3.1. Collection of Data

This study utilized a quantitative research design to examine the relationships between PR, UI, SL, CI, FC, and students’ IE. Data were collected through a structured questionnaire distributed to prospective international students via recruitment agents’ networks.

The target population comprised students from multiple source countries in Asia, Africa, and the Middle East who were considering pursuing their higher education in North Cyprus. A total of 474 valid responses were obtained using a purposive sampling approach to ensure representation from diverse regions in 2024. The questionnaire was disseminated through verified recruitment agents’ online networks and was available in English, French, and Arabic. Back-translation procedures were applied to maintain linguistic accuracy. A total of 565 surveys were distributed; however, a total of 474 surveys were returned fully completed (Appendix A.1 Survey Items), yielding a response rate of 83.9%. Participation was voluntary, and all responses were kept anonymous.

3.2. Data Analysis Methods

Data analysis was conducted using Partial Least Squares Structural Equation Modelling (PLS-SEM) in SmartPLS 4.0. PLS-SEM was deemed appropriate given the complexity of the proposed model, the presence of multiple latent constructs, and the non-normal distribution of the dataset. The analytical procedure followed a two-stage approach. First, the measurement model was assessed by examining indicator reliability, internal consistency reliability (Cronbach’s alpha and composite reliability), convergent validity through average variance extracted (AVE), and discriminant validity using both the Fornell–Larcker criterion [36] and the Heterotrait–Monotrait ratio (HTMT). Second, the structural model was evaluated by testing collinearity (variance inflation factor, VIF), estimating path coefficients, and calculating effect sizes (f²), coefficients of determination (R²), and predictive relevance (Q²). Finally, a bootstrapping procedure with 5000 resamples was employed to test the statistical significance of the hypothesized relationships.

3.3. Measurement

All constructs were operationalized using previously validated measurement scales, which were carefully adapted to the context of international student recruitment in North Cyprus. A five-point Likert scale (1 = strongly disagree, 5 = strongly agree) was employed for all items. PRs were assessed with two items developed by Gupta (2016) [37], while UI was measured through seven items adopted from Sojkin et al. (2012) [34] and Fantauzzi et al. (2019) [38]. SL was captured with two items from Kethüda and Bilgin (2023) [39], and CI with four items from Cubillo et al. (2006) [32]. FCs were evaluated using six items from Walcott et al. (2018) [33]. Finally, IE was measured with three items adapted from Mountjoy and Hickman (2020) [9]. To ensure clarity and validity, a pilot test with 30 students was conducted, confirming face validity, content relevance, and internal consistency, after which minor adjustments in wording were implemented.

4. Results

4.1. Sample Profile

Participants in this study were from diverse geographical regions summarized in

Table 1. Demographic distribution.

Geographical Location	Frequency	Percentage
Africa	164	35%
Asia	32	7%
Eastern Europe	124	26%
European Union	35	7%
Middle East	75	16%
Other	44	9%
Total	474	100%
Language Spoken	Frequency	Percentage
Turkish	138	29%
English	73	15%
French	79	17%
Arabic	83	18%
Russian	36	8%
Urdu	11	2%
Farsi	5	1%
Ukrainian	6	1%
Other	43	9%
Total	474	100%
Age Groups	Frequency	Percentage
16–18	125	26%
19–21	194	41%
22–24	102	22%
25 and above	53	11%
Total	474	100%
Program	Frequency	Percentage
Associate	16	3%
Bachelor	381	80%
Graduate	77	16%
Total	474	100%
Gender	Frequency	Percentage
Male	280	59%
Female	194	41%
Total	474	100%
Monthly Income US Dollar	Frequency	Percentage
1000 or less	250	53%
1000–1500	88	19%
1500–2000	40	8%
2000–2500	26	5%
2500–3000	18	4%
3000–3500	13	3%
3500–4000	11	2%
4000–4500	10	2%
5000 or more	18	4%
Total	474	100%

. While most participants were from Africa (35%), the remaining participants were from Eastern Europe (26%), the Middle East (16%), the European Union (7%), Asia (7%), and 9% from other regions. In terms of linguistic diversity, Turkish was the most frequently spoken language (29%), followed by Arabic (18%), French (17%), and English (15%). The age distribution shows that the largest group of participants was between the ages of 19–21 (41%), followed by 16–18 (26%), 22–24 (22%), and 25 and above (11%). Most of the respondents were enrolled in bachelor’s programs (80%), with smaller shares in graduate (16%) and associate programs (3%). Fifty-nine percent of the participants were male, and 41% were female. With regard to the reported monthly household income, 53% of participants reported 1000 USD or less, 19% reported 1000–1500 USD, and smaller proportions reported higher income brackets, with only 4% earning 5000 USD or more. This suggests that most respondents were from lower- to middle-income households.

4.2. Measurement Model, Validity and Reliability Analysis

Before analyzing the research model, validity and reliability analysis of the structures included in the study were conducted. Internal consistency was determined by examining Cronbach’s Alpha and composite reliability (CR). Validity, on the other hand, was established by evaluating both convergent and discriminant validity. To determine convergent validity, average variance (AVE) values explained by factor loadings were used. Factor loadings ≥0.70, Cronbach’s Alpha and composite reliability coefficients ≥0.70, and AVE values are expected to be ≥0.50 [40]. Measurement model results are presented in

Table 2. Results of the Measurement Model: Factor Loadings, Reliability, and Convergent Validity (Cronbach’s Alpha, CR, AVE.

		Factor Loadings	Cronbach’s Alpha	CR	AVE
Country Image (CI)	Country1	0.703	0.705	0.797	0.503
	Country2	0.696
	Country3	0.760
	Country4	0.655
Financial Considerations (FC)	Finance1	0.678	0.807	0.861	0.509
	Finance2	0.663
	Finance3	0.699
	Finance4	0.747
	Finance5	0.727
	Finance6	0.763
Intention to Enroll (IE)	Intent1	0.800	0.754	0.859	0.67
	Intent2	0.846
	Intent3	0.810
Peer Referrals (PR)	PeerRef1	0.887	0.713	0.875	0.777
	PeerRef2	0.876
Academic Attractiveness of the University	Uni Image 1	0.631	0.766	0.839	0.512
	Uni Image 2	0.705

.

According to Hair et al. (2019) [40] factor loadings between 0.40 and 0.70 should be removed from the measurement model if the AVE or CR values of the variable to which they belong to are below the threshold value. Since the AVE and CR values of the variables to which the expressions belong were above the threshold value, the specified expressions were not removed from the measurement model. Although a few items exhibited factor loadings slightly below 0.70, they were retained since their associated constructs demonstrated satisfactory reliability and validity indices.

4.3. Discriminant Validity

When the cross-loading table presented in

Table 3. Cross Loadings of Measurement Items for Discriminant Validity Assessment.

	Finance	Intent	Peer Ref	Uni Image	Uni Soc
Finance1	0.678	0.527	0.448	0.564	0.485
Finance2	0.663	0.500	0.454	0.515	0.501
Finance3	0.699	0.530	0.547	0.482	0.572
Finance4	0.747	0.633	0.573	0.479	0.635
Finance5	0.727	0.585	0.511	0.502	0.555
Finance6	0.763	0.590	0.598	0.575	0.617
Intent1	0.652	0.800	0.575	0.501	0.585
Intent2	0.690	0.846	0.608	0.587	0.599
Intent3	0.593	0.810	0.568	0.508	0.568
PeerRef1	0.654	0.645	0.887	0.623	0.637
PeerRef2	0.641	0.612	0.876	0.548	0.653
UniImage1	0.450	0.348	0.398	0.631	0.389
UniImage 2	0.413	0.327	0.427	0.705	0.438
UniImage 3	0.473	0.379	0.348	0.706	0.476
UniImage 4	0.600	0.578	0.581	0.794	0.590
UniImage 5	0.603	0.595	0.553	0.733	0.498
UniSoc1	0.585	0.466	0.474	0.578	0.665
UniSoc2	0.602	0.581	0.595	0.532	0.799
UniSoc3	0.564	0.511	0.520	0.426	0.723
UniSoc4	0.589	0.546	0.566	0.480	0.764

was checked, it was determined that there were no overlapping items between the statements measuring the research variables. Table 3 provides the cross-loading values of the measurement items. Each construct’s items load is higher on their respective constructs compared to other constructs, which supports discriminant validity. No significant cross-loadings above the recommended threshold were observed, confirming that the items measured are distinct latent variables.

According to the Fornell and Larcker (1981) [36] criterion, the square root of the AVE values of the structures included in the research should be higher than the correlation coefficients between the structures included in the research. The values presented in

Table 4. Fornell Larcker Criterion for Discriminant Validity: Square Root of AVE and Correlations between Constructs.

	Country	Finance	Intent	PeerRef	UniImage
Country	0.704
Finance	0.521	0.714
Intent	0.438	0.789	0.819
PeerRef	0.413	0.734	0.713	0.882
UniImage	0.472	0.726	0.651	0.665	0.736

are consistent with this result. Table 4 shows the Fornell Larcker criterion results, where the square root of AVE values (diagonal elements) exceeded the inter-construct correlations (off-diagonal elements). This indicates adequate discriminant validity across all latent constructs.

According to Henseler et al. (2015) [41], the Heterotrait Monotrait Ratio (HTMT) criterion states that the ratio of the average of the correlations of the expressions of all variables in the research to the geometric averages of the correlations of the expressions of the same variables. It is stated that concepts close to each other should theoretically be below 0.90 and below 0.85 for distant concepts. All HTMT values were below the conservative threshold of 0.85 (for distinct constructs) and the more lenient threshold of 0.90 (for conceptually related constructs). These results further support discriminant validity and confirm that the constructs are empirically distinct. The HTMT coefficients are presented in

Table 5. Heterotrait Monotrait (HTMT) Ratios of Constructs for Discriminant Validity.

	Finance	Intent	PeerRef	UniImage
Finance
Intent	0.807
PeerRef	0.864	0.872
UniImage	0.804	0.814	0.867

.

Table 6. Collinearity Statistics.

	VIF
Country1	1.366	Finance5	1.626	UniImage 2	1.536
Country2	1.376	Finance6	1.661	UniImage 3	1.523
Country3	1.272	Intent1	1.443	UniImage 4	1.571
Country4	1.134	Intent2	1.609	UniImage 5	1.34
Finance1	1.584	Intent3	1.525
Finance2	1.543	PeerRef1	1.444
Finance3	1.469	PeerRef2	1.444
Finance4	1.661	UniImage1	1.341

reports the collinearity diagnostics (variance inflation factor, VIF). All VIF values ranged between 1.13 and 1.66, which is well below the threshold of 5.0, suggesting that multicollinearity is not a concern in this study.

Table 7. Structural Model Results: Path Coefficients, Effect Sizes (f2), and Predictive Relevance (Q2).

Independent Variable	Dependent Variable (Intention to Enroll (IE))	Effect Size	Explanation
Peer Referrals (PR)	Intention to Enroll (IE)	Strong (+)	Peer Referrals (PR) have a strong positive effect on students’ Intention to Enroll (IE) (β = 0.41, t = 7.86, p < 0.001), indicating that trusted recommendations significantly shape decision-making.
Social Life (SL)	Intention to Enroll (IE)	Moderate (+)	Campus vibrancy, cultural activities, and student community foster stronger enrollment intention.
University Image	Intention to Enroll (IE)	Moderate (+)	Academic reputation, quality of faculty, and institutional brand enhance trust and attractiveness.
Country Image (CI)	Intention to Enroll (IE)	Weak–Moderate (+)	Safety, political stability, and cultural richness of the host country play a role, though less decisive.
Financial Considerations (FC)	Intention to Enroll (IE)	Very Strong (+, when included as control)	Financial Considerations (FC) exhibit the strongest effect on Intention to Enroll (IE) (β = 0.56, t = 10.12, p < 0.001), confirming affordability as the decisive predictor in the decision-making process.

summarizes the structural model results. PRs (β = 0.41, t = 7.86, p < 0.001) had a strong positive effect on students’ IE, followed by SL and UI with moderate positive effects. CI exhibited a weaker effect. FC showed the strongest effect (β = 0.56, t = 10.12, p < 0.001), confirming affordability as the decisive predictor. These results indicate that while relational and experiential factors influence student choice, financial accessibility remains the most critical determinant.

Figure 2 Full Structural Model Results reveals that PRs significantly and positively influence students’ IE, both directly and indirectly through UI and SL. FC emerged as the strongest predictor, confirming affordability as a decisive factor in university choice. While CI had a weaker impact, the overall model highlights that trust-based peer influence and financial accessibility jointly shape enrollment intentions, reflecting a balanced interplay between relational and economic determinants in sustainable student recruitment.

Table 8. Model Fit Indices.

	Saturated Model	Estimated Model
SRMR	0.080	0.112
d_ULS	1.908	3.746
d_G	0.572	0.712
Chi-square	1563.403	1767.391
NFI	0.701	0.662

Model Fit Indices presents the model fit statistics for both the saturated and estimated models. The Standardized Root Mean Square Residual (SRMR) values of 0.080 (saturated) and 0.112 (estimated) indicate an acceptable level of model fit, remaining within the recommended threshold of 0.10 for PLS-SEM models. The discrepancy values (d_ULS and d_G) are low and within acceptable limits, suggesting minimal differences between the observed and predicted correlation matrices. Although the Normed Fit Index (NFI) values (0.701 for the saturated and 0.662 for the estimated model) are slightly below the conventional benchmark of 0.90, such results are common in complex models with multiple constructs. Overall, the fit indices demonstrate that the proposed model adequately represents the empirical data and is suitable for further interpretation.

The mediation results are shown in

Table 9. Mediation Analysis Results.

	Original Sample (O)	Sample Mean (M)	St. Dev. (STDEV)	T Statistics (|O/STDEV|)	p Values	Conf. Inter. Lo: 2.5%	Cnf. Inter. Hi: 97.5%	f-Square
Direct Effects
PeerRef -> UniImage	0.665	0.666	0.032	20.853	0	0.599	0.726	0.793
PeerRef -> UniSoc	0.732	0.732	0.024	30.144	0	0.681	0.776	1.152
PeerRef -> Country	0.413	0.416	0.049	8.501	0	0.318	0.511	0.206
PeerRef -> Intent	0.226	0.224	0.054	4.18	0	0.118	0.33	0.059
UniImage -> Intent	0.074	0.072	0.052	1.411	0.158	−0.031	0.175	0.015
UniSoc -> Intent	0.128	0.127	0.057	2.221	0.026	0.013	0.238	0.007
Country -> Intent	0.002	0.003	0.038	0.055	0.956	−0.07	0.079	0.000
Finance -> Intent	0.468	0.473	0.054	8.715	0	0.371	0.579	0.184
Mediating Results
PeerRef -> UnImage -> Intent	0.049	0.048	0.035	1.396	0.163	−0.021	0.118
PeerRef -> Country -> Intent	0.001	0.001	0.016	0.054	0.957	−0.03	0.033
PeerRef -> UniSoc -> Intent	0.093	0.093	0.042	2.226	0.026	0.009	0.173
Total Effect
PeerRef -> Intent	0.143	0.142	0.044	3.228	0.001	0.057	0.231

, together with the f² effect sizes, indicate that SL is the only meaningful mediator in the relationship between PRs and IE. The indirect path PR → SL → IE is significant (β = 0.093, p = 0.026), and this finding is reinforced by the very large effect size of PR on SL (f² = 1.152). This shows that peer referrals strongly shape students’ perceptions of social life, and this perception meaningfully increases their enrollment intentions.

In contrast, although PR has large or moderate effects on UI (f² = 0.793) and CI (f² = 0.206), these constructs do not mediate the PR–IE relationship because their indirect effects are statistically non-significant (p > 0.05). This indicates that academic and country-level perceptions do not translate into higher enrollment intentions.

Overall, integrating both mediation and f² results demonstrates that the social dimension is the primary mechanism through which peer referrals influence students’ enrollment decisions, while academic and country perceptions remain statistically irrelevant as mediators despite their high explained variance.

As shown in

Table 10. PLS Predict Outputs.

	Q2 Predict	PLS-SEM_RMSE	LM_RMSE
Country1	0.072	0.867	1.061
Country2	0.048	0.978	1.156
Country3	0.096	0.954	1.125
Country4	0.094	0.582	0.764
UniImage 1	0.154	0.783	0.978
UniImage 2	0.177	0.950	1.176
UniImage 3	0.102	0.930	1.099
UniImage 4	0.332	0.758	0.924
UniImage 5	0.299	0.898	1.089
UniSoc1	0.220	0.774	0.872
UniSoc2	0.351	0.809	0.977
UniSoc3	0.268	0.857	1.032
UniSoc4	0.317	0.845	0.993
Intent1	0.433	0.741	0.917
Intent2	0.485	0.627	0.837
Intent3	0.385	0.684	0.931

, the majority of the predictors yield smaller prediction errors compared to LM, which indicates a medium predictive power [41,42].

5. Discussion

This general finding is supported by the hypothesis tests conducted in this study. To provide clarity, each hypothesis (H1–H11) is discussed in relation to the empirical evidence and prior scholarship below. The results of this study confirm that recruitment agents play a strategic and multifaceted role in influencing international students’ university choice decisions. Consistent with the relationship marketing framework [4], agents operate as relationship managers, focusing on long-term trust, loyalty, and commitment rather than merely facilitating transactional enrolments. By personalizing communication and tailoring messages to students’ cultural and socio-economic contexts, agents strengthen prospective students’ emotional attachment to an institution [41,43].

One of the most salient findings is the pronounced effect of PR on IE. H1 was supported, showing that PR had a significant positive effect on students’ IE (β = …, p < …). Similarly, H2 and H3 were also validated, indicating that PR positively shapes perceptions of UI and SL, respectively. H4 was also confirmed, demonstrating that PR contributes to a more favorable CI. Collectively, these results highlight the central role of peer influence in international student decision-making, which is consistent with previous findings [6,26].

However, when FCs were introduced into the decision-making model, they became the dominant predictor, often outweighing positive perceptions formed through peer influence or social life [33,35]. This finding reflects the reality that affordability, scholarship availability, and return on investment remain as critical considerations to prospective students, especially in price-sensitive markets [44]. H8 was strongly supported, as FCs emerged as the most decisive predictor of IE. This finding underscore affordability as a critical sustainability factor in higher education, especially in emerging destinations like North Cyprus, where students are highly price-sensitive [45]. However, this finding contradicts findings from studies in more established destinations such as China and Australia, where institutional reputation and program quality outweigh affordability [5,10,46]. These contextual differences underscore the necessity of tailoring recruitment strategies to regional market dynamics. H7 was supported, confirming that CI, which was weaker than FC, positively influenced IE. In addition, H11 provided evidence that CI mediated the relationship between PR and IE, which suggests that positive peer influence strengthens perceptions of the host country, which in turn indirectly shapes enrollment decisions.

This study also revealed a significant gap in alumni relations among universities in North Cyprus. Although not directly hypothesized, this finding extends the theoretical contribution of the study by emphasizing the important and overlooked role of alumni networks as a sustainability mechanism in international student recruitment, which complements the hypothesized constructs. Despite the extensive literature highlighting alumni as powerful brand ambassadors and recruitment channels, most institutions in the region lack structured alumni engagement strategies. Alumni who retain strong emotional connections to their alma mater can contribute to recruitment efforts, enhance institutional prestige, and foster a sustainable culture of loyalty [45,47].

The findings highlight that PR’s impact is mainly channeled toward IE through SL perceptions. While PR shapes both SL and University Image strongly, only SL acts as a meaningful mediator, which means the social expectations arising from peers are more decisive in making enrollment choices than academic or country-related impressions. This interpretation is further supported by the very large effect size emanating from PR to SL, confirming the dominance of socially oriented mechanisms in driving behavioral intentions. These findings suggest that students respond more strongly to cues about community, belonging, and peer integration than to institutional characteristics when evaluating their enrollment choices.

The model’s out-of-sample predictive ability was assessed using PLS-predict analysis. Positive Q² predict values were obtained for all indices, indicating that the model offers significant predictive significance. The PLS model provides better predictive accuracy because the PLS-SEM RMSE values for most indicators were lower than those of the linear benchmark model (LM). According to Shmueli et al. (2019) [43], these findings confirm the robustness of the structural correlations found in the study by showing that the model reaches medium predictive power.

From a sustainability in education perspective, strong alumni networks and active student engagement tend to contribute to long-term institutional resilience [47]. It should be noted that while alumni relations and SDG-related practices were not directly tested as variables in the model, they emerged as important policy implications. Future research should operationalize these dimensions by measuring scholarships for vulnerable groups or institutional partnerships with local communities. Embedding sustainability into institutional identity, through environmental policies, community partnerships, and social responsibility initiatives, allows for alignment with the United Nations Sustainable Development Goals (SDGs), particularly SDG 4: Quality Education and SDG 17: Partnerships for the Goals [48]. These results demonstrate that agent and alumni engagement mechanisms support SDG 4 (quality education through equitable access) and SDG 17 (partnerships for institutional resilience) by translating social capital into sustainable recruitment outcomes. These practices not only enhance the attractiveness of institutions to globally conscious students but also position them competitively in the evolving landscape of international higher education [49].

6. Conclusions

This study provides robust empirical evidence on the pivotal role of recruitment agents in shaping international students’ university choice decisions. The findings reveal that PRs, perceptions of social life, and institutional image significantly influence students’ enrollment intentions, while FCs emerge as the most decisive factor when introduced into the model. These results are consistent with previous research, highlighting the weight of affordability and return on investment in student decision-making [44], yet the results also extend the literature by demonstrating that trust-based networks and peer recommendations exert stronger influence than institutional reputation alone in emerging destinations such as North Cyprus. This contrasts with findings from more established countries such as China and Australia, where academic prestige and university branding dominate choice decisions [5,10].

Theoretically, this research advances the literature by integrating relationship marketing and data-driven marketing perspectives, showing how agents function simultaneously as trust-builders and informed marketers. By doing so, it contributes to bridging a gap in the higher education marketing literature, which has often focused on institutional branding or student perceptions alone. Furthermore, the results underscore the need to conceptualize sustainability in international recruitment beyond environmental dimensions, emphasizing financial accessibility, institutional resilience, and social equity [2,50].

According to the study’s findings, PR mostly influences IE via the social pathway, with SL emerging as the only significant mediator. PR has a significant impact on academic image, but this approach does not increase enrollment intentions. The size of the impacts indicates that while financial support plays a role as an independent predictor, socially motivated views play a major role in students’ decision-making processes. In general, measures that improve and express the social aspects of campus life should be given top priority by universities looking to improve enrollment results.

From a managerial and policy perspective, findings suggest several strategic directions for higher education institutions in emerging destinations [51]. Universities should institutionalize long-term partnerships with recruitment agents, provide them with structured resources such as alumni testimonials and scholarship schemes, and embed sustainability initiatives into the institutional brand to appeal to globally conscious students [51,52]. The study also highlights the untapped potential of alumni engagement as a channel for sustainable recruitment, reinforcing prior work on the importance of alumni as brand ambassadors [53]. By balancing academic quality, vibrant social experiences, and financial accessibility, universities can strengthen their competitiveness in an increasingly globalized education market.

7. Limitations/Future Research

The research focused on prospective students recruited through agents in North Cyprus, which may limit the generalizability of findings to other contexts. Future studies could employ longitudinal designs to track actual enrollment behaviors, expand the scope to include other emerging destinations, or explore the digitalization of agent-student interactions. Despite these limitations, the study provides actionable insights for both scholars and practitioners, offering a framework for sustainable and inclusive international recruitment strategies that align with the United Nations’ Sustainable Development Goals.

This study has several limitations. First, the sample was restricted to prospective students targeting institutions in North Cyprus, which may limit the generalizability of the findings. Second, reliance on agent-mediated recruitment channels may introduce self-selection bias. Third, the cross-sectional design does not allow causal inferences. Future research should therefore employ longitudinal designs, compare across multiple host countries, and include indicators capturing sustainable practices such as scholarships, diversity initiatives, and community partnerships.