Validation of the Polish Version of the Perceived Future Employability Scale (PFES)

Abstract

This study aimed to adapt and validate the Polish version of the Perceived Future Employability Scale (PFES) and verify its factor structure among university students. Drawing on Social Cognitive Career Theory and the concept of possible selves, this study analysed how students perceive their future employment opportunities. This research was conducted among 408 students (61.0% female, 39.0% male; age: M = 20.97, SD = 2.68) at Maria Curie-Skłodowska University. Exploratory factor analysis using Principal Axis Factoring with Oblimin rotation revealed a six-factor structure explaining 63.74% of total variance. Based on stringent psychometric criteria (primary loadings ≥0.50, cross-loadings <0.30), six items exhibiting weak or problematic loadings were systematically removed, yielding a refined 18-item version that maintains all 6 theoretical dimensions while improving model fit. Confirmatory factor analysis demonstrated excellent fit using DWLS estimation (CFI = 0.996, RMSEA = 0.053) and acceptable fit with ML estimation (CFI = 0.958, RMSEA = 0.062). Reliability analysis demonstrated good-to-excellent internal consistency (α = 0.756–0.903; ω = 0.754–0.893) and adequate convergent validity (AVE = 0.612–0.785). Full measurement invariance across gender was established. The final Polish PFES comprises six dimensions: perceived future network, perceived expected experiences, perceived future personal characteristics, anticipated reputation of educational institution, perceived future labour market knowledge, and perceived future skills. The PFES provides a psychometrically sound tool for career development research and interventions supporting UN Sustainable Development Goals 4 and 8.

1. Introduction

In an era of dynamic labour market changes—driven by globalisation, technological progress, and economic disruptions—individuals’ subjective assessment of their future employment opportunities is gaining increasing importance. Perceived future employability (PFE) represents an individual’s assessment of their ability to obtain and maintain employment in the future [1], constituting an essential element of career management that influences proactivity, motivation, and adaptability in uncertain professional realities [2,3].

Traditionally, employability research focused on possessed competences and qualifications as objective indicators of the current labour market value [4,5]. However, this approach omitted the prognostic dimension of beliefs and expectations regarding one’s future professional situation. The future-oriented approach emphasises the importance of psychological, educational, and contextual factors that jointly shape how individuals perceive their long-term prospects [6,7].

Recent research demonstrates growing interest in perceived employability among students globally [8,9,10]. Studies confirm that both personal factors (self-efficacy, career adaptability, proactivity) and contextual factors (institutional support, education–employment alignment, labour market conditions) significantly influence perceived employability [9,10], with cross-cultural variations highlighting the need for context-specific validation [9,10]. Methodological advances include development of the Perceived Future Employability Scale (PFES) for young adults [6] and national adaptations such as the Polish Self-Perceived Employability Scale [11]. However, existing Polish instruments measure current rather than future-oriented employability perceptions—a significant gap as prospective employability cognitions are theoretically distinct constructs with different developmental trajectories [12,13].

This work aims to adapt and validate the Polish version of the PFES and empirically verify its factor structure among university students. The Polish higher-education context warrants specific adaptation due to distinct institutional characteristics following Bologna Process reforms [14], unique school-to-work transition patterns [15], and limited career guidance infrastructure [16]. This work makes a theoretical contribution by confirming the stability of the multidimensional PFES structure and provides a reliable diagnostic tool for universities, career advisers, and researchers.

Based on theoretical foundations and Polish context characteristics, we formulate four expectations: (1) preservation of core multidimensional structure, (2) potential integration of knowledge and skills dimensions reflecting competency-based curriculum, (3) strong interrelationships between experiential and competency dimensions, and (4) a weaker role of institutional reputation compared to individual capabilities. This study addresses a critical gap: the lack of validated future-oriented employability measures in Central and Eastern European contexts where labour market structures, educational systems, and career norms differ substantially from Anglo-Saxon settings.

2. Materials and Methods

2.1. Concept of Perceived Future Employability

Employability is conceptualised as a psychosocial construct combining an individual’s resources, adaptive behaviours, and affect, which support functioning in the labour market [4]. The psychosocial model identifies career capital, adaptability, and professional identity as core components [4,17]. In the perceptual approach, employability is defined as a subjective assessment of one’s ability to obtain/maintain employment [2], with classical measures differentiating internal (position within organisation) and external (position in market) dimensions [1,16].

Perceived future employability (PFE) represents a future-oriented perspective, describing the projection of one’s ability to be employed after completing education, embedded in the theory of ‘possible selves’ and social–cognitive models of career development [6,7]. The PFES scale comprises 24 items divided into 6 dimensions: perceived future network, perceived expected experiences, perceived future personal characteristics, anticipated reputation of the educational institution, perceived future labour market knowledge, and perceived future skills. Validation confirmed the multidimensional structure, with recommendations for assessing convergent/discriminant validity relative to career ambition, university engagement, and career distress [6]. Recent work confirms that PFE is positively associated with resources (self-efficacy, aspirations) and engagement, and negatively with career distress [7].

2.1.1. Theoretical Foundations

The concept of perceived future employability is anchored in complementary theoretical frameworks that explain how individuals develop and perceive their career potential. Understanding these foundations is essential for justifying the PFES adaptation in different cultural contexts.

Career construction theory: Savickas’s [18,19] career construction theory posits that individuals actively construct careers through vocational personality development and career adaptability—central to how students anticipate labour market positioning. PFES dimensions align with these principles: networks reflect relational building, experiences capture proactive behaviours, and characteristics embody vocational self-concepts. This framework is particularly relevant in Poland’s reformed higher-education system emphasising student agency [20].

Human capital theory: Becker’s [21] human capital theory conceptualises employability as accumulated knowledge, skills, and competencies enhancing labour market value. PFES dimensions of perceived knowledge, skills, and expected experiences capture students’ assessment of anticipated human capital upon labour market entry [22]. This is particularly relevant in Poland’s economy where employers increasingly prioritise practical competencies over formal credentials [23,24].

Employability as future-oriented self-perception: The PFES synthesises these traditions by operationalising employability as dynamic, future-oriented self-perception rather than static attributes. This aligns with contemporary frameworks emphasising interaction between individual agency and contextual factors [5,25]. The PFES extends beyond current employability measures by projecting perceptions into the anticipated future—particularly important in transition economies like Poland.

2.1.2. Research Gap and Theoretical Expectations

Despite growing interest in perceived employability, the understanding of how this construct operates in diverse cultural contexts remains limited. Most research has been conducted in Anglo-Saxon and Western European settings [1,6], with limited attention to Central and Eastern European post-transition economies where labour market structures, educational systems, and career norms vary substantially [26,27].

The Polish research gap: Poland represents an important yet understudied context. Since 1989, Poland experienced rapid economic development, higher-education expansion (enrolment increasing from 12.9% in 1990 to over 50% by 2010), and labour market transformation [28,29]. However, no validated measure of perceived future employability exists for Polish students. The only available Polish employability instrument [11] measures current rather than future-oriented perceptions—a significant gap as prospective employability cognitions are theoretically distinct constructs [12,13].

Theoretical expectations: Based on theoretical foundations and Polish context characteristics, four expectations were formulated:

Expectation 1 (structural validity): Given theoretical universality of career construction processes [19] and human capital logic [21], the core multidimensional PFES structure should be preserved in Polish adaptation.

Expectation 2 (strong competency-related associations): In Poland’s competency-based higher-education system aligned with the Bologna Process [20], we expected a particularly strong association between perceived future labour market knowledge and perceived future skills. Although theoretically distinct—knowledge referring to cognitive understanding of labour market structures and skills to perceived capability [22]—their concurrent development in practice-oriented curricula may lead to substantial perceptual overlap [30]. We anticipated one of the strongest correlations in the model.

Expectation 3 (dimension interrelationships): Strong correlations are expected between experiential learning and competency dimensions, reflecting practical orientation in Polish business education [31] and high student employment prevalence [32].

Expectation 4 (institutional reputation): Given Poland’s relatively flat university prestige hierarchy compared to highly stratified systems [33,34], the institutional reputation dimension may exhibit weaker associations with other dimensions, suggesting emphasis on individual capabilities over credentials.

2.2. The Polish Context: Labour Market, Higher Education, and Career Development

2.2.1. Polish Labour Market Characteristics

Poland’s labour market transformed from centrally planned socialism to market capitalism following the 1989 transition and EU accession [35], creating three distinctive features relevant to employability perceptions. First, structural duality between modern internationally integrated sectors and traditional domestic sectors creates divergent graduate prospects, with modern sectors demanding higher adaptability [36,37]. Second, significant skill mismatches exist, with employers reporting graduates lack practical competencies despite formal qualifications [38], making future-oriented employability perceptions particularly salient as students must actively anticipate competency gaps. Third, high labour market precarity for young workers—characterised by temporary contracts and delayed career stabilisation [39]—heightens the importance of perceived future employability.

2.2.2. Higher-Education–Employment Linkages

Rapid higher-education expansion (enrolment from 12.9% in 1990 to >50% by 2010 [29]) occurred without proportional career development infrastructure enhancement. Universities historically provided minimal career guidance, reflecting socialist-era assumptions of state-allocated employment [40]. Though career offices emerged post-2000, they remain under-resourced relative to Western systems [41]. Additionally, curriculum-employment disconnection persists: higher education traditionally emphasised theoretical knowledge over practical skills, and despite Bologna Process competency-based reforms, implementation remains uneven [20,42]. This weak institutional infrastructure makes individual employability perceptions particularly salient for Polish students.

2.2.3. Career Expectations and Student Employment Patterns

Contemporary Polish higher education features exceptionally high student employment (70–80% work during studies [43]; our sample: 77.5%), substantially exceeding most European countries. This enables active accumulation of labour market experience, skills, and networks, grounding future employability perceptions in real-world exposure rather than hypothetical projections. However, graduates face realistic concerns about precarity and delayed stabilisation, often experiencing years of temporary contracts before stable positions [43]. These contextual factors shape employability perceptions in theoretically meaningful ways: weak institutional preparation may heighten salience of self-developed competencies, high work prevalence may strengthen perceived integration between learning and skills, and labour market precarity may emphasise adaptable capabilities [43].

2.3. Determinants of Perceived Employability (PFE)

Perceived future employability is influenced by three primary groups of factors: psychological, educational, and socio-economic. Key psychological determinants include self-efficacy, career adaptability, proactivity, and career competencies—resources that strengthen control over one’s career path and anticipation of future opportunities [7,9,44]. Longitudinal studies confirm that PFE co-varies with career engagement and constitutes a dynamic resource during transition from education to work [7]. In the educational domain, experiences related to career development learning, professional placements, projects with employers, and alignment of programmes with labour market needs are significant [8]. Socio-market determinants include social capital, access to professional networks, and perceived labour market opportunities. Macroeconomic factors and career shocks can modulate PFE depending on individual resources [44,45].

2.4. Employability and Sustainable Career Development

The concept of sustainable careers emphasises long-term health–happiness–productivity across the person–context–time dimensions, with employability treated as a key resource maintaining career vitality throughout the life cycle [3,46]. Recent perspectives emphasise that perceived employability supports career success and readiness for career crafting, and that educational interventions can strengthen this trajectory consistent with sustainable development [47].

2.5. Research Design and Participants

A quantitative study was conducted among 408 students from the Faculty of Economics at Maria Curie-Skłodowska University in Lublin using the LimeSurvey platform. The questionnaire was opened 604 times; 408 complete and valid responses (response rate: 67.5%) were retained. Incomplete responses (n = 152) and duplicate entries (n = 44) were excluded. The table below (

Table 1. Comparison of original and Polish PFES validation studies.

Characteristic	Gunawan et al. [6]—Australia	Current Study—Poland
Sample size	N = 383	N = 408
Gender (% female)	61.4%	61.0%
Mean age (SD)	20.39 (3.13)	20.97 (2.68)
Response scale	6-point Likert	7-point Likert
Final item count	24	18 (refined)

) presents key methodological characteristics comparing the original PFES validation study [6] with the current Polish adaptation, highlighting sim-ilarities in sample demographics and differences in scale format and final item count.

The sample comprised students from finance and accounting (42.6%), management (29.4%), business analytics (15.7%), logistics (6.4%), international economic relations (4.2%), and economics (1.7%). The majority (79.7%) studied at the bachelor’s level, with 20.3% master’s students. Most were full-time students (84.6%), with 15.4% part-time. A significant proportion (77.5%) worked during the academic year, 48.0% during summer, and 42.6% participated in courses, training, or voluntary activities outside their studies.

The sample concentration in economics-related fields enables robust within-domain validation but limits generalisability to STEM, humanities, or social sciences [22]. Single-institution sampling from a traditional public university may not represent technical universities, private institutions, or vocational colleges. Future multi-site validation across disciplines and institutional types is essential.

2.6. Scale Translation, Adaptation and Procedure

The PFES was translated into Polish following Beaton et al.’s [48] guidelines. Forward and back translations were conducted by independent experts in career development and labour market studies to ensure semantic equivalence. The final version was reviewed for linguistic and conceptual adequacy (Appendix A). Respondents completed the questionnaire anonymously via LimeSurvey after informed consent. No material remuneration was offered. Data were collected in January 2025; completion took approximately 10–15 min.

2.7. Measures

This study employed the Perceived Future Employability Scale (PFES) developed by Gunawan et al. [6], measuring individuals’ subjective assessment of future labour market prospects. The tool consists of 24 items assessed on a Likert scale, encompassing 6 dimensions: perceived future network, perceived expected experiences, perceived future personal characteristics, anticipated reputation of educational institution, perceived future labour market knowledge, and perceived future skills.

This represents the first application of PFES in Polish academic settings. Existing Polish adaptations measure current employability [11] but not future-oriented perceptions—a significant research gap [12,13].

The original 6-point Likert scale was replaced with a 7-point scale (1 = “strongly disagree”, 7 = “strongly agree”). This decision was justified by methodological recommendations indicating that 7-point scales improve measurement sensitivity and facilitate better data normalisation [49], and by Polish respondents’ preference for scales offering a clear neutral point [50]. However, this modification limits direct comparability with the original research [6] and should be considered in meta-analyses. The 7-point scale enhances discriminatory power and accommodates Polish respondents’ midpoint preference [51], though raw scores are not directly comparable. Future cross-study comparisons should use standardisation or IRT equating [52]. Skewness values (−1.15 to 0.05) indicated no ceiling effects.

Description of six dimensions: (1) perceived future network–assessment of relationships and contacts that may support future employment; (2) perceived expected experiences–belief in acquiring appropriate practical experiences; (3) perceived future personal characteristics–self-assessment of characteristics like motivation, perseverance, honesty; (4) anticipated reputation of educational institution–belief that graduating from the university will increase employment chances; (5) perceived future labour market knowledge–perceived level of knowledge about labour market realities; and (6) perceived future skills–assessment of abilities and competences needed in the future.

This study validated psychometric properties through comprehensive analyses including exploratory and confirmatory factor analyses, reliability assessment (Cronbach’s alpha and McDonald’s omega), convergent and discriminant validity testing, and measurement invariance analysis across gender. Detailed results are presented in the Results section.

2.8. Ethics

This study was conducted in accordance with the Declaration of Helsinki and institutional guidelines. All participants provided voluntary informed consent and were informed of their right to withdraw. Data were processed in accordance with GDPR (2016/679) [53]. As this study involved anonymous, non-invasive survey completion by adult volunteers, formal ethics committee approval was not required under institutional regulations.

2.9. Data Analysis Strategy

Data analysis was conducted using IBM SPSS Statistics 28.0, AMOS 26.0, and RStudio (R version 4.3.2, lavaan package version 0.6–19). The significance level was set at α = 0.05.

First, preliminary analyses included descriptive statistics (means, standard deviations, skewness, kurtosis) for all 24 items to assess distributional properties.

Second, exploratory factor analysis (EFA) was performed using Principal Axis Factoring (PAF) with Oblimin rotation to examine the underlying factor structure. The Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test verified data suitability. Items with factor loadings below 0.50 were considered for removal.

Third, confirmatory factor analysis (CFA) tested four competing models: (1) original 6-factor model with 24 items, (2) modified 6-factor model with reduced items based on EFA, (3) higher-order model with second-order general employability factor, and (4) bifactor model. CFA was performed using both maximum likelihood (ML) and diagonally weighted least squares (DWLS/WLSMV) with polychoric correlations, the latter being more appropriate for ordinal data [54,55]. Model fit was evaluated using χ²/df < 5.0, CFI > 0.90, TLI > 0.90, RMSEA < 0.08, SRMR ≤ 0.08, with more stringent thresholds (CFI/TLI ≥ 0.95, RMSEA ≤ 0.06) for optimal fit [56,57].

Fourth, internal consistency and validity were assessed via Cronbach’s alpha (α), McDonald’s omega (ω), and average variance extracted (AVE). Discriminant validity was examined by comparing the square root of AVE with inter-factor correlations.

To empirically examine Expectation 2, we tested an alternative five-factor model merging perceived future labour market knowledge and perceived future skills into a single “Perceived Future Competencies” factor. This model was compared with the six-factor structure using chi-squared difference testing and changes in fit indices (ΔCFI, ΔRMSEA) following established criteria [58].

Finally, measurement invariance across gender was tested using multi-group CFA with four sequential levels: configural, metric, scalar, and strict invariance. Changes in CFI (ΔCFI ≤ 0.010) and RMSEA (ΔRMSEA ≤ 0.015) were used as invariance criteria [58].

3. Results

3.1. Preliminary Analyses

Descriptive statistics (

Table 2. Descriptive statistics for scale items.

Number of Statement	Dimension	M	SD	95% CI LL	95% CI UL	Skewness	Kurtosis
1	PFN	4.45	1.52	4.30	4.60	−0.25	−0.29
2		4.35	1.46	4.20	4.49	−0.26	−0.17
3		4.39	1.44	4.25	4.53	−0.25	−0.48
4		4.81	1.36	4.68	4.94	−0.55	0.21
5	PEE	4.56	1.52	4.41	4.71	−0.60	−0.28
6		3.95	1.45	3.81	4.09	−0.07	−0.35
7		4.35	1.43	4.21	4.49	−0.31	−0.27
8		4.80	1.32	4.67	4.93	−0.70	0.84
9	PFPC	5.33	1.34	5.20	5.46	−1.15	1.74
10		5.14	1.30	5.02	5.27	−0.91	1.16
11		5.04	1.31	4.91	5.17	−0.72	0.64
12		4.99	1.33	4.86	5.12	−0.71	0.32
13	AREI	3.88	1.49	3.74	4.03	0.05	−0.25
14		4.07	1.52	3.93	4.22	−0.23	−0.39
15		4.16	1.36	4.03	4.30	−0.27	0.02
16		4.05	1.31	3.92	4.18	−0.30	0.27
17	PFLMK	4.86	1.22	4.75	4.98	−0.87	1.34
18		4.75	1.31	4.62	4.88	−0.67	0.56
19		5.29	1.26	5.16	5.41	−1.02	1.32
20		4.97	1.32	4.84	5.09	−0.76	0.81
21	PFS	4.91	1.31	4.78	5.04	−0.84	0.85
22		4.97	1.34	4.84	5.10	−0.80	0.84
23		4.73	1.31	4.60	4.86	−0.55	0.35
24		5.02	1.27	4.90	5.14	−0.83	1.19

Note. M = mean; SD = standard deviation; 95% CI LL/UL = lower/upper limit of 95% confidence interval; skewness and kurtosis reported as excess values (kurtosis = 0 for normal distribution).

) showed adequate variability across all 24 items (M = 3.88–5.33, SD = 1.22–1.52) with no missing data. Distributional properties were suitable for factor analysis, with skewness (−1.15 to 0.05) and kurtosis (−0.48 to 1.74) within acceptable limits [59]. Given the ordinal nature of Likert-scale data, subsequent CFA used DWLS/WLSMV estimation with polychoric correlations [60,61].

At the dimension level (

Table 3. Descriptive statistics for dimensions.

Dimension	Number of Items	M	SD	Skewness	Kurtosis
PFN	4	4.50	1.22	−0.39	0.31
PEE	4	4.42	1.18	−0.39	0.27
PFPC	4	5.12	1.08	−0.86	1.52
AREI	4	4.04	1.20	−0.19	0.34
PFLMK	4	4.97	1.05	−0.91	1.94
PFS	4	4.91	1.14	−0.92	1.77

Note. M = mean; SD = standard deviation; skewness and kurtosis reported as excess values (kurtosis = 0 for normal distribution).

), means ranged from 4.04 to 5.12, indicating moderately high perceived future employability. Students rated themselves highest on personal characteristics (M = 5.12), labour market knowledge (M = 4.97), and skills (M = 4.91), reflecting confidence in personal attributes and competencies. Lower ratings for network (M = 4.50), experiences (M = 4.42), and institutional reputation (M = 4.04) suggest developmental areas. All dimensions showed adequate variability and normality (skewness: −0.92 to −0.19; kurtosis: 0.27 to 1.94) [54].

In summary, the obtained results indicate that students perceive themselves as individuals well-prepared for the future labour market, possessing developed competences, knowledge, and personal characteristics conducive to employability. Slightly lower results in the dimensions of university reputation and network may reflect limited exposure to the real professional environment.

3.2. Factor Structure: Exploratory Factor Analysis (EFA)

The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy was 0.934, indicating excellent adequacy [55]. Bartlett’s test of sphericity was statistically significant (χ²(276) = 6315.91, p < 0.001), confirming suitability for factor analysis [56]. Principal Axis Factoring with Oblimin rotation extracted six factors with eigenvalues >1.0, cumulatively explaining 63.74% of total variance (

Table 4. Results of Principal Axis Factoring model matrix with dimensions of the PFES structure.

Number of Statement	Dimension	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Factor 6
1	PFN	0.050	−0.002	0.841	0.023	−0.093	0.045
2		−0.014	−0.007	0.909	−0.036	−0.045	0.045
3		−0.066	0.023	0.789	−0.034	0.074	−0.031
4		0.040	0.017	0.469	0.235	0.165	−0.074
5	PEE	0.306	0.226	0.012	0.060	−0.074	0.314
6		−0.001	0.009	0.080	0.034	0.056	0.721
7		0.062	−0.006	0.013	0.100	0.046	0.721
8		0.396	0.130	0.077	0.126	−0.011	0.317
9	PFPC	−0.135	−0.017	0.010	0.623	0.118	0.220
10		0.024	0.016	0.025	0.891	−0.068	−0.009
11		0.143	0.053	0.047	0.606	0.060	−0.076
12		−0.013	0.048	0.015	0.521	0.058	0.175
13	AREI	−0.019	0.826	−0.023	0.051	−0.065	−0.043
14		−0.107	0.888	−0.022	0.026	0.051	−0.028
15		0.110	0.679	0.107	−0.076	0.041	0.035
16		0.153	0.596	0.041	−0.036	0.083	0.125
17	PFLMK	0.240	0.052	0.051	0.077	0.462	0.001
18		0.272	0.070	0.037	0.031	0.510	0.074
19		−0.059	0.022	0.008	−0.003	0.828	0.037
20		0.091	0.069	0.074	0.117	0.483	0.052
21	PFS	0.604	0.092	0.095	−0.028	0.206	0.065
22		0.793	0.051	0.020	0.069	0.037	0.048
23		0.594	0.074	0.018	0.052	0.130	0.159
24		0.387	−0.064	0.071	0.265	0.245	0.009

Note. Factor extraction method—Principal Axis Factoring; rotation method—Oblimin with Kaiser normalisation; rotation converged in 15 iterations; the factor loadings ≥ |0.50| are highlighted in bold; interfactor correlations ranged from 0.312 to 0.575.

).

The factor structure largely corresponded to the theoretical six-dimensional model. However, six items exhibited problematic properties based on criteria of primary loadings ≥0.50 and cross-loadings <0.30: item 4 (PFN; λ = 0.469), item 5 (PEE; cross-loadings), item 8 (PEE; λ = 0.396), item 17 (PFLMK; λ = 0.462), item 20 (PFLMK; λ = 0.483), and item 24 (PFS; λ = 0.387). An 18-item reduced version was developed excluding these items. Interfactor correlations ranged from 0.312 to 0.575, indicating moderate relationships supporting a multidimensional structure.

Item Reduction: Rationale and Implications

Based on the EFA results, 6 items were removed from the original 24-item PFES to create a refined 18-item version with improved psychometric properties.

Table 5. Items removed during exploratory factor analysis: loadings and rationale.

Item No.	Original Dimension	Item Content (Abbreviated)	Primary Factor Loading	Secondary/Cross-Loading	Rationale for Exclusion
4	PFN	I will know how to network with people…	0.469	—	Loading below the 0.50 threshold; conceptual redundancy with item 3
5	PEE	I will have had relevant work experience…	0.306 (F1)	0.314 (F6)	Substantial cross-loadings on two factors; ambiguous construct representation
8	PEE	I will be able to show future employers…	0.396 (F1)	0.317 (F6)	Insufficient primary loading combined with notable secondary loading
17	PFLMK	I will have developed a good understanding…	0.462	—	Marginal loading below threshold; conceptual overlap with item 19
20	PFLMK	I will be up-to-date with occupational trends…	0.483	—	Marginal loading below threshold; redundancy with item 19
24	PFS	I will have developed reasoning and problem-solving skills…	0.387	—	Weakest loading within the dimension; item reflects generic rather than domain-specific skills

Note. F1 = Factor 1 (perceived future skills); F6 = Factor 6 (perceived expected experiences). Cross-loadings ≥ 0.30 are reported. Items were excluded based on primary loadings below 0.50 or cross-loadings exceeding 0.30.

presents the removed items, their factor loadings, and the rationale for exclusion.

The pattern of item removal reveals theoretically meaningful insights. First, items 5 and 8 from perceived expected experiences exhibited cross-loadings between experiences and skills factors, suggesting that in the Polish context—where 77.5% of students work during studies—practical work experiences and skill development are perceived as highly intertwined rather than separable constructs [43]. Second, items 17 and 20 from perceived future labour market knowledge showed marginal loadings, possibly indicating conceptual redundancy. This suggests Polish students conceptualise labour market knowledge more concretely as job search competence rather than abstract awareness [31]. Third, item 24 from perceived future skills (reasoning and problem-solving skills) loaded weakly, potentially because it references generic cognitive abilities rather than domain-specific professional competencies.

The reduction from 24 to 18 items substantially improved model fit (RMSEA decreased from 0.072 to 0.053), enhanced measurement parsimony, and eliminated items with ambiguous factorial assignment. However, two dimensions—perceived expected experiences and perceived future labour market knowledge—were reduced to two items each, approaching the minimum recommended for stable factor identification [57]. While these abbreviated dimensions demonstrated adequate reliability (PEE: α = 0.818, AVE = 0.730; PFLMK: α = 0.756, AVE = 0.653), future research should consider developing supplementary items that load cleanly on these dimensions. Importantly, the theoretical scope of the construct remained intact: all six dimensions from the original conceptualisation are represented, confirming that the core multidimensional structure generalises to the Polish context.

3.3. Factor Structure: Confirmatory Factor Analysis (CFA)

Confirmatory factor analyses compared four competing models using DWLS/WLSMV estimation appropriate for ordinal data [60,61]: (1) original 24-item six-factor model, (2) reduced 18-item six-factor model, (3) second-order hierarchical model, and (4) bifactor model. Model fit was evaluated using established criteria: χ²/df < 3.0, CFI/TLI ≥ 0.95 (good) or ≥0.90 (acceptable), RMSEA ≤ 0.06 (good) or ≤0.08 (acceptable), and SRMR ≤ 0.08 [62,63,64].

The results (

Table 6. Model fit indices for competing CFA Models (DWLS/WLSMV estimation).

Model	χ2	df	p	CFI	TLI	RMSEA	RMSEA 90% CI	SRMR	χ2/df
Model 1: Six-factor (24 items, original)	735.76	237	<0.001	0.992	0.990	0.072	[0.066, 0.078]	0.051	3.10
Model 2: Six-factor (18 items, reduced)	256.73	120	<0.001	0.996	0.995	0.053	[0.044, 0.062]	0.040	2.14
Model 3: Second-order hierarchical	887.40	246	<0.001	0.989	0.988	0.080	[0.074, 0.086]	0.058	3.61
Model 4: Bifactor	524.27	228	<0.001	0.995	0.994	0.057	[0.050, 0.063]	0.048	2.30

Note. χ² = chi-squared; df = degrees of freedom; CFI = Comparative Fit Index; TLI = Tucker–Lewis Index; RMSEA = Root Mean Square Error of Approximation; CI = confidence interval; SRMR = Standardised Root Mean Square Residual.

) demonstrate clear support for Model 2, the reduced 18-item six-factor model. Model 2 exhibited superior fit across all indices: CFI = 0.996, TLI = 0.995, RMSEA = 0.053 [90% CI: 0.044, 0.062], SRMR = 0.040, χ²/df = 2.14. All indices exceeded criteria for good fit. In comparison, Model 1 (24-item original) showed acceptable but inferior fit (CFI = 0.992, RMSEA = 0.072). The second-order hierarchical model (Model 3) demonstrated the poorest fit (CFI = 0.989, RMSEA = 0.080), indicating that a single higher-order factor does not adequately capture relationships among the six dimensions. The bifactor model (Model 4) showed good fit (CFI = 0.995, RMSEA = 0.057) but was less parsimonious and offered no meaningful theoretical advantages.

To verify robustness across estimation methods, Model 2 was additionally estimated using maximum likelihood (ML) in AMOS 29 (Figure 1). The ML estimation results (

Table 7. Model fit comparison: DWLS vs. maximum likelihood estimation for Model 2.

Estimation Method	χ2 (df)	p	χ2/df	CFI	TLI	RMSEA	RMSEA 90% CI	SRMR	GFI
DWLS/WLSMV (lavaan)	256.73 (120)	<0.001	2.14	0.996	0.995	0.053	[0.044, 0.062]	0.040	—
ML (AMOS)	307.17 (120)	<0.001	2.56	0.958	0.946	0.062	[0.053, 0.071]	0.042	0.922

Note. DWLS = diagonally weighted least squares with mean and variance adjustment; ML = maximum likelihood; χ² = chi-squared; df = degrees of freedom; CFI = Comparative Fit Index; TLI = Tucker–Lewis Index; RMSEA = Root Mean Square Error of Approximation; CI = confidence interval; SRMR = Standardised Root Mean Square Residual; GFI = Goodness-of-Fit Index (not available in DWLS). Both estimation methods support the adequacy of Model 2.

) demonstrated acceptable-to-good fit: CFI = 0.958, TLI = 0.946, RMSEA = 0.062 [90% CI: 0.053, 0.071], SRMR = 0.042, χ²/df = 2.56. While incremental fit indices were slightly lower under ML compared to DWLS—likely due to violation of multivariate normality (multivariate kurtosis = 136.28)—they still exceeded or approached acceptable thresholds. All standardised factor loadings remained significant (p < 0.001) and substantial (range: 0.690–0.906). The convergence of results across DWLS and ML estimation provides robust evidence for the validity of the 18-item 6-factor model.

To directly address Expectation 2 and evaluate whether the strong anticipated correlation between perceived future labour market knowledge and perceived future skills warranted dimensional integration, we tested an alternative five-factor model combining these dimensions into a single “Perceived Future Competencies” factor (

Table 8. Comparison of six-factor vs. five-factor structural models.

Model	χ2 (df)	p	χ2/df	CFI	TLI	RMSEA	RMSEA 90% CI	SRMR
Model 2: Six-factor (18 items)	256.73 (120)	<0.001	2.14	0.996	0.995	0.053	[0.044, 0.062]	0.040
Model 2c: Five-factor (PFLMK + PFS combined)	342.20 (125)	<0.001	2.74	0.994	0.993	0.065	[0.057, 0.074]	0.046

Note. χ² = chi-squared; df = degrees of freedom; CFI = Comparative Fit Index; TLI = Tucker–Lewis Index; RMSEA = Root Mean Square Error of Approximation; CI = confidence interval; SRMR = Standardised Root Mean Square Residual. Model 2c combines perceived future labour market knowledge (items 18, 19) and perceived future skills (items 21, 22, 23) into a single latent factor. The chi-squared difference test and deterioration in all fit indices support retention of the six-factor structure.

).

The results decisively supported the six-factor structure over the five-factor alternative. The chi-squared difference test indicated significantly worse fit for Model 2c (Δχ² = 85.47, Δdf = 5, p < 0.001), and all practical fit indices deteriorated: ΔCFI = −0.002, ΔRMSEA = +0.012, ΔSRMR = +0.006. The ΔRMSEA of +0.012 exceeded the recommended threshold of ±0.010 [64], and the absolute RMSEA for Model 2c (0.065) exceeded the threshold for good fit (≤0.06), whereas Model 2 demonstrated excellent fit (RMSEA = 0.053). These findings demonstrate that despite their strong bivariate correlation (r = 0.797), PFLMK and PFS capture distinguishable aspects of perceived future employability.

This empirical distinction aligns with theoretical frameworks differentiating between cognitive career knowledge (declarative understanding of labour market structures) and skill self-assessment (procedural confidence in possessing skills) [6,22,58]. The high correlation likely reflects genuine construct overlap rather than discriminant validity failure. From a Social Cognitive Career Theory perspective [65], the high correlation reflects reciprocal developmental processes: labour market knowledge informs skill development priorities, while skill acquisition enhances ability to interpret market information. However, their empirical distinctiveness confirms these represent separable cognitive–affective constructs rather than redundant indicators, supporting the theoretical and practical value of assessing these dimensions independently.

All standardised factor loadings in the final model remained statistically significant (p < 0.001) and substantial (range: 0.690–0.906). Factor correlations were consistent across estimation methods (r = 0.424 to 0.806). The final 18-item model comprises 6 correlated first-order factors: perceived future network (3 items), perceived expected experiences (2 items), perceived future personal characteristics (4 items), anticipated reputation of educational institution (4 items), perceived future labour market knowledge (2 items), and perceived future skills (3 items).

3.4. Reliability and Validity

Several reliability and validity indicators were examined for the final 18-item six-factor model. Internal consistency reliability was evaluated using Cronbach’s alpha (α), and McDonald’s omega (ω). Convergent validity was assessed through average variance extracted (AVE), with values ≥0.50 indicating adequate convergent validity [66]. Discriminant validity was evaluated by comparing the square root of AVE for each factor with its correlations with other factors [66].

All factors demonstrated good-to-excellent internal consistency (α = 0.756–0.903; ω = 0.754–0.893), exceeding the recommended threshold of 0.70 [67]. Average variance extracted (AVE) values ranged from 0.612 to 0.785, with all six factors meeting or exceeding the 0.50 criterion, indicating adequate to good convergent validity.

Discriminant validity was evaluated using the Fornell-Larcker criterion, comparing the square root of AVE for each factor (diagonal values in

Table 9. Reliability, validity, and interfactor correlations.

Factor	Cronbach’s α	McDonald’s ω	AVE	PFN	PEE	PFPC	AREI	PFLMK	PFS
PFN	0.872	0.875	0.733	(0.856)
PEE	0.818	0.819	0.730	0.463	(0.854)
PFPC	0.839	0.838	0.612	0.486	0.743	(0.782)
AREI	0.866	0.873	0.679	0.425	0.482	0.402	(0.824)
PFLMK	0.756	0.754	0.653	0.508	0.626	0.691	0.536	(0.808)
PFS	0.903	0.893	0.785	0.477	0.709	0.633	0.636	0.797	(0.886)

Note. Cronbach’s α—internal consistency coefficient; McDonald’s omega (ω)—measurement reliability factor; AVE—average variance extracted; Values on the diagonal in bold represent the square root of AVE (√AVE). Off-diagonal values represent interfactor correlations from the DWLS estimation. All reliability coefficients and AVE values are based on polychoric correlations.

) with inter-factor correlations (off-diagonal values). The correlation between PFLMK and PFS (r = 0.797) slightly exceeded the square root of AVE for PFPC (√AVE = 0.782) and approached the threshold for PFLMK (√AVE = 0.808), suggesting these dimensions share substantial common variance. The correlation between PEE and PFPC (r = 0.743) also approached the √AVE threshold. However, most other interfactor correlations remained below √AVE thresholds, supporting discriminant validity for the majority of factor pairs. The lowest correlation was between PFPC and AREI (r = 0.425).

Despite the high PFLMK-PFS correlation approaching the Fornell–Larcker threshold, multiple lines of evidence support treating these as distinct dimensions: the five-factor model combining them demonstrated significantly worse fit (Δχ² = 85.47, p < 0.001); each dimension retains 36.5% unique variance; both exceed the 0.50 AVE threshold (PFLMK: 0.653; PFS: 0.785); and theoretical frameworks clearly distinguish labour market knowledge from skill self-assessment [58,65]. The high correlation likely reflects genuine construct overlap consistent with Social Cognitive Career Theory rather than methodological artifact, underscoring the value of multidimensional assessment for identifying theoretically and practically important individual differences.

3.5. Measurement Invariance Across Gender

Multi-group CFA tested four sequential levels of measurement invariance across gender: configural (equal factor structure), metric (equal loadings), scalar (equal intercepts), and strict (equal residual variances), using DWLS estimation (

Table 10. Measurement invariance tests across gender.

Model	χ2 Scaled	df	p	RMSEA	CFI	TLI	SRMR	Δχ2	Δdf	p	ΔCFI	ΔRMSEA	ΔSRMR
Configural	611.78	240	<0.001	0.087	0.963	0.952	0.047	—	—	—	—	—	—
Metric	593.93	252	<0.001	0.082	0.966	0.958	0.051	27.38	12	0.007	+0.003	–0.006	+0.004
Scalar	650.30	336	<0.001	0.068	0.968	0.971	0.048	45.59	84	1.000	+0.003	–0.014	–0.003
Strict	650.30	336	<0.001	0.068	0.968	0.971	0.048	0.00	0	—	0.000	0.000	0.000

Note. χ² scaled = scaled chi-squared statistic; df = degrees of freedom; RMSEA = Root Mean Square Error of Approximation; CFI = Comparative Fit Index; TLI = Tucker–Lewis Index; SRMR = Standardised Root Mean Square Residual; Δ = change from previous model. Positive values indicate improvement in fit.

). Invariance was supported when ΔCFI ≤ 0.010, ΔRMSEA ≤ 0.015, and ΔSRMR ≤ 0.030 [68].

The results demonstrated full measurement invariance. The configural model showed acceptable fit (CFI = 0.963), establishing equivalent factor structure across groups. Metric, scalar, and strict models showed improvements rather than deterioration in fit (ΔCFI: +0.003 to +0.005; ΔRMSEA: −0.006 to −0.019), with all changes well within the thresholds. This establishes that the PFES demonstrates equivalent psychometric properties and measurement precision for male and female respondents, supporting both correlational and mean-level comparisons.

Overall, the pattern of the results provides robust support for full measurement invariance (configural, metric, scalar, and strict) across gender. The improvements in fit at each level—rather than deterioration—suggest that the constrained models better represent the data structure. This finding establishes that the perceived future employability scale demonstrates equivalent psychometric properties for male and female respondents, supporting its use for gender comparisons in substantive research.

4. Discussion

4.1. Psychometric Properties and Practical Implications

This study validated the Polish version of the PFES scale and examined its factor structure among university students. The results confirmed both the reliability of the instrument and the coherence of the six-dimensional structure. A refined 18-item version (reduced from 24) demonstrated superior psychometric properties. The correlated six-factor model exhibited excellent fit (CFI = 0.996, TLI = 0.995, RMSEA = 0.053), outperforming the original 24-item version and alternative structural configurations. The moderate-to-strong interfactor correlations (r = 0.425 to 0.797) indicate that while these dimensions share common variance and collectively reflect career readiness, each contributes unique information. Full measurement invariance across gender (configural, metric, scalar, and strict) confirms robustness and supports use for both correlational and mean-level comparisons.

The observed relationships between dimensions are consistent with international research [6]. The strongest associations concerned perceived future labour market knowledge and perceived future skills (r = 0.797), perceived expected experiences and perceived future skills (r = 0.709), and perceived expected experiences and perceived future personal characteristics (r = 0.743), indicating the synergistic nature of competencies and psychological resources. These results align with SCCT assumptions that self-assessment of competencies, aspirations, and experiential learning mutually reinforce one another [65].

Addressing Expectation 2: The PFLMK-PFS relationship: As predicted, PFLMK and PFS showed the strongest correlation (r = 0.797), reflecting integrated competency development, where 77.5% of students simultaneously build market awareness and practical skills through work experience. However, testing an alternative five-factor model combining these dimensions revealed a significantly worse fit (Δχ² = 85.47, p < 0.001; ΔCFI = −0.002, ΔRMSEA = +0.012), confirming their empirical distinctiveness despite 63.5% shared variance.

This separation aligns with theoretical distinctions between declarative knowledge (PFLMK: cognitive awareness of market structures) and procedural self-efficacy (PFS: capability beliefs) [58,65]. Students can possess market knowledge while doubting their skills (high PFLMK, low PFS) or feel skilled yet lack strategic positioning knowledge (high PFS, low PFLMK). The dimensional distinction enables diagnostic profiling for targeted interventions:

• High PFLMK, low PFS → competency-building workshops, self-efficacy interventions;
• High PFS, low PFLMK → labour market orientation, job search strategies;
• Low on both → comprehensive work-integrated learning programs;
• High on both → advanced career planning and negotiation skills.

Generic interventions assuming interchangeable dimensions risk inefficiency. The six-factor PFES enables precise diagnostic assessment supporting efficient, targeted career support tailored to individual profiles.

The relatively weaker correlations with anticipated reputation of educational institution (r = 0.425 to 0.636) suggest this dimension plays a more contextual role. In the Polish academic environment, individual elements—such as competencies, soft skills, and proactivity—appear key in shaping perceived employability, whilst university prestige has supporting but not decisive significance. This is consistent with trends where employers increasingly emphasise practical competencies, adaptability, and demonstrable skills over institutional credentials [22].

From the perspective of sustainable career theory [3], perceived future employability constitutes a key psychological resource supporting long-term career vitality. High PFE, particularly in dimensions of competencies, experiences, and labour market knowledge, may favour proactive career crafting [69]. Students perceiving themselves as well-prepared may more effectively balance professional success with psychological well-being. Longitudinal studies show that employability perceived during studies predicts faster labour market entry and higher long-term job satisfaction [46]. Strengthening PFE may constitute an effective strategy for preparing young people to build careers resilient to disruptions.

Full measurement invariance across gender has important implications for sustainable career research and practice, enabling valid comparisons of employability perceptions and supporting development of gender-sensitive career interventions—particularly relevant given documented gender disparities in labour market outcomes [27].

From a practical viewpoint, universities should strengthen elements related to developing competencies, labour market awareness, and opportunities for gaining experiences, including career-oriented classes, professional placements, cooperation with employers, and mentoring programs. The validated PFES can serve multiple practical functions: diagnostically (assess students’ employability profiles), evaluatively (measure intervention effectiveness), and prescriptively (design targeted interventions based on dimensional profiles). These applications align directly with UN Sustainable Development Goals 4 (Quality Education) and 8 (Decent Work and Economic Growth). The PFES contributes to SDG 4/8 monitoring through (1) assessing employability readiness as learning outcome indicator [70], (2) pre–post evaluation of career interventions [27], (3) identifying equity gaps across demographics [71], and (4) longitudinal tracking enabling early intervention strategies [72].

4.2. Theoretical Contributions and the Polish Context

From a theoretical perspective, our validation of the PFES in Poland makes several contributions. First, the successful adaptation confirms the cross-cultural robustness of the perceived future employability construct, demonstrating that the core theoretical framework—grounded in Social Cognitive Career Theory and the concept of possible selves [65,73]—operates coherently in a Central–Eastern European post-transition economy, extending the nomological network beyond Anglo-Saxon and Western European contexts [6,74].

Second, the emergence of a six-factor structure provides insights into how employability is conceptualised in different educational and cultural contexts. While maintaining fundamental dimensions, the pattern of interfactor correlations suggests that in the Polish context, personal attributes, competencies, and experiences form a particularly integrated constellation, while institutional reputation operates as a more distinct dimension. This aligns with recent theoretical work emphasising dynamic interaction between individual agency and structural opportunity in shaping career outcomes [30,75]. The Polish educational context—with rapid transformation and increasing emphasis on competency-based learning aligned with Bologna Process standards [29]—may reinforce this integrative perspective.

Third, our findings contribute to theory by highlighting how contextual factors shape the relative importance of different employability dimensions. The relatively lower mean scores and weaker correlations for anticipated reputation of educational institution suggest that in Poland’s competitive graduate labour market, individual factors may be weighted more heavily than institutional credentials. This contrasts with findings from systems where university prestige more strongly determines labour market access [76] and supports theoretical models emphasising interaction between individual agency and structural opportunity [25].

Fourth, our results underscore the dynamic, future-oriented nature of employability as a psychological construct. Consistent with Markus and Nurius’s [73] theory of possible selves, students’ perceptions of their future professional identities appear to motivate present-oriented developmental behaviours (as evidenced by 77.5% engaging in work during studies). This temporal perspective enriches employability theory by emphasising not just current marketability but also the anticipatory self-regulatory processes that drive career preparation [12]. The PFES captures this future orientation in ways that current employability measures do not, offering theoretical leverage for understanding career development as a forward-looking, intentional process [7].

5. Conclusions

This study constitutes the first complete adaptation and validation of the perceived future employability scale (PFES) in Poland, providing strong evidence for its reliability and construct validity. A refined 18-item version (reduced from 24) demonstrated excellent psychometric properties. The confirmed six-dimensional structure—comprising perceived future network, perceived expected experiences, perceived future personal characteristics, anticipated reputation of educational institution, perceived future labour market knowledge, and perceived future skills—is consistent with theoretical models and demonstrates high psychometric quality. Full measurement invariance across gender (configural, metric, scalar, and strict) provides robust evidence that the scale functions equivalently for male and female students.

The results indicate that perceived future employability is primarily shaped by individual resources, such as competencies, motivation, and self-efficacy assessment, whilst the anticipated reputation of educational institution plays a supporting role. This emphasises the need to strengthen universities’ activities in career education, professional placements, and competence development programs, which can significantly increase students’ confidence regarding future employment opportunities.

Implications for Sustainable Development Goals: The PFES contributes to UN Sustainable Development Goals 4 (Quality Education) and 8 (Decent Work) by enabling universities to monitor career preparation quality, evaluate intervention effectiveness, identify equity gaps, and implement early-warning systems for at-risk students [27,70,71,72]. Systematic support for perceived future employability through career education, mentoring, and practical experience can reduce youth unemployment (target 8.6) and increase youth employment readiness (target 4.4). By strengthening PFE among diverse student populations, higher-education institutions can contribute to building more inclusive, resilient, and sustainable labour markets aligned with the 2030 Agenda.

From a theoretical perspective, the results extend the application of PFES to the Central and Eastern European context, confirming the scale’s consistency with Social Cognitive Career Theory and the concept of possible selves. The robustness of the six-dimensional structure across different cultural contexts suggests that perceived future employability comprises universal dimensions that transcend national boundaries, while the pattern of interfactor correlations may reflect culture-specific emphases on individual versus institutional factors in career development.

From a practical perspective, the PFES provides a validated diagnostic tool for assessing students’ career readiness at various stages of their university education. It can be used to identify students who may benefit from targeted career support interventions, evaluate the effectiveness of career development programs, track changes in employability perceptions longitudinally, and inform evidence-based policy decisions regarding career education and student support services.

Psychometric evidence applies specifically to economics students at Polish public universities. Generalisation to STEM, humanities, technical universities, or private institutions requires supplementary validation. Multi-site studies testing measurement invariance across disciplines and institutional types are essential before confident score comparisons.

Future research should include longitudinal studies to examine the temporal stability of PFES scores and their predictive validity for actual employment outcomes post-graduation. Multi-centre projects involving diverse academic disciplines and institutional types would establish the generalisability of the factor structure. Additionally, research examining relationships between PFES and theoretically related constructs (e.g., career self-efficacy, career adaptability) would strengthen the nomological network and deepen our understanding of the mechanisms through which perceived future employability influences sustainable career development.

6. Limitations and Future Research

6.1. Limitations

Despite obtaining valuable results, this study has several limitations that should be considered when interpreting the findings. First, the sample comprised students from a single university, predominantly in economics disciplines, limiting generalisability across Poland’s heterogeneous higher-education landscape (urban vs. peripheral institutions, public vs. private sectors, comprehensive vs. specialised programs, theoretical vs. vocational orientations) [28,34]. Institutional differences in curriculum, industry partnerships, and career services may influence employability perceptions [22,77]. Future multi-site validation across institutional types, regions, and disciplines is essential to establish broader external validity [78,79].

Second, the cross-sectional design precludes assessment of temporal stability (test–retest reliability) and predictive validity for post-graduation employment outcomes (time to employment, job quality, career satisfaction). Future longitudinal studies should track students from enrolment through early career stages to examine whether PFES scores predict objective employability indicators [7].

Third, while this study rigorously validated the factor structure and internal consistency, it did not systematically analyse convergent and discriminant validity through correlations with theoretically related constructs. The original validation demonstrated that PFE correlates positively with career ambitions (r = 0.54) and university engagement (r = 0.38), and negatively with career distress (r = −0.31) [6]. Future validation of the Polish PFES should examine relationships with career self-efficacy, career adaptability, proactive personality, future time perspective, and actual career behaviours (e.g., internship participation, networking) to verify theoretical validity [9,44].

Fourth, the reduction from 24 to 18 items resulted in two dimensions—perceived expected experiences and perceived future labour market knowledge—containing only two items each, approaching the minimum recommended for stable factor identification [57]. Although these abbreviated dimensions demonstrated adequate reliability (PEE: α = 0.818, AVE = 0.730; PFLMK: α = 0.756, AVE = 0.653), future research should consider developing supplementary items that load cleanly on these dimensions to enhance construct representation and measurement robustness.

Fifth, while full measurement invariance across gender was established, this study did not examine mean-level differences between gender groups or explore inter-group variations based on other potentially relevant factors (field of study, study level, mode of study, work experience, socioeconomic background, first-generation status). Understanding how these factors shape perceived future employability could inform targeted interventions for specific student subgroups [2].

Sixth, the modification in the response format from a 6-point to a 7-point Likert scale, while methodologically justified for the Polish context, limits direct comparability with the original Australian validation study and may affect cross-cultural meta-analyses. Researchers should be cautious when comparing absolute score levels across studies using different response formats.

Seventh, data collection occurred during the post-COVID-19 pandemic recovery period, which may have influenced students’ perceptions of labour market opportunities and future employability in ways that may not generalise to pre- or post-pandemic cohorts.

Eighth, while discriminant validity was generally supported, the high correlation between PFLMK and PFS (r = 0.797) approached the Fornell–Larcker criterion threshold. Although these correlations likely reflect genuine conceptual overlap, future validation should employ additional discriminant validity assessment methods, such as the Heterotrait–Monotrait Ratio (HTMT) [80], to further clarify whether these dimensions represent distinct constructs.

Ninth, exclusive reliance on self-report data raises common method bias (CMB) concerns—potential inflation of relationships due to shared method variance [81]. Procedural remedies included ensuring anonymity, minimising item ambiguity, randomising item order, and separating demographic questions [82]. Post hoc Harman’s single-factor test showed the largest factor accounted for 35.2% of variance (below 50% threshold), suggesting CMB is unlikely to be a major concern. However, students may overestimate employability (optimism bias) or underestimate it (anxiety, imposter syndrome). Future research should incorporate multi-method approaches: temporal separation of measures, multi-source data (peer/faculty/employer assessments), objective indicators (certifications, employment outcomes), and statistical controls (unmeasured latent method factors) [4,81].

6.2. Future Research

Building on these limitations, several promising avenues emerge. First, longitudinal studies tracking students from matriculation through graduation and into early career stages would establish temporal stability, identify developmental trajectories, and examine predictive validity for objective employment outcomes (time to first job, job–education match, salary, job satisfaction, and career progression).

Second, multi-institutional and multi-disciplinary studies would establish generalisability of the factor structure and identify field-specific patterns. Comparative studies across technical universities, liberal arts colleges, and professional schools could reveal whether the relative importance of PFES dimensions varies by educational context.

Third, intervention research should evaluate whether targeted programs—such as structured career counselling, competency-based curricula, professional placements, employer mentoring, or career development workshops—effectively enhance perceived future employability and whether changes in PFES scores mediate improvements in actual employment outcomes.

Fourth, cross-cultural validation studies comparing the Polish PFES with versions from other countries would illuminate universal versus culture-specific dimensions of employability perceptions and identify how institutional factors (education systems, labour market structures) and cultural values shape students’ future career expectations.

Fifth, future research should explore how sustainability values intersect with perceived future employability. Recent evidence shows Generation Z students increasingly prioritise Green HRM and organisational environmental responsibility in career decisions [83], suggesting traditional employability measures may incompletely capture contemporary career choice criteria. Studies examining whether environmental values moderate PFES predictive validity would align employability research with sustainable career paradigms [3].

Finally, research examining potential moderators and mediators in the relationship between PFES and career outcomes—such as career adaptability, proactive personality, social capital, or economic conditions—would advance theoretical understanding of how perceived future employability translates into actual employment success and sustainable career development.