The purpose of this paper is to explore how graduate employability is conceptualized and operationalized in higher education through educational interventions. It provides a comprehensive review of how these interventions are implemented and the results they report as well as the levels of governance that motivate their implementation.
A systematic review of peer-reviewed publications was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
Interventions offering practical experiences had the greatest positive impact on graduate employability. Programs focused on employability skills and career planning showed mixed results, with limited links to employment outcomes. Interventions motivated via direct labor market demands were more effective, suggesting strong communication between higher education institutions and employers is pivotal. Conversely, government-mandated interventions often had inconclusive results, possibly due to weak institutional collaboration.
The study highlights the need for standardized evaluation of graduate employability interventions to enhance comparability and effectiveness.
Our findings should inform future graduate employability policy designs. Specifically, policies should promote effective communication and collaboration among policymakers, institutions and employers to improve outcomes, rather than relying solely on higher education institutions to achieve these goals.
This review offers insights for educators in selecting effective graduate employability interventions. It underscores the importance of effective communication and collaboration between policymakers, institutions and employers to improve outcomes.
Graduate employability (GE) has become a central theme in higher education (HE) policy discussions on human capital development and institutional strategy. The Bologna process, which expanded and standardized HE, has led to a surge in degree holders and heightened competition in the labor market (Leoni, 2014; Teichler, 2006, 2011). Consequently, there is growing concern that graduates must offer more than just academic credentials to attract employers (Jackson and Bridgstock, 2021; Mason et al., 2003).
McQuaid and Lindsay (2005) define GE broadly as the ability to secure employment post-graduation. While this definition holds general validity, a universal definition remains elusive due to variations across industries, required skills and (regional) labor market demands. Nevertheless, despite the absence of a universally accepted definition, policymakers continue to expect higher education (HE) institutions to integrate employability objectives into their curricula (Daubney, 2022; Masduki et al., 2022), even though the effectiveness of this approach remains uncertain (e.g. Morley, 2001; Teichler, 2006, 2011). How to operationalize GE within educational curricula remains largely at the discretion of the HE institutions, requiring these institutions to independently determine how to best achieve and measure GE improvement – e.g. successful labor market entry after graduation (e.g. Teichler, 2006, 2011).
As a result, HE institutions have been attempting to operationalize GE by implementing a wide array of interventions in the form of changes that are made to the curriculum on program or course level (Mason et al., 2009). These interventions, likely underpinned by diverse interpretations of what constitutes GE include, for example, the incorporation of practical elements into the curriculum (e.g. internships) or the development of specific soft skills (e.g. communication or teamwork skills) in classroom settings (e.g. Blackmore et al., 2016; Holmes, 2013; Mason et al., 2009; Noah and Aziz, 2020; Setor and Joseph, 2021). Because of these manifold operationalization attempts, the types of HE interventions and their effects on GE remain elusive.
This review aims at elucidating how GE is operationalized in HE by creating a comprehensive overview of the types of employability interventions implemented and their reported effectiveness. The study thereby distills how GE is conceptualized within interventions from the way it is put into practice. To achieve this, the review focuses exclusively on studies that explicitly examine employability enhancement as the primary objective of the intervention's design.
Such an overview is essential for the prioritization of efficacious GE interventions by HE institutions, the avoidance of the adoption of ineffective GE approaches and policy designs and to coin in on the role of HE in the implementation of GE in a providing role. Our study thereby addresses a gap identified by Scandurra et al. (2023), who highlighted the lack of such an overview of the types and effectiveness of employability interventions in HE. The overview will also provide insight into how HE institutions assess the effectiveness of the GE interventions they implemented and whether the outcome measures used to assess effectiveness align with the general GE definition of “the ability to obtain work after graduation.”
Additionally, this study also aims to identify the level of governance driving GE interventions in HE to assess potential patterns between the motivation behind an intervention and its reported outcomes. This inquiry adopts an exploratory approach, proposing that different motivating factors are likely to result in varying types of interventions and, consequently, diverse outcomes. The following four categories of governance levels – ranging from truly external to truly internal motivations – are inquired: (1) policy-driven GE interventions (Cairns et al., 2018), (2) government-driven interventions communicated to HE institutions without formal policy, (3) direct labor market demand, where institutions collaborate with local labor market agents and (4) inferred labor market demand, where HE institutions implement GE interventions based on their own judgment.
To facilitate a comparative analysis between the GE interventions in HE institutions, the current paper uses theory of change as a foundational principle. Theory of change is a general research principle by which an intervention is designed on the premise of achieving a specific (intended) outcome, and whereby actual and causal effects are observable and analyzable through evidence that the implementation process generates (e.g. Reinholz and Andrews, 2020; Rossi et al., 2018; UNDG, 2017). For example, a career management program might be designed to increase graduate employment rates, where the program’s success is measured by the number of graduates securing a job, demonstrating the causal link between the intervention and the desired outcome (Taylor and Hooley, 2014). Using the theory of change as a foundational principle accommodates the inclusion of a broad range of GE intervention studies without being limited to a specific pedagogical methodology or educational principle. Given the wide array of interventions that HE institutions have implemented through curricular adjustments – ranging from program-level to course-level changes (Mason et al., 2003) – such a general principle is essential. It allows for an inquiry into the overarching goals of GE interventions, including their design, contents and intended and reported outcomes. Using the theory of change framework also allows for an assessment of the adequacy of applied designs and conceptualizations of GE in included articles and their effectiveness, including the results reported – i.e. whether they show positive, inconclusive or negative results.
With the objective of uncovering how GE interventions are evoked and operationalized in HE, this study reviews peer-reviewed publications of implemented and evaluated interventions aimed at promoting GE within the confines of HE and the level of governance that underlies their emergence. The overview will also give insight into the effectiveness of the interventions discussed. In order to meet the mentioned objectives, the paper is guided by the following research questions:
What types of GE interventions are implemented in HE?
By which theory of change are the interventions implemented?
Which level of governance motivates their implementation?
Which results do the interventions report?
Methodology
Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021b), a systematic literature review was conducted, aiming at identifying all relevant publications of GE interventions that have been implemented within the confines of HE.
Search strategy and study selection
Three databases were searched to identify eligible publications across educational and multidisciplinary fields, listed in order from closely related to educational sciences to broader databases: ERIC, Scopus and Web of Science (Figure 1). This approach ensures a comprehensive review by also capturing relevant employability interventions published outside the behavioral sciences.
The PRISMA flow diagram titled “Identification of studies via databases and registers” and shows three section headings arranged vertically on the left side: “Identification,” “Screening,” and “Included.” The flowchart is divided into two columns. The left column contains five text boxes, which are labeled as follows: Text box 1: Records identified from asterisk: Databases (n equals 3.365). Text box 2: Records screened (n equals 2.325). Text box 3: Reports sought for retrieval (n equals 433). Text box 4: Reports assessed for eligibility (n equals 405). Text box 5: Studies included in review (n equals 72). The right column contains four text boxes, which are labeled as follows: Text box 6: Duplicate records removed through EndNote (n equals 1.039). Text box 7: Records excluded double asterisk (n equals 1.893). Text box 8: Reports not retrieved triple asterisk (n equals 28). Text box 9: Reports excluded: Out of Scope (n equals 37), Wrong Population (n equals 9), Conceptual Papers (n equals 113), Measuring Perceived Employability (n equals 146), Outside Curriculum (n equals 10), Comparative Studies (n equals 18). The text boxes 1 and 6 are placed under the heading “Identification,” the text boxes 2, 3, 4, 7, 8, and 9 are placed under the heading “Screening,” and the text box 5 is placed under the heading “Included.” Text box 1 is connected to text box 2 with a downward arrow. Text box 2 is connected to text box 3 with a downward arrow. Text box 3 is connected to text box 4 with a downward arrow. Text box 4 is connected to text box 5 with a downward arrow. Text box 1 is connected to text box 6 with a rightward arrow. Text box 2 is connected to text box 7 with a rightward arrow. Text box 3 is connected to text box 8 with a rightward arrow. Text box 4 is connected to text box 9 with a rightward arrow.Study selection process following PRISMA 2020 guidelines (Page et al., 2021a). Note: *ERIC (n = 540); Web of Science (n = 1.274) and Scopus (n = 1.551). **Papers screened for relevance by title and abstract. ***Not retrievable or later identified duplicates. The study selection flow diagram was created in accordance with the PRISMA 2020 statement
The PRISMA flow diagram titled “Identification of studies via databases and registers” and shows three section headings arranged vertically on the left side: “Identification,” “Screening,” and “Included.” The flowchart is divided into two columns. The left column contains five text boxes, which are labeled as follows: Text box 1: Records identified from asterisk: Databases (n equals 3.365). Text box 2: Records screened (n equals 2.325). Text box 3: Reports sought for retrieval (n equals 433). Text box 4: Reports assessed for eligibility (n equals 405). Text box 5: Studies included in review (n equals 72). The right column contains four text boxes, which are labeled as follows: Text box 6: Duplicate records removed through EndNote (n equals 1.039). Text box 7: Records excluded double asterisk (n equals 1.893). Text box 8: Reports not retrieved triple asterisk (n equals 28). Text box 9: Reports excluded: Out of Scope (n equals 37), Wrong Population (n equals 9), Conceptual Papers (n equals 113), Measuring Perceived Employability (n equals 146), Outside Curriculum (n equals 10), Comparative Studies (n equals 18). The text boxes 1 and 6 are placed under the heading “Identification,” the text boxes 2, 3, 4, 7, 8, and 9 are placed under the heading “Screening,” and the text box 5 is placed under the heading “Included.” Text box 1 is connected to text box 2 with a downward arrow. Text box 2 is connected to text box 3 with a downward arrow. Text box 3 is connected to text box 4 with a downward arrow. Text box 4 is connected to text box 5 with a downward arrow. Text box 1 is connected to text box 6 with a rightward arrow. Text box 2 is connected to text box 7 with a rightward arrow. Text box 3 is connected to text box 8 with a rightward arrow. Text box 4 is connected to text box 9 with a rightward arrow.Study selection process following PRISMA 2020 guidelines (Page et al., 2021a). Note: *ERIC (n = 540); Web of Science (n = 1.274) and Scopus (n = 1.551). **Papers screened for relevance by title and abstract. ***Not retrievable or later identified duplicates. The study selection flow diagram was created in accordance with the PRISMA 2020 statement
Keywords related to HE interventions and their synonyms were used in the search (see Supplementary Material 1 for the full search string). No commonly assumed components of employability were included as keywords, as the review focused exclusively on studies examining interventions explicitly designed to enhance employability itself, regardless of how each study operationalized the concept. This approach aligns with the study’s objective of analyzing how educationalists define and implement employability through specific HE interventions.
The literature search was limited to peer-reviewed, full-text scientific articles in the English language, published in the period between January 2000 and April 2022. The starting point of January 2000 was chosen as employability policies emerged in frequency after the Bologna convention in 1999 (McQuaid and Lindsay, 2005; Polenta, 2018; Teichler, 2011).
Out of the 2,325 articles from the initial database extraction that were screened based on title and abstract, 433 articles were retrieved for full-text screening. Only publications that report empirically tested educational interventions designed to enhance students’ employability within the confines of the HE curriculum were included in the analyses. N = 72 publications were found eligible for analysis. GE interventions that lie beyond the control of HE institutions – e.g. student jobs – or interventions that were only proposed conceptually but not tested empirically were excluded (n = 187). Studies solely measuring perceived employability were excluded (n = 146) since this metric only correlates with actual employment when students are adequately informed about their labor market prospects (Caricati et al., 2016). Since it remains unclear whether such information is provided before interventions, studies using objective measures – such as labor market outcomes, standardized tests and GPA – were prioritized.
Both stages of the selection process were conducted independently and verified by two researchers. A sample coding method was applied to 15% of the publications in the database to validate the selection process for fitness and demonstrability (Belur et al., 2021; Gwet, 2014; Marston, 2010). The interrater reliability score was calculated through the kappa statistic, resulting in κ = 0.752, which can be interpreted as substantial (as cited in McHugh, 2012; Marston, 2010).
The articles included in the analysis either measured enhanced employability as a direct (i.e. intervention to enhance employability) or indirect outcome of their interventions (e.g. intervention to enhance critical thinking skill, which enhances employability by implication) with diverse metrics, ranging from placement uptake to specific skill improvements.
Methodological quality assessment
In order to examine the methodological quality of the included articles, an adapted version of the Medical Education Research Study Quality Instrument (MERSQI) checklist, originally designed to assess the quality of medical education research studies, was used (Reed et al., 2007a). A total of 18 points can be achieved across six domains (see Table 1). One item in the MERSQI tool, which specifically inquires about study outcomes related to medical interventions, was replaced with labor market outcomes as the most direct approximation of GE. Individual MERSQI scores of publications have not been factored into the analysis; they serve as an indicator of the informative value of each analyzed publication (individual scores for each publication are visualized in Table 2 and Supplementary Material 2–10).
Paper count per item of the MERSQI tool (Adapted from Reed et al., 2007b)
| Domain | MERSQI item | Points per item | Paper count per item n = 72 | |
|---|---|---|---|---|
| Study design | 1. Study design | Single group cross-sectional or single group posttest only | 1 | 32 |
| Single group pretest and posttest | 1.5 | 11 | ||
| Nonrandomized, 2 group | 2 | 21 | ||
| Randomized controlled trial | 3 | 7 | ||
| Sampling | 2. No. of institutions | 1 | 0.5 | 56 |
| 2 | 1 | 1 | ||
| >2 | 1.5 | 15 | ||
| 3. Response rate in % | not applicable | 4 | ||
| <50 or not reported | 0.5 | 24 | ||
| 50–74 | 1 | 14 | ||
| >75 | 1.5 | 30 | ||
| Type of data | 4. Type of data | Assessment by study participant | 1 | 37 |
| Objective measurement | 3 | 35 | ||
| Validity of evaluation instruments | 5. Internal structure | not applicable | 2 | |
| not reported | 0 | 23 | ||
| Reported | 1 | 47 | ||
| 6. Content | not applicable | 1 | ||
| not reported | 0 | 7 | ||
| Reported | 1 | 64 | ||
| 7. Relationships to other variables | not applicable | 1 | ||
| not reported | 0 | 51 | ||
| Reported | 1 | 20 | ||
| Data analysis | 8. Appropriateness of analysis | Data analysis inappropriate for study design or type of data | 0 | 7 |
| Data analysis appropriate for study design and type of data | 1 | 65 | ||
| 9. Complexity of analysis | Descriptive analysis only | 1 | 34 | |
| Beyond descriptive analysis | 2 | 38 | ||
| Outcomes | 10. Outcomes | Satisfaction, attitudes, perceptions, opinions and general facts | 1 | 21 |
| Knowledge, skills | 1.5 | 13 | ||
| Behaviors | 2 | 11 | ||
| Labor market outcome | 3 | 27 | ||
| Maximum total points | 18 | Point Average: 11.5 |
Note(s): Adapted from “Association between funding and quality of published medical education research,” by Reed et al. (2007b)
Intervention core components and definitions
| Intervention core component | Definition (cumulative) | Intervention items (as reported) | Studies and MERSQI score |
|---|---|---|---|
| Linguistic Skill Development | Interventions that shall aid to foster verbal and written communication | Language Training; Semantic Enhancement | Dash et al. (2020) (14/18) Hiew et al. (2021) (13/18) Luka and Seniut (2019) (10.5/18) Murphy and Cunningham (2012) (14.5/18) Zakaria and Nair (2019) (11/18) |
| Didactical Innovation | In-classroom interventions that are specified to alter the teaching/delivery method of knowledge and skill, which fosters tacit knowledge and skill development that enhances employability skill | Teaching Innovation; Skill Assessment/Exam; Motivational Nudging | Cloete (2018) (14.5/18) Fowlie and Forder (2020) (10/18) Garcia et al. (2016) (14/18) Hyett et al. (2019) (8.5/18) Lim (2015) (9/18) Lopez-Carril et al. (2021) (12/18) Santos Rego et al. (2021) (12/18) Skaife et al. (2016) (7/18) Zulkifli et al. (2021) (13/18) |
| Advanced Programs | Programs for Students that exceed the generic curricular activity, and have a more extensive or intensive syllabus | Scholarship Program; Honors Program; Advanced Program; Interdisciplinary Training | Alexandre et al. (2009) (17/18) Clemente et al. (2020) (14/18) Felicetti (2017) (13.5/18) Liu et al. (2015) (10/18) Nghia et al. (2019) (15/18) Noah and Aziz (2020) (8/18) Pidgen and Moore (2018) (13/18) |
| Practical Experience | Practice in real life work placements that are embedded in the curriculum, and are built in collaboration with (local) enterprises | Internship; Field Work; Work Integrated Learning (WIL) | Bennett and Low (2021) (7/18) Alwi et al. (2020) (8/18) Bates et al. (2018) (11.5/18) Chen et al. (2018) (10.5/18) Chen (2022) (10/18) Feijoo et al. (2019) (9/18) Ferreras-Garcia et al. (2019) (10/18) Garwe (2020) (12.5/18) Goodenough et al. (2020) (13/18) Lloyd et al. (2019) (6.5/18) Morgan et al. (2018) (9.5/18) Purdie et al. (2013) (12/18) Rodzala and Saat (2018) (10.5/18) Setor and Joseph (2021) (16/18) Thune and Støren (2015) (13.5/18) Vamsi Krishna and Selwyn Babu (2021) (16/18) Watkins et al. (2018) (12/18) |
| International Experience | Knowledge and skill transfer of cultural and linguistic content other than domestic, in the form of a study semester at a university abroad | International Placement | DiPietro (2015) (13/18) Liwinski (2019) (16/18) Pinto (2022) (15/18) Potts and Kim (2021) (12/18) |
| Practical Simulation | Simulations of real-life work situations as a low-cost and low-risk alternative to full practical experience | Job Shadowing; Problem/Project Based Learning (PBL); Practical Simulation | Chad (2020) (10/18) Clarke et al. (2020) (8/18) Clausen and Andersson (2019) (6/18) Han et al. (2021) (12.5/18) Ismail et al. (2019) (10/18) Klegeris (2021) (11/18) Peng et al. (2021) (12/18) Power (2010) (12/18) Setor and Joseph (2021) (16/18) Wang et al. (2020) (11/18) |
| Career Planning Programs | Courses and content specifically designed to teach the navigation of job search as a skill | Preparatory Courses/Modules/Workshops for Labor Market Participation | Chang et al. (2022) (11/18) Cornish (2019) (7/18) Dray et al. (2011) (9/18) El-Temtamy et al. (2016) (14/18) Green et al. (2020) (14/18) Taylor and Hooley (2014) (14/18) Teychenne et al. (2019) (10/18) Von Konsky and Oliver (2012) (9.5/18) Williams et al. (2022) (12/18) Zakaria and Nair (2019) (11/18) |
| Personality Development Programs | Teaching methods and content that is designed to develop personality/personal qualities and competences which foster attractiveness of the individual in the labor market | Personality Courses/Modules | Caballero-Garcia and Sanchez Ruiz (2021) (15/18) Dacre Pool and Qualter (2012) (14/18) Gibbs and Slevitch (2019) (12.5/18) Gilar-Corbi et al. (2018) (14/18) Gomes da Costa et al. (2021) (14/18) Santos Rego et al. (2021) (12/18) Yousef (2021) (13.5/18) |
| Employability Skill Programs | Courses and modules within the curriculum that are designed around the employability topic through content. Such as employability skill development | Employability Courses/Modules/Workshops | Bradley et al. (2022) (9.5/18) Ito (2014) (9.5/18) Maguire and Guyer (2004) (8.5/18) Noah and Aziz (2020) (8/18) Secundo et al. (2019) (10.5/18) Shahabadkar et al. (2021) (12.5/18) Tomasson Goodwin et al. (2019) (11/18) Watkins et al. (2018) (12/18) |
Note(s): The definitions of the core components were synthesized by the authors from the intervention descriptions provided in the analyzed publications
Analysis
Guided by the theory of change framework, the papers were analyzed via a customized form and recorded with the Atlas.ti tool (Atlas.ti, version 5.8.0). The publications were examined for population and study design, intervention content, intended outcome, outcome measure and results reported (see Supplementary Material 2–10). The reported results were screened for having reported (1) positive-, (2) inconclusive or (3) negative results, based on whether intended outcomes and results matched fully, partially or whether inadvertent outcomes were reported. These results were further grouped by whether labor market metrics were tested or alternative measures within the curriculum were used. The level of governance behind each intervention was categorized into four groups, from external to internal motivation. Publications were classified based on keywords explaining the intervention's motivation: (1) Policy issued – if a specific policy was cited; (2) Governmental demand – if government involvement was mentioned; (3) Direct labor market demand – if employing institutions were partners and (4) Inferred labor market demand – if HE institutions implemented GE interventions based on their own judgment.
Given the variety of study designs, intended outcomes, content and metrics reported, a more parsimonious overview of the interventions was needed to create a level of comparability between the studies. Using an inductive approach, types of interventions were classified into distinct intervention core components based on the descriptions provided in the analyzed publications. Interventions were grouped into the same category if they shared a consistent and identifiable theme, such as focusing on practical experience or international exposure. The definitions of these core components, as outlined in Table 2, offer a comprehensive synthesis of the intervention descriptions found in the analyzed publications. Some studies explored multiple interventions, both within and across different intervention groups, and are thus listed multiple times in Table 2.
To further explore research question 4, three Fisher’s exact tests were executed to identify associations among intervention core components, reported results and level of governance. These tests yield statistical insights into the effectiveness of specific intervention core components. To exemplify, if statistical significance is absent, outcomes tend to be comparable across all intervention core components, thereby reducing the importance of selecting a specific intervention approach. Due to the small sample sizes within core components (e.g. n > 5), Fisher’s tests using Monte Carlo simulations of 1e+07 replicates were performed using R Studio (Version, 2023.03.1 + 446).
Results
Classification of intervention core components and level of governance
To address research question (1) and analyze the types of GE interventions implemented in higher education, the nine intervention core components identified will be discussed in the following chapter, in the sequence presented in Table 2. Additionally, the theory of change (Research Question 2) and the level of governance (Research Question 3) that informed their implementation will also be examined. A comprehensive overview of the relevant studies can be found in Supplementary Material 2–10.
Five studies implemented interventions for linguistic skill development, which were aimed at enhancing students’ verbal and written communication in English. All but one study were embedded in teaching English as a second language courses, postulating English language proficiency as a major contributor to GE. The intervention contents were specified for English language training, including grammar, vocabulary and speaking ability (e.g. Dash et al., 2020). As an example, Murphy and Cunningham (2012) administered daily crosswords for a month, arguing that semantic verbal fluency can be enhanced via simple, non-subject-specific interventions. Linguistic Skill Development interventions were implemented under various levels of governance, including direct policies, governmental demands or direct labor market demands.
Didactic innovation interventions (n = 9) were applied exclusively in theclassroom and ranged from applications of IT tools over diversified assessment tools, teaching innovations and peer activities. For example, Hyett et al. (2019) conducted a virtual intercultural learning activity to enhance students’ intercultural communication skills, which is assumed to enhance GE. Zulkifli et al. (2021) implemented an intervention on reciprocal peer tutoring via Facebook group chats for critical thinking skill development. Fowlie and Forder (2020) applied motivational nudges via personalized letters to prompt students to engage in industry placements. This strategic nudging, along with the resulting increase in placement uptake is expected to positively affect GE. None of the didactic innovation interventions were initiated via policies (i.e. truly externally motivated interventions). Five out of nine interventions were implemented as a response to inferred labor market demands.
Studies that applied a form of advanced program intervention (n = 7) include scholarship programs, interdisciplinary training, single and double-honors and advanced programs. To illustrate, Clemente et al. (2020) tested academic performance as a predictor of employability in 2,983 undergraduates across five degree courses offering programs for high academic achievers in a University in Spain. A similar study by Felicetti (2017) was conducted, testing a scholarship program for academically talented and high-achieving students. Overall, the studies are heterogeneous in terms of level of governance and applied interventions but similar in the supposition that advanced programs (i.e. programs that exceed standard education) improve GE.
The most frequent interventions tested were inclusions of practical experience in the HE curriculum (n = 17), including internships, field work or types of work integrated learning (WIL) of various durations (e.g. Vamsi Krishna and Babu, 2021; Thune and Støren, 2015). The inclusion of practical experiences in the curriculum is generally acknowledged to enhance GE by providing students with work experience during their studies and the opportunity to apply theoretical knowledge and practical habits in real-life settings. Bates et al. (2018) applied WIL in the form of an industry placement during study to develop self-efficacy and career management skills in students. Garwe (2020) tested whether WIL at different phases of the study program had an impact on labor market outcomes of graduates. The interventions in this core component were rarely imposed via policies.
Four studies examined the effects of international experience on GE. That is study placements (e.g. ERASMUS) of short-term or long-term durations abroad (e.g. DiPietro, 2015). The articles stated no specific contents of the international exposure other than duration but concur that the intervention enhances GE. Liwinski (2019), for example, examined whether international placements enhance the signaling ability of intercultural communication skills in graduates. All studies measured labor market outcomes of students that partook in mostly policy-initiated international placements.
Practical simulations (n = 10) are interventions in the form of job shadowing, problem- or project-based learning (PBL) or simulated WIL and considered an effective substitute for practical experiences. Holding the potential to deliver comparable benefits to GE by enabling students to apply theoretical knowledge within simulated real-life contexts. The interventions were mostly integrated in a course module, often with a group work component, simulating real-life work situations for students to apply their learned skills. Klegeris (2021), for example, applied small group problem solving in a PBL-model, postulating it enhances students' problem solving skills. Wang et al. (2020) implemented a problem-based learning intervention to cultivate self-efficacy and career management skills in students. The studies followed either a case- or a cohort study design. None of the studies were policy-initiated and were mostly applied as an institutional response to an inferred labor market demand (n = 6).
Ten studies tested career planning programs as employability interventions. The interventions mainly consisted of workshops and seminars of various durations or course modules, aimed at enhancing skills tied to job search activities. For example, Williams et al. (2022) held a two-hour workshop to develop career management skills in first year students. The workshops often combined IT-tools into the activity. For example, Teychenne et al. (2019) offered an online career planning module to second year students. Dray et al. (2011) held a preparatory training seminar titled “Enhanced Employability Event” for nurse graduates, aimed at enhancing their job applications skills. None of the identified interventions were policy-prompted.
Personality development programs, exclusively conducted within the classroom, aim to enhance students’ employability, focusing on improving emotional intelligence, critical thinking or interpersonal skills. To illustrate, Yousef's (2021) course for critical thinking skill included cooperative learning, think-pair-share and reading aloud. While all other studies in this group relied on various validated tests, Gibbs and Slevitch (2019) tested skill acquisition based on customer’s evaluation of students in a combination course for emotional- and technical competence training. Personality development interventions were exclusively motivated and implemented by an inferred labor market demand.
Eight studies implemented employability skill programs – i.e. workshops and course modules, varying from one-time meetups up to six-month programs, dedicated toward preparing students’ attractiveness in the labor market. For example, Shahabadkar et al. (2021) applied a preparatory training workshop to develop industry-networking skills in students. Tomasson Goodwin et al. (2019) applied the STAR technique (Situation, Task, Action and Result) – commonly used in work settings – to develop students’ signaling ability to the labor market. The studies in this group were generally vague in reporting intervention content, execution and the exact mechanism for enhancing GE. About 80% of these interventions (n = 6) were initiated through governmental demand.
Table 3 displays the distribution of interventions falling into the nine intervention core components across the four different levels of governance inquired. The statistically significant association between these core components and the level of governance (p < 0.001) suggests that the nature of the intervention is associated with the level of governance that motivated their implementation.
Level of governance per intervention core component, row proportions
| Intervention core component | Policy issued intervention | Institutional response to labor market demand | Governmental demand | Response to inferred labor market demand |
|---|---|---|---|---|
| No. of studies | 7 | 17 | 20 | 28 |
| Linguistic skill development | 20.00% | 60.00% | 20.00% | 0.00% |
| Didactical innovation | 0.00% | 22.22% | 22.22% | 55.56% |
| Advanced programs | 14.29% | 28.57% | 42.86% | 14.29% |
| Practical experience | 11.11% | 27.78% | 27.78% | 33.33% |
| International experience | 75.00% | 0.00% | 25.00% | 0.00% |
| Practical simulation | 0.00% | 11.11% | 11.11% | 77.78% |
| Career planning programs | 0.00% | 40.00% | 40.00% | 20.00% |
| Personality development programs | 0.00% | 0.00% | 0.00% | 100.00% |
| Employability skill programs | 0.00% | 12.50% | 75.00% | 12.50% |
Note(s): Values represent row proportions, indicating, how many of the included studies were implemented through each level of governance. Proportions do not sum to 100% across intervention core component groups, but instead reflect distributions within each component
Intervention effectiveness by core component
To assess which results implemented interventions report (Research Question 4), the following section discusses the intervention core components and their respective reported results (see Table 4 for a full overview) divided into six subgroups: positive-, inconclusive- and negative intervention outcomes (i.e. effects on skill or behavioral development within higher education) and positive-, inconclusive- and negative labor market outcomes (i.e. placement uptake post-graduation). The numbers of studies and results in Table 4 do not represent the total number of publications analyzed, as papers reported on multiple interventions and diverse results thereof. Fisher’s exact test concluded with a value of p < 0.001, indicating a significant association between the intervention core components and their reported results. Following, the results will be discussed in the order presented in Table 4.
Intervention core component-results matrix
| In-Classroom | Post-Graduation | ||||||
|---|---|---|---|---|---|---|---|
| Intervention core component | No. of inter-ventions | Positive intervention outcomes | Inconclusive intervention outcomes | Negative intervention outcomes | Positive labor market outcomes | Inconclusive labor market outcomes | Negative labor market outcomes |
| Linguistic skill development | 5 | 40.00% | 0.00% | 0.00% | 40.00% | 20.00% | 0.00% |
| Didactic innovation | 9 | 100.00% | 0.00% | 0.00% | 0.00% | 0.00% | 0.00% |
| Advanced programs | 7 | 44.44% | 0.00% | 0.00% | 22.22% | 22.22% | 11.11% |
| Practical experience | 17 | 52.94% | 5.88% | 0.00% | 41.18% | 0.00% | 0.00% |
| International experience | 4 | 0.00% | 0.00% | 0.00% | 75.00% | 25.00% | 0.00% |
| Practical simulation | 10 | 50.00% | 10.00% | 10.00% | 30.00% | 0.00% | 0.00% |
| Career planning programs | 10 | 27.27% | 27.27% | 9.09% | 36.36% | 0.00% | 0.00% |
| Personality development programs | 7 | 100.00% | 0.00% | 0.00% | 0.00% | 0.00% | 0.00% |
| Employability skill programs | 8 | 44.44% | 22.22% | 11.11% | 22.22% | 0.00% | 0.00% |
| Total | 77 | ||||||
Note(s): Values represent row proportions, indicating, how many of the included studies reported each type of result. Proportions do not sum to 100% across intervention core component groups, but instead reflect distributions within each component
Linguistic skill development interventions report largely positive results (40%, respectively, for intervention – as well as labor market outcomes (e.g. Dash et al., 2020; Murphy and Cunningham, 2012). In contrast, Hiew et al. (2021) reported inconclusive labor market outcomes post-administration of an advanced English language course intervention, finding that the employment rate among graduates was lower than the unemployment rate post-intervention.
Didactic innovation interventions (n = 9) exclusively reported positive intervention results measured in-classroom. Therefore, whether or not the positive results of didactic innovation interventions transfer to the labor market remains yet unknown.
Studies that applied a form of advanced program intervention showed ambiguous results, with partial positive outcomes. Nghia et al. (2019) report potential for the development of human capital through advanced program interventions, but could not verify this notion with equally increased labor market success of the student cohort. Contrastingly, Pigden and Moore (2018) reported a consistent negative gap for joint honors students compared to single honors graduates’ employment rates.
The most frequent interventions tested were inclusions of practical experience, reporting 94% positive study outcomes, of which 41% are substantiated with positive labor market outcomes for participants. A study by Feijoo et al. (2019) shows a 63% positive response of companies toward students after an industrial internship program.
The majority of studies that examined the effects of international experience demonstrate positive labor market outcomes. Potts and Kim (2021) show that long-to-medium duration (i.e. six months and longer) of international experience leads to better labor market performances of graduates than short-term international placements. Liwinski (2019) reports positive effects on labor market entry for graduates, but exclusively for those with at least one highly educated parent.
Practical simulation interventions (n = 10) in the form of problem- or project-based-learning (PBL) largely show positive results, both, in-classroom and in labor market outcomes (e.g. Clarke et al., 2020; Klegeris, 2021). However, Han et al. (2021) conclude from their findings that improvements of individual skills (i.e. enhanced creativity) did not correlate with improved academic performance. Peng et al. (2021) even report a potential negative impact of problem-oriented teaching on employability of students, whereas employability-focused pedagogy appears to have positive effects.
Of the ten studies that tested career planning programs for employability, 36% reported positive labor market outcomes. In-classroom tested interventions, however, show mixed results, 27%, respectively. Von Konsky and Oliver (2012) report no statistically significant difference in self-reported tech skill post-intervention. Williams et al. (2022) report that their Life Design Workshop had no impact on self-efficacy or career optimism.
Exclusively testing outcomes in-classroom, personality development programs yielded consistently positive results.
Of studies that implemented employability skill programs, 66% showed positive results for labor market outcomes and in-classroom intervention outcomes, however, 22% of interventions in-classroom were inconclusive in outcomes. For example, Ito (2014) reports that their employability seminar only improved about half of the skills that were tested for. Secundo et al. (2019) report that social networks of participants were negatively related to self-reported career management skills.
Table 5 illustrates the distribution of results for each identified level of governance in the study. Interventions initiated through policy and those responding directly to labor market demands predominantly yielded positive results. Among the 20 interventions arising from direct governmental demand, outcomes were often inconclusive or negative. Meanwhile, out of the 28 studies responding to inferred labor market demands, most exhibited positive intervention outcomes and otherwise ambiguous results. A Fisher’s exact test resulted in a value of p < 0.001, indicating a significant relationship between the reported results and the level of governance by which the interventions were issued.
Level of governance-results matrix, row proportions
| In-Classroom | Post-Graduation | ||||||
|---|---|---|---|---|---|---|---|
| No. of interventions | Positive intervention outcomes | Inconclusive intervention outcomes | Negative intervention outcomes | Positive labor market outcomes | Inconclusive labor market outcomes | Negative labor market outcomes | |
| Policy issued intervention | 7 | 42.86% | 0.00% | 0.00% | 42.86% | 14.29% | 0.00% |
| Institutional response to labor market demand | 17 | 52.94% | 0.00% | 0.00% | 47.06% | 0.00% | 0.00% |
| Governmental demand | 20 | 30.00% | 25.00% | 10.00% | 35.00% | 10.00% | 5.00% |
| Response to inferred labor market demand | 28 | 79.31% | 6.90% | 3.45% | 6.90% | 3.45% | 0.00% |
Note(s): The values represent row proportions, illustrating the distribution of results for each identified level of governance in the study
Discussion
This systematic review aimed to assess the approaches of HE institutions to operationalize and enhance GE in its elements: the underlying theories of change guiding these interventions, the levels of governance motivating their implementation and the results they report. Consequently, the review also serves as a comprehensive inventory of HE interventions designed to enhance graduate employability.
The analysis shows that interventions providing practical experiences to students seem most beneficial for GE enhancement, given the positive labor market- and intervention outcomes reported. These results indicate HE institutions’ understanding of the usefulness of labor market experience before graduation. This notion has become evident in leading literature, stating favorable employment outcomes as well as benefits for educational purposes when incorporating practical elements into the study course (e.g. Helyer and Lee, 2014; Jackson, 2024; Jackson and Bridgstock, 2021; Passaretta and Triventi, 2015).
Didactic innovations and personality development interventions scored exclusively positive results; however, in the present sample, they were not tested based on labor market outcomes. This observation might be attributable to a common perception that classroom and personality training may not be directly correlated with labor market outcomes. Conversely, interventions involving practical experience or international placements – are understood to have a more direct link to labor market outcomes for graduates. A previous study by Mason et al. (2009) concurs this notion by identifying personal professional ability as a contributing factor to GE but finding no evidence for classroom practices to affect employability of graduates. Our present data does not display results that confirm this. Instead, it highlights a gap in mutual understanding of the potential effects of classroom interventions on GE between HE institutions and labor market players. This lack of understanding regarding the potential distal benefits of didactic interventions may also stem from challenges in accessing or generating the necessary data for testing them on the basis of labor market outcomes.
Employability skill programs and career planning programs yielded mixed results in their endeavors to enhance GE through various approaches. Unlike practical experiences, interventions in these categories lack a clear link to favorable employment outcomes. The informative value of the identified studies may be hindered by variations in study design quality and clarity of scope and aims.
Advanced programs, career planning programs, employability skill programs and practical simulations reported negative results, frequently also reporting inconclusive results. The negative results identified in this review could point to interventions being subjected to negatively contributing factors outside the scope of course designs, which cannot be foreseen or factored into the operation. The findings do not suggest that these interventions generally display negative results, but that surrounding circumstances are substantial to the success of the intervention.
It is a noteworthy observation that the number of studies observing labor market outcomes of their employability objectives is low – out of 72 included articles, only 27 used labor market metrics – i.e. placement uptake, employment or salary metrics. This factor can be attributed to the inherent challenges associated with procuring access to the study cohort once they graduate (e.g. Peng et al., 2021; Teychenne et al., 2019). The matter further points to a communication gap between HE institutions and the labor market, potentially due to their distinct and separate functions.
Considering the level of governance, the ambiguous results of interventions stemming from governmental demand can be linked to a potential lack of understanding by both the requesting government and HE institutions in charge of the operationalization about the specific elements required for effective GE interventions. In contrast, the favorable outcomes observed in interventions responding to direct labor market demands indicate successful communication between higher education institutions and the employing institutions, emphasizing that the GE topic should not be addressed in isolation but in collaboration with all stakeholders (i.e. HE institutions, employers and policymakers). This also presupposes the acknowledgment and understanding of diverging interest of stakeholders. Future policies should facilitate structured communication between these actors, rather than expecting HE institutions to independently address GE challenges. Policies should be: (1) directive (clearly defined objectives), (2) locally tailored (aligned with regional labor market needs) and (3) inclusive (accounting for multiple perspectives). Effective policy design also requires a shared understanding of GE, including its definition, scope and practical limitations within HE. Policies must balance the competing interests of stakeholders to ensure feasible and sustainable GE strategies (e.g. Kinash et al., 2016).
Our analysis revealed that the operationalization of HE interventions for GE exhibit considerable variability and lack uniformity, producing diverse and heterogeneous results across different metrics. This variability confirms that HE institutions independently explore and implement tailored solutions for integrating employability goals into their curricula. It also underscores the broader challenge of achieving and measuring improvements in GE, which remains largely at the discretion of individual institutions. The degree to which not only individual interventions are universally effective remains obscured, in part by the variety of design and reporting style of publications but also by the large heterogeneity of conceptualization of the GE topic. This is displayed in the large variety of interventions applied and their intended outcomes (i.e. the theory of change), but also in the limited insight of the conditions which brought about the need for an intervention (i.e. the level of governance). This notion also has been noted in similar studies reviewing entrepreneurship education (e.g. Lorz et al., 2013; Pittaway and Cope, 2007).
Study strength, limitations and implications
The present study excluded papers on perceived employability enhancement, which significantly reduced the sample size (n = 146). Perceived employability correlates with actual employment rates in master's students only when they are sufficiently informed about their labor market realities (Caricati et al., 2016). From our analysis, the student’s knowledge of their potential labor market outcomes was not identifiable. However, these papers hold potential informative value; therefore, we plan to include studies on perceived employability interventions in future research.
The initial sample included n = 72 papers; however, the classification process significantly reduced the sample size within intervention core components, limiting the generalizability of the findings. A future study should expand the inclusion criteria by incorporating identified intervention core components as keywords, potentially increasing the sample size and substantiate the findings.
We conclude that establishing a standardized research framework, grounded in the theory of change, for testing GE interventions would significantly enhance the quality, replicability and comparability of educational studies. Wherever the intended outcome and outcome measures do not correspond directly, stated results become less transparent (e.g. Liu et al. (2015) measuring placement uptake for enhanced emotional competence and practical abilities). Interventions that were tested via pre- and post-test, including standardized testing and control groups, provide higher clarity and information and are therefore favorable (see Rossi et al., 2018).
This review found a variety of methods for operationalizing HE, highlighting the absence of a unified approach across HE institutions. These variations are driven by varying local labor market needs, non-directive policies and a unilateral focus on HE in GE implementation. While some consensus exists on the benefits of practical experience during study – likely by fostering communication across stakeholders – variations in the definition of GE will remain. HE institutions are recommended to shift the focus on school-to-work transitions outcomes of HE interventions, as it offers a more tangible and comparable framework for evaluating employability outcomes as opposed to the elusive GE concept.
We gratefully acknowledge the Teaching Academy Groningen for their valuable support.
The supplementary material for this article can be found online.
