Introduction
The higher education system is considered to be one of the key factors for the competitive position of an economy1, where higher education is increasingly expected to engage with the challenges of the contemporary world2,3. To ensure the efficiency of economy, quality education is one of the conditions that must be met because through improving tertiary education, it is possible to move the economy to a higher level of development4. Therefore, higher education institutions should be considered as an element that not only stimulates and encourages the connection between learning, research and innovation, but also transfers knowledge to the economy5.
Higher education is considered an important factor for improvement of youth employability, with many studies showing that young people who have a university degree tend to face lower unemployment rates compared to those without a tertiary level of education6. However, the increased participation in higher education in recent years was faced with a challenging phenomenon, namely the mismatch between the skills of graduates and the needs of the labor market6,7. According to this, a university degree is not sufficiently a criterion for employment, but what matters instead are the practical skills and experience of graduates, transferable skills and the compatibility of the graduate's professional profile with market requirements.
In this situation, higher education institutions receive criticism for not preparing graduates for the real context involved in professional practices8. Therefore, employability turns into the main issue that drives the mission of higher education institutions9. The incorporation of employability as a key component in the mission of higher education prompts changes in teaching, learning and assessment methods, and in particular causing the adaptation of curriculum design and assessment methods in line with labor market demands10.
When assessing quality of higher education, perspectives of graduates and other stakeholders should be taken in consideration, given the critical importance of the learning processes and competences of graduates11. The understanding and study of employability varies across disciplines, representing one of the missing links between the professional context and higher education institutions12. Higher education institutions should include employability competencies in the learning process to prepare students for the labor market and develop competencies that shape their readiness for work13, in particular since traditional theory based teaching in higher education has led to a mismatch between the skills of graduates and the labor market needs.
The fact that a gap has emerged between competencies in higher education and labor market expectations has created a need for further research that studies this phenomenon14. According to studies, the skills gap is the lack of cognitive, technical, and interpersonal skills that are required by the labor market, but that are not sufficiently developed by higher education institutions. It is assessed that study programs are disconnected from labor market requirements, with universities promoting theoretical knowledge, at a time when the labor market values practical competencies and adaptation to a dynamic, constantly changing work environment15.
Even though environmental factors cannot be controlled by higher education institutions, students' competencies can be improved through teaching and learning experiences provided by universities. In relation to this, higher education institutions can take initiatives that would strengthen personal capital and career development to improve their employability16. This would reduce the gap between the preparation of graduates and the employers’ unrealistic expectations for the graduates to have applicable skills in different contexts immediately upon entering the labor market17. The implications of this gap in employability skills go beyond academic considerations with a direct impact on the employment opportunities offered to graduates, but also improving their professional longevity outlooks. The importance of addressing this phenomenon can reduce unemployment, increase productivity without stagnating the economy, and can contribute to the regional and national context. In this regard, understanding the complexity of the mismatch between the supply of higher education and the demand of the labor market can influence the proposal of viable solutions to reduce this gap18.
The Kosovo context is a notable example of the challenges and incompatibilities between higher education and the labor market. At the outset of the third decade of this century, Kosovo saw a significant increase in participation in higher education, culminating at around 73% of young people aged 18-22 enrolled in higher education - one of the highest rates in Europe19. As a result, thousands of graduates tend to join the labor market annually. Another key tendency in Kosovo higher education in recent decades has been the rapid increase in the number of students enrolled in private higher education institutions (43 % of the total number of students in 202320), which came as result of increased popular demand for higher education qualifications and a limited provision by public institutions. However, despite the growing percentages of persons with university qualifications, Kosovo continues to struggle with high unemployment rates (19.5% and low employment rates 38.6%21) and a mismatch between higher education provision and employment.
As a result, this research aims to measure and analyze the impact of higher education on the employability of young people and to identify the factors that influence the mismatch between the educational provision and the demands of the labor market. A significant perspective into the impact of higher education on the labor market is obtained by measuring and analyzing the influence of the type of ownership of the HEIs as an independent variable. Subsequently, and based on the findings, the research aims to inform social partners (universities, industry, and authorities), about the ways to fill the gap between the supply of higher education and the needs of the labor market at home and in Europe as the main channels of employment of Kosovo graduates. The purpose is to come up with well-grounded recommendations and sustainable mechanisms, which would guide the relations between stakeholders and the design and provision of study programs.
Literature review
Employability has been a hot topic in higher education policy and research in recent years, with the emphasis on graduates having the skills to contribute immediately and productively to the economy. This would effectively reposition higher education as a means to an end, namely decent employment, rather than as an end in itself, namely the development of intellectually well-rounded individuals22. Higher education has recently been considered of little use in predicting the specific skills that are expected to be required for a job in the future, as a result, it is mandatory to stimulate the competencies of students in order to take advantage of opportunities for coping and creating self-employment opportunities after completing their studies23. This means that education cannot directly prepare graduates for the specific demands that the labor market may have, since labor market demands are commonly short-term and do not consider the general skills needed to cope with long-term developments and changes. Therefore, it may be more opportune if the orientation of higher education institutions focused on increasing employability of graduates by preparing them to quickly learn specific workplace competencies24.
The way in which the mismatch is interpreted depends on the understanding of theoretical knowledge and learning and the expectations for the function that higher education should have in relation to the labor market. The fact that higher education is heterogeneous and has different functions in relation to the labor market, where different jobs require different competencies according to sectors, the nature of the programs affects the applicability of competencies in the real labor market25.
Many studies have addressed what students should learn and what they actually learn in relation to the demands of the job market26. Depending on the sectors and diversity of enterprises, workplace requirements also vary, as there are no uniform competencies required by all jobs, with the exception perhaps of general skills that are constantly discussed as required for the labor market27. The presence of technological innovation and organizational diversity creates a highly dynamic demand in the labor market, where the main challenge for graduates is to be up-to-date and ready to have a greater diversity of competencies than they had before28.
A study dealing with graduate employability posits that labor market demands are related to observable and unobservable skills, defining these in an inclusive group as general skills that are transferable and useful for almost any job, with the exception of specific requirements that a job may have29. So, employability is considered to be related to the study programs, development of soft skills, and the necessary personal development in order to work within a team23.
Despite the challenges that higher education institutions face, their goal is to prepare graduates with professional and general skills, however, discrepancies and gaps emerge between what is offered and what is required. For example, in a study by Nilsson, graduates in fields such as engineering and economics tended to have a wider range of positions in the labor market immediately after graduation, while graduates in professional programs such as law and medicine found the labor market to be more limited, with homogeneity of horizontal and vertical distribution for graduates in these fields identified25. As a result, higher education institutions can take various measures to improve employability despite the inconsistencies and problems of horizontal and vertical distribution, such as curriculum modifications, providing various trainings in terms of personal attributes, problem-solving skills, interpersonal skills and communication skills, job placement through internship opportunities, and providing learning from the experiences of successful former students, such as Alumni activities30. For this to be possible, faculty must be trained in these skills so that they are then able to steer students through ways to improve their entrepreneurial skills required by the business and workplace environment.
Kosovo context
In Kosovo, there is a vivid gap between the demand of the labor market and higher education provision, a fact that has been negatively affecting the employability of young people. This discrepancy is assessed as a failure of educational structures to function in accordance with the socio-economic needs of the country, with higher education institutions focusing mainly on theoretical content void of practical components that provide students with skills to face challenges of the labor market. Curricula in the large part of the programs offered by higher education do not reflect a technological, economic and social perspective. In addition, this gap is further deepened, due to the lack of institutionalized mechanisms of cooperation between higher education and the labor market, such as advisory/industrial boards, and as a result, graduates do not possess applicable and transferable skills in real work contexts31. This mismatch between labor market demand and higher education is a proven phenomenon where despite the growth in the number of higher education institutions and the increase in the number of graduates, the labor market is failing to absorb them, with a fairly high proportion of graduates remaining unemployed or even employed in fields unrelated to their qualifications, proving a functional gap between the needs of the labor market and the education system32.
The analysis of the trend in the number of graduates and their employability presented in Figure 1 shows that in the period 2012/2013 to 2023/2024, the number of active students increased continuously until 2016/2017, culminating at 123 243 students33, after which a sharp decline began, dropping to 71 11734 in the 2023/2024 academic year. This change indicates a shift from massive expansion to a narrowing of the student base, related to demographic factors, emigration, saturation of the demand for higher education, and a lack of motivation for higher education due to high levels of graduate unemployment. The number of graduates reached an early peak in 2014/2015 with 21 364 graduates but then fell significantly and remained stable at a lower average. At the same time, unemployment among higher education graduates increased sharply, peaking in 2017/2018 with 31 436 unemployed, more than the number of graduates that year, which signals a build-up of unemployment from previous years and an increased mismatch with the labor market. After 2021/2022, unemployment gradually decreased, related to the reduction in educational supply, improved market adaptation, or even graduate emigration. The latter could be linked to the EU visa liberalization for Kosovo in 2023 and introduction of working visas for Kosovo labor force to Western Countries, as a result of which Kosovo has been losing between 25 000 and 40 000 inhabitants annually for the last five years35. Overall, the data reflects a cycle of transition from quantitative expansion to an effort for structural adjustment and higher effectiveness of higher education in relation to market demands.
Figure 1. Number of graduated students and unemployment of graduates on HE.
Source: Kosovo Agency of Statistics, 2025.
The World Bank and UNDP reports emphasize that higher education in Kosovo is not making a sufficient contribution to improving employability. This is related to the lack of cooperation and student internships with industry to develop transversal competencies, as a result of which graduates often do informal work, or are looking for opportunities to migrate for better opportunities36,37.
In the same context, the European Training Foundation (ETF) emphasizes that policy and operational measures should be taken to create a more integrated system of education and employment to enable the harmonization of academic content with labor market requirements, defining the function of higher education as a catalyst for economic development38.
Methods
To accomplish the research objectives, a quantitative method was used, with the population consisting of public and private university graduates of the last 5 years (2019/20-2023/24) as shown on Table 1 (data obtained from the Kosovo Agency of Statistics). The sample for this study consists of 300 graduates, selected through simple random selection, in accordance with the real distribution of graduates in Kosovo during the years 2019/2020 - 2023/2024. Out of a total of 64,784 graduates, 42% belonged to private institutions and 58% to public ones. Therefore, the sample was also divided proportionally with 127 from the private sector and 173 from the public sector, ensuring fair representation for statistical analysis.
Table 1. Graduate students in the last five years (KAS).
| 2019/20 | 2020/21 | 2021/22 | 2022/23 | 2023/24 | Total | Percentage | Sample |
|---|
| Private | 2 270 | 6 094 | 5 257 | 7 665 | 6 070 | 27 356 | 42% | 127 |
| Public | 6 216 | 7 205 | 7 892 | 7 272 | 8 843 | 37 428 | 58% | 173 |
| | | | Total | 64 784 | Total | 300 |
Informed consent procedures
The survey was conducted online using Microsoft Forms and was fully anonymous. Before accessing the questionnaire, participants were presented with an information page describing the purpose of the study, the voluntary nature of participation, data confidentiality, and their right to withdraw at any time. Proceeding to the first question required selecting “Yes, I agree to participate”, which constituted electronic written informed consent. No minors were involved in the study, and no personal or identifying data were collected.
The study instrument was structured and divided into three sections. The first section contained demographic questions and Graduate Profile Characteristics, the second section included five measurement dimensions for higher education (study programs (SPH), teaching quality (TQ), university-industry collaboration (CUB), flexibility of academic programs (FAP), infrastructure and resources (HEIR)) constructed according to the Likert scale and part of the third section was the employability of graduates treated with the same scale. Figure 2 presents the research methodology.
Figure 2. Flowchart of research methodology.
This paper has presented two objectives and three hypotheses:
Objective 1: To measure the impact of higher education on the employability level of graduates;
H1: Higher education has a statistically significant impact on the employability of graduates;
Objective 2: To assess whether there is a statistically significant difference in the employability of graduates depending on the ownership of the institution (private and public);
H2: There is a statistically significant difference in the level of employability of graduates depending on the type of higher education institution (public or private);
Objective 3: To analyze the level of employability of graduates depending on graduate profile characteristics;
H3: There is a statistically significant difference in the level of employability of graduates depending on the graduate profile characteristics;
Objective 4: To measure the impact of external drivers of employability on the employability of graduates;
H4: Drivers of employability external to HE has a statistically significant impact on the employability of graduates;
To see the relationship between the variables, refer to the framework below Figure 3:
Figure 3. Framework of research.
To determine the distribution of data, the Kolmogorov Smirnov test and the Shapiro Wilk test were used. According to the test results, the significance value is p>0.05, which means that we have a normal distribution of data and the condition for using the Pearson correlation and parametric tests for comparing means is met. Also, the internal consistency of the measuring instrument was measured through Cronbach's alpha coefficient, where the value α=.875 argues that the questionnaire has acceptable reliability. This test is done to assess whether the questions of a measurement scale function as a connected group and accurately measure a single construct. The calculation of this coefficient was done through the following equation:
Also, exploratory factor analysis (EFA), a statistical technique of multiple variables, is used to reduce the number of variables that are found in relation to each other to a smaller number of significant variables and independent of each other. Factor analysis includes different techniques, but at the same time they are related to each other. The first test applied is Kaiser Meyer Olkin – KMO and Bartlett's Test of Sphericity. KMO argues the suitability of the selected sample for factor analysis, where the KMO level should be above 0.50, the higher the coefficient, the better the data set is to do factor analysis. While Bartlett's Test of Sphericity shows whether the data set is suitable and if there is a proper theoretical basis for factor analysis. Such measurements are made through the following equations:
Through communality, the contributing variables in the construction of the latent structure are assessed to identify which ones represent weak loadings for interpretation. Thus, it measures the degree to which a variable is represented by the common factors of the factorial model. The following equation represents the mathematical calculation of communality.
Through correlation analysis, the relationship between the measured variables of the higher education dimensions (SPH, CUB, TQ, FAP, HEIR, LMD, RTI) and youth employability (EY) was tested. The Pearson correlation coefficient measures the direction and strength of the linear relationship between the variables treated in the paper, where depending on the value of r, we determine the level of the relationship. The Pearson correlation was calculated through the following equation:
The F statistics are used in the regression model to measure whether the variation between groups is statistically greater than within them. So, through the equation below, it is tested whether the regression model is valid as well as to measure whether the groups in the ANOVA differ for the dependent variable:
To measure what percentage of the variance of the dependent variable is explained by the independent variable, R Square was used, while avoiding overfitting the model when many variables are added, Adjusted R Square was used. Calculations were made based on the following equations:
The important test is the Durbin Watson test, through which we tested whether autocorrelation exists in our model. The test values should be between 1.5 and 2.5 to assess that the model does not have autocorrelation problems. This measurement is made through the following equation:
The equation below represents the linear regression, which was used to measure the impact of higher education on employability. Beta, as a standardized coefficient, was used to measure the relative impact of each independent variable. So, through B, we measured which factors of higher education have the greatest impact on the employability of young people:
The t-test is used to investigate the difference between two sample groups in terms of means. The test determines whether there is a significant difference between the mean of one group and the mean of the other group. In this case, through the equation below, we tested whether students' perceptions regarding employability and the measurement dimensions of higher education differ between private and public higher education institutions in Kosovo:
Analysis of variance is used to test the hypothesis about whether there is a significant difference between two or more means. One Way ANOVA is a simple analysis of variance and consists of two variables. One of these variables has categorical characteristics (independent variable) and the other variable has metric characteristics (dependent variable). In analysis of variance, the F value is used to test the hypothesis through the following equation:
When a significant difference is found between groups in the analysis of variance, the Post Hoc test is important to see which group the difference originates from. The test used to identify where the difference originates within the Post Hoc tests is the Tukey HSD test. Through this test, it was identified which areas have higher employability, by grouping them into 3 subsets. This measurement was made through the following equation:
Results and discussion
The demographic data in Table 2 represents a clear distribution of educational attainment by age group, reflecting the actual profile of graduates in Kosovo over the past five years. The sample is dominated by individuals aged 24–32 (over 70%), where a shift from Bachelor to master’s degrees is evident, aligned with increasing labor-market demands for advanced qualifications. The 21–23 age group consists mainly of Bachelor graduates, while those aged 27–32 are almost entirely Master’s graduates, indicating professional consolidation and academic maturity. Doctoral degrees appear only after age 33, with the highest concentration in the “36 and above” group, consistent with the timeline required to complete third-cycle studies.
Table 2. Demographic data.
| Age Group (Years) | Bachelor’s Degree | Master’s Degree | Doctorate (PhD) | Total | % |
|---|
| 21–23 | 66 | 0 | 0 | 66 | 22.0 |
| 24–26 | 36 | 67 | 0 | 103 | 34.3 |
| 27–29 | 6 | 45 | 0 | 51 | 17.0 |
| 30–32 | 0 | 56 | 0 | 56 | 18.7 |
| 33–35 | 0 | 6 | 3 | 9 | 3.0 |
| 36 and above | 0 | 6 | 9 | 15 | 5.0 |
| Total | 108 | 180 | 12 | 300 | 100.0 |
Figure 4 shows perceptions of graduates from HEIs in Kosovo regarding the dimensions influencing the relevance of higher education to the needs of the labor market. The obtained data show differing results, with of the dimensions receiving positive responses from graduates, whereas some others reflect to the gaps and systematic challenges, which affect the functioning of higher education in Kosovo and determine the employability of young people. Even though study programs are assessed as partly aligned with the labor market, there is still no coherence between the objectives of higher education and the labor market. As a result, higher education graduates appear more like passive consumers of theoretical knowledge than as agents ready to face the challenges of the labor market. Cooperation between universities and businesses indicates to fundamental issues, which showcase systemic weaknesses of higher education. The dimension of cooperation turns out to be one of the weakest links in the chain, reflecting a functional lack of connection with industry and the institutions, that ultimately limit students’ opportunities to transform theoretical knowledge into practice, and thus further increasing the gap between theoretical knowledge and applicable skills in the reality of the labor market.
Figure 4. Degree of compliance for variables.
Source: Authors’ calculations based on survey data39.
Another worrying result concerns the role of technology and innovation as the most critical indicator in terms of overall compatibility, with data showing a significant lack of integration of technology and innovation in the content of study programs, turning into a widening gap between the content of study programs and the exponentially changing features and needs of the labor market. Regarding employability, here too, a fragmentation appears, which shows that (a) some graduates manage to enter the labor market through the education received from higher education, (b) another part has had to receive additional professional training to enter/stay in the labor market, while (c) the third are employed in jobs as overqualified or outside their field of study. This result casts doubt on the connection between higher education policies and labor market demands, as well as on the relevance of the university diploma.
The research results pointed out overwhelmingly to graduates’ long periods of unemployment upon graduation (46%), as well as to the lack of relevant skills obtained from university studies and the need to go through additional training as a precondition for employment (58%). Such results speak of serious issues regarding the interface between HE qualifications and the needs in the labor market. The conservative nature (in particular of the public) higher education institutions and the resulting slow response to rapidly changing trends in the new technologies and in the labor market, present a call for alarm for education authorities to review the higher education responsiveness and provision in Kosovo. It should be added that most of the graduates were initially engaged in a job other than their field of study (49%), requiring lower qualifications than their degree (41%) and therefore being underpaid when compared to their university qualification (38%). Moreso, despite their formal ‘overqualification’ they had to undergo additional training to enjoy more adequate earnings in their job place.
These indications call for universities to pay more attention to the needs and trends in the labor market and to design more flexible and responsive study programs for their students and future graduates. This specifically points to the need for engaging in closer dialogue and partnerships with the industry and providing shorter, more flexible, and most notably, tailor-made training and study programs/qualifications for their students and industry partners. The results show a dual tendency, whereby some HE dimensions tend to align (to a lesser or greater extent) with labor market demands, whereas, important dimensions such as university cooperation with industry and the inclusion of technology and innovation in study programmes present negative results. This duality makes it necessary to reform higher education policies and approaches in order to move towards a coherent, innovative and practical model that will influence the increase in youth employability.
Figure 5 presents the comparison between private and public higher education institutions regarding graduates’ perceptions on the dimensions influencing the relevance of higher education related to the needs of the labor market. In the first dimension, that of higher education programs (SPH), private institutions perform better than public ones. Graduates from private institutions had a higher tendency of compliance with the study programs, assessing them as more up to date, oriented towards labor market requirements, and supporting the development of practical skills. On the other hand, graduates from public institutions had neutral or negative attitudes, highlighting the mismatch of the curriculum with labor market requirements and the lack of inclusion of the practical work component in academic curricula. The second dimension represents the cooperation of higher education institutions with industry (CUB), where the difference between private and public institutions is even more striking than in programs dimension (SPH). Private institutions were assessed to be more involved in establishing collaborations with industry to enable transfer of theoretical skills into practice. In public institutions, a significant lack of cooperation with industry and a low level of international cooperation for training opportunities and capacity development were identified, further widening the mismatch of higher education with market demands. Teaching quality (TQ) is another dimension addressed within the framework of higher education, a dimension in which private education institutions yet again stood in a more favorable position than the public ones. Graduates of public institutions stated that teaching methods are oriented more in the theoretical aspect and less in solving concrete problems and lacked the professional orientation of students. In this context, students of private institutions stated that teaching methods are contemporary, with teachers engaged in research regarding labor market demands, and bringing the current labor market into the teaching process. The flexibility of academic programs (FAP) is another dimension perceived more positively by graduates of private institutions through providing the opportunity to choose appropriate modules by adapting their skills to market demands, recognizing practical experience as part of academic credits and including courses on new technologies. Meanwhile, public institutions, according to the graduates’ perceptions, had a more negative approach, assessing that public institutions have a bureaucratic structure, are not aligned with the demands of the labor market, and show significant lack of flexibility to adapt to individual student needs. If we refer to the infrastructure and resources of higher education (HEIR), we find that public institutions have a serious problem in this context, as a result of the lack of investment in laboratories and a modern environment for practical learning, lack of access to advanced resources such as electronic libraries, scientific databases, professional software and lack of orientation to offer courses or certifications to help students be more competitive in the market. In this line, private institutions resulted in infrastructure investments oriented towards quality culture.
Figure 5. Degree of compliance for variables - comparison between private and public institutions.
Source: Authors’ calculations based on survey data39.
These findings posit that private institutions are more oriented towards preparing students in line with the demands of the labor market than public ones. This difference reveals the need for substantial reform of public institutions in terms of updating curricula, infrastructure investments, engaging in industry cooperation, increasing the flexibility of academic programs and the quality of teaching, aiming to eliminate the gap between higher education and labor market demands, and avoid consequences on employability.
Figure 5 shows also the comparability between private and public institutions in terms of employability (EY), matching labor market requirements (LMD) and the role of technology and innovation (RTI). It resulted that graduates from private institutions evaluate the education received as more useful in employment, while in public institutions, preparation for the labor market seems more uneven. Even in labor market requirements, the perception is more positive among graduates of private institutions than public ones, where public institutions show a lack of alignment of academic curricula with the dynamic nature of the labor market, mainly in the practical skills required by employers. Interesting results for both types of institutions were obtained regarding the role of technology and innovation in higher education, viewing them as ignored or marginalized factors in academic curricula, further proving that preparing students for market needs in terms of technology is a challenge throughout the higher education system.
Factor analysis was used as a statistical technique of multiple variables to reduce the number of variables that are found in relation to each other to a smaller number of significant variables and independent of each other. For the assessment of the adequacy of the data, the statistical indicators Kaiser-Meyer-Olkin (KMO) and Bartlett's Test of Sphericity come into consideration. Bartlett's Test of Sphericity tests the possibility of testing high correlations between at least a part of the variables in the correlation matrix, while KMO is an index that compares the size of the observed correlation coefficient with the size of the partial coefficient. The statistics presented in Table 3, with the KMO index of 0.889, indicate sample adequacy, which means that the relationships between the variables are sufficient to obtain reliable factors in the analysis. Also, based on the values (Chi-Square = 1308.312; df = 28; p < .001) of Bartlett's Test, we confirm that there is a significant correlation between the variables and as a result the data is suitable for exploratory factor analysis (EFA).
Table 3. KMO and Bartlett's Test.
| KMO and Bartlett's Test |
|---|
Kaiser-Meyer-Olkin Measure of Sampling
Adequacy. | 0.889 |
| Bartlett's Test of Sphericity | Approx. Chi-Square | 1308.312 |
| Df | 28 |
| Sig. | 0.000 |
Communalities statistics are an indicator of the percentage of variance of each variable that is explained by the factorial structure as is showed in Table 4. Since all values are greater than 0.5, we say that the suitability of the variables in this factorial model is evident and one of the key criteria for model acceptance is met. Referring to the Rotated Component Matrix, the factorial structure is statistically acceptable and provides support from high Communalities and clarity of factor loadings. The variables SPH, CUB, TQ, FAP, HEIR, EY, LMD, have strong loadings and are representatives of a latent factor that argues that the model is adequate. Also, according to the Varimax rotation, no correlation has been identified between the factors, which means that they are orthogonal, which means that there is no overlap of information between the factors.
Table 4. Communalities and Rotated Component Matrixa.
| Rotated Component
Matrixa |
|---|
| Communalities | Component |
|---|
| Initial | Extraction | | 1 | 2 |
|---|
| SPH | 1.000 | 0.583 | SPH | 0.712 | 0.277 |
| CUB | 1.000 | 0.724 | CUB | 0.811 | 0.258 |
| TQ | 1.000 | 0.718 | TQ | 0.829 | 0.174 |
| FAP | 1.000 | 0.772 | FAP | 0.874 | 0.089 |
| HEIR | 1.000 | 0.744 | HEIR | 0.861 | 0.046 |
| EY | 1.000 | 0.713 | EY | 0.758 | -0.372 |
| LMD | 1.000 | 0.677 | LMD | 0.714 | -0.408 |
| RTI | 1.000 | 0.793 | RTI | 0.161 | 0.876 |
Extraction Method: Principal
Component Analysis. | Extraction Method: Principal
Component Analysis.
Rotation Method: Varimax
with Kaiser Normalization. |
| a. Rotation converged in 3
iterations. |
To analyze the relationship between the independent and dependent variables, Pearson Correlation was used, where the correlation of SPH with EY (r=.425; p<0.01) is a moderate positive linear relationship, which means that the structure of study programs is influential in preparing for work. As shown in Table 5, according to the values (r=.468; p<0.01), the correlation between CUB and EY is a moderate positive linear relationship, where cooperation between universities and businesses improves the possibility of employment. The correlation between TQ and EY (r=.513; p<0.01) is estimated to be a moderate positive relationship, which means that the quality of teaching significantly affects preparation for the labor market. The correlation (r=.606; p<0.01) between FAP and EY and (r=.625, p<0.01) between LMD and EY is indicative of a strong positive linear relationship between the variables. So, the more flexible the academic programs are, the higher the perception of employability will be, and compliance with labor market requirements is a key factor for employability. According to the values (r=.546; p<0.01) between HEIR and EY there is a moderately strong relationship, where infrastructural conditions support professional preparation. The RTI variable is excluded from the model due to the increased significance value since p>0.01.
Table 5. Correlation matrix.
| SPH | CUB | TQ | FAP | HEIR | LMD | RTI | EY |
|---|
| SPH | 1 | | | | | | | |
| CUB | .576** | 1 | | | | | | |
| TQ | .548** | .688** | 1 | | | | | |
| FAP | .589** | .679** | .711** | 1 | | | | |
| HEIR | .514** | .727** | .705** | .715** | 1 | | | |
| LMD | .403** | .390** | .443** | .539** | .556** | 1 | | |
| RTI | .276** | .248** | .213** | .204** | .147* | -0.075 | 1 | |
| EY | .425** | .468** | .513** | .606** | .546** | .625** | -0.078 | 1 |
According to the results of the simple linear regression, R=.606 shows that there is a strong overall relationship between higher education and employability, while R2=.367 means that 36.7% of the variance in youth employability is described by the HE measurement dimensions, as shown in Table 6. Adjusted R2=.365 proves that the model is stable. Another important indicator is the Std. Error of the Estimate (Std. Err=0.51898) which proves that the predictions are accurate in relation to the degree of data variation. The F value (1, 298) = 172.967, and sigF<0.05 confirms that the model is statistically significant and that the impact of higher education on employability is significant. The Durbin-Watson test (2.073) shows that the model does not have an autocorrelation problem and that the assumption of independence of errors is met.
| Model | R | R Square | Adjusted R
Square | Std. Error of the
Estimate | Change Statistics | Durbin-
Watson |
|---|
R Square
Change | F Change | df1 | df2 | Sig. F
Change |
|---|
| 1 | .606a | 0.367 | 0.365 | 0.51898 | 0.367 | 172.967 | 1 | 298 | 0.000 | 2.073 |
Based on the model below, the value β0= 1.217, describes the expected value of youth employability when the value of HE components is zero, for each increased unit of HE, youth employability will increase by 0.600 units. Std. Error=0.046, t=13.152 and p<0.01 emphasize that the model is statistically highly significant, as it is shown in Table 7. The regression results show that HE is an important and influential factor in youth employability, where the increase in the perception of the quality and relevance of higher education is influential in increasing the opportunity for employment.
Model:
Table 7. Coefficients.
| Model | Unstandardized
Coefficients | Standardized
Coefficients | t | Sig. |
|---|
| B | Std. Error | Beta |
|---|
| 1 | (Constant) | 1.217 | 0.155 | | 7.876 | 0.000 |
| Higher education | 0.600 | 0.046 | 0.606 | 13.152 | 0.000 |
| a. Dependent Variable: Employability of youth |
To measure the individual impact of each HE descriptive factor on employability, seven separate simple regression models were conducted with employability as the dependent variable and independent variables (SPH, CUB, TQ, FAP, HEIR, LMD, RTI) tested separately in regression analysis.
Table 8 presents the simple linear regression, based on R2=0.18 (β=0.422; R=0.425; t=8.099; Sig.<0.05), 18% of youth employability is positively affected by higher education study programs (SPH). The value R2=0.219 (β=0.372; R=0.468; t=9.13; Sig.<0.05) indicates that 21.9% of employability is positively affected by higher education-industry cooperation (CUB). The value R2=0.263 (β=0.466; R=0.513; t=10.304; Sig.<0.05) shows that 26.3% of employability is positively affected by teaching quality (TQ). The value R2=0.367 (β=0.493; R=0.606; t=13.15; Sig.<0.05) shows that 36.7% of employability is positively affected by the flexibility of the study program (FAP). The value R2=0.298 (β=0.404; R=0.546; t=11.259; Sig.<0.05) shows that 29.8% of employability is positively affected by the resources and infrastructure of higher education institutions (HEIR). The value R2=0.39 (β=0.527; R=0.625; t=13.812; Sig.<0.05) shows that 39% of employability is positively affected by labor market demand (LMD). Due to the significance of p=0.18>0.05, the role of technology and innovation is excluded from the model.
The regression data show that the most important and influential determinants in youth employability are labour market demand (LMD), flexibility of academic programs (FAP), Higher Education Infrastructure and Resources (HEIR), and teaching quality (TQ). Technology and innovation (RTI), because they did not show a significance level, proves the findings of Pearson correlation and factor analysis, even though RTI in theory has a significant impact, it is not effectively included in practices affecting employability.
Table 8. Simple linear regression for dimensions.
| Variables | β | R | R Square | Adjusted R
Square | t | Sig. | Std.
error | Durbin
Watson |
|---|
| SPH | 0.422 | 0.425 | 0.180 | 0.178 | 8.099 | 0.00 | 0.053 | 1.679 |
| CUB | 0.372 | 0.468 | 0.219 | 0.216 | 9.130 | 0.00 | 0.041 | 1.888 |
| TQ | 0.466 | 0.513 | 0.263 | 0.26 | 10.304 | 0.00 | 0.045 | 2.013 |
| FAP | 0.493 | 0.606 | 0.367 | 0.365 | 13.150 | 0.00 | 0.037 | 1.99 |
| HEIR | 0.404 | 0.546 | 0.298 | 0.296 | 11.259 | 0.00 | 0.036 | 2.159 |
| LMD | 0.527 | 0.625 | 0.390 | 0.388 | 13.812 | 0.00 | 0.038 | 2.004 |
| RTI | -0.072 | 0.078 | 0.006 | 0.003 | -1.3430 | 0.18 | 0.054 | 1.577 |
Table 9 presents the averages of the variables treated in the paper comparing institutions by ownership. The average SPH for private is x=3.62 and SD=0.55 and public x=3.40 and SD=0.69, showing an average difference of 0.22 in favor of the private sector, a not very large difference but the high consistency in the assessments in the private sector proves that the study programs are evaluated more positively, are more standardized for students. The private sector has problems with fragmentation of study programs, lack of updating, and lack of practical integration. The average CUB for privates is x=3.43 and SD=0.55 and public x=3.39 and SD=0.69, showing an average difference of 0.57 in favor of private institutions wherewith students graduating from private institutions report that the university's cooperation with industry is stronger, suggesting a more practical orientation. According to statistics, public universities were found to be more bureaucratic and without flexible mechanisms for cooperation with industry. The average TQ for privates is x=3.74 and SD=0.47 and public x=3.05 and SD=0.71, with an average difference of 0.69 in favor of private institutions, ranking the quality of teaching higher in private institutions with a lower standard deviation, indicating more stable assessments. The average FAP for private is x=3.78 and SD=0.74 and public x=3.24 and SD=0.77, with an average difference of 0.54 in favor of private institutions, which showed greater adaptability by adapting to market demands, factors that reflect the functional orientation of the curriculum. The average HEIR for private institutions is x=4.07 and SD=0.54 and public x=3.07 and SD=0.82, with an average difference of 1.00, representing the highest difference between institutions from all variables treated. According to this result, public institutions do not have an orientation to invest sustainably in modern infrastructure, reporting a lack of laboratories, functional spaces for learning, digitalization, etc., a situation that reduces the capacity for providing quality. The average EY for private is x=3.58 and SD=0.58 and public x=3.03 and SD=0.61, with an average difference of 0.55 in favor of private institutions. Such perception regarding employability is as an accumulated result of all the factors discussed above, where in each of them private institutions had the highest score. The average LMD for private is x=3.75 and SD=0.61 and public x=3.26 and SD=0.79, with an average difference of 0.49 in favor of private institutions which are perceived to have a higher match to the demands of the labor market.
The evident differences between private and public institutions present structural, strategic and managerial problems of public higher education in Kosovo. One reason that private institutions stand better in each component is because of competition that effectuated their orientation towards matching the market demands and focusing on more functional preparation of students.
Table 9. Descriptive results for private and public institutions.
| Private | Public |
|---|
| N | Mean | Std.
Deviation | | N | Mean | Std.
Deviation |
|---|
| SPH | 130 | 3.6150 | 0.55325 | SPH | 170 | 3.3975 | 0.69147 |
| CUB | 130 | 3.4325 | 0.59761 | CUB | 170 | 2.8613 | 0.84833 |
| TQ | 130 | 3.7350 | 0.47250 | TQ | 170 | 3.0513 | 0.70946 |
| FAP | 130 | 3.7800 | 0.73879 | FAP | 170 | 3.2400 | 0.77038 |
| HEIR | 130 | 4.0700 | 0.54712 | HEIR | 170 | 3.0688 | 0.82438 |
| EY | 130 | 3.5780 | 0.58296 | EY | 170 | 3.0277 | 0.60573 |
| LMD | 130 | 3.7550 | 0.61132 | LMD | 170 | 3.2675 | 0.79434 |
| RTI | 130 | 2.1950 | 0.49439 | RTI | 170 | 2.2500 | 0.77896 |
To determine whether the differences identified between private and public institutions are significant, a t-test was used for all variables included in the model. The results of these analyses are summarized in Table 10. The results of the t-test (t (298) = –2.737, p = 0.007) for SPH show that the difference in the mean is significant, in which case it is confirmed that private institutions are marginally (0.23) better than public ones in terms of study programs. For the CUB variable, the results of the t-test (t (298) = –6.025, p < 0.05) prove that the difference between private and public institutions is significant, with private institutions being more structured and effective in the labor market. Similarly, the results of the t-test for TQ (t (298) = –8.716, p < 0.05) confirm that the difference in the mean of 0.68 in favor of private institutions is significant, which reinforces the presence of teaching standards. The results of the t-test for FAP (t (298) = –5.801, p < 0.05) also confirm the fact that private institutions offer more opportunities for curriculum adaptation as the results were significant. The variable with the largest difference turned out to be HEIR, where the results of the t-test (t (298) = –10.991, p < 0.05) confirm that the difference in the mean of 1.00 is significant, presenting a deep structural inequality between private and public institutions. The dependent variable EY also referred to the t-test (t (298) = –7.511, p < 0.001), identifies a statistically significant difference of 0.55 in favor of the private sector, a result which summarizes the cumulative impact of other factors addressed in the research, making private institutions have a higher affinity to prepare students for the labor market.
Table 10. T-test for differences between private and public institutions in the dimensions of higher education and employability.
| Levene's Test for Equality of Variances | t-test for Equality of Means |
|---|
| F | Sig. | t | df | Sig. (2-tailed) | Mean Difference | Std. Error Difference | 95% Confidence Interval of the Difference |
|---|
| Lower | Upper |
|---|
| SPH | Equal variances assumed | 4.966 | 0.027 | -2.737 | 298 | 0.007 | -0.21750 | 0.07946 | -0.37388 | -0.06112 |
| Equal variances not assumed | | | -2.946 | 240.924 | 0.004 | -0.21750 | 0.07383 | -0.36294 | -0.07206 |
| CBU | Equal variances assumed | 13.411 | 0.000 | -6.025 | 298 | 0.000 | -0.57125 | 0.09481 | -0.75783 | -0.38467 |
| Equal variances not assumed | | | -6.746 | 265.107 | 0.000 | -0.57125 | 0.08467 | -0.73797 | -0.40453 |
| TQ | Equal variances assumed | 14.882 | 0.000 | -8.716 | 298 | 0.000 | -0.68375 | 0.07845 | -0.83813 | -0.52937 |
| Equal variances not assumed | | | -9.922 | 274.479 | 0.000 | -0.68375 | 0.06891 | -0.81942 | -0.54808 |
| FAP | Equal variances assumed | 0.369 | 0.544 | -5.801 | 298 | 0.000 | -0.54000 | 0.09308 | -0.72319 | -0.35681 |
| Equal variances not assumed | | | -5.883 | 205.668 | 0.000 | -0.54000 | 0.09179 | -0.72097 | -0.35903 |
| HEIR | Equal variances assumed | 21.221 | 0.000 | -10.991 | 298 | 0.000 | -1.00125 | 0.09110 | -1.18053 | -0.82197 |
| Equal variances not assumed | | | -12.524 | 275.026 | 0.000 | -1.00125 | 0.07995 | -1.15864 | -0.84386 |
| EY | Equal variances assumed | 0.127 | 0.422 | -7.511 | 298 | 0.000 | -0.55033 | 0.07327 | -0.69453 | -0.40614 |
| Equal variances not assumed | | | -7.608 | 205.013 | 0.000 | -0.55033 | 0.07234 | -0.69296 | -0.40771 |
To measure the difference in the employability of young people (EY) depending on the three variables Level of education (LE), Grade point average (GPA), Field of study (FS), One Way ANOVA was performed. The results of these analyses are presented in Table 11. Given that the results for LE (F(2,297)=1.850; sig=0.159>0.05) and the results for GPA (F(3,296)=1.811; sig=0.145>0.05) are not statistically significant, we conclude that there is insufficient evidence that the employability of young people differs depending on the level of education and that the level of the average grade is not a significant indicator of employability. Unlike the two variables above, field of study (FS) (F (8,291) =10.043; sig=0.000) presents a statistically significant result, which means that the perception of employability varies depending on the field of study. Post Hoc tests were used to identify fields of study that have differences in employability.
Table 11. Results of One-way ANOVA.
| Employability of youth | Sum of
Squares | Df | Mean
Square | F | Sig. |
|---|
| LE | Between Groups | 1.561 | 2 | 0.780 | 1.850 | 0.159 |
| Within Groups | 125.291 | 297 | 0.422 | | |
| GPA | Between Groups | 2.286 | 3 | 0.762 | 1.811 | 0.145 |
| Within Groups | 124.566 | 296 | 0.421 | | |
| FS | Between Groups | 27.445 | 8 | 3.431 | 10.043 | 0.000 |
| Within Groups | 99.407 | 291 | 0.342 | | |
| Total | 126.852 | 299 | | | |
Since the ANOVA results highlighted statistically significant differences regarding the field of study, the detailed results of the Post Hoc Tukey HSD test are presented in Table 12. Through Post Hoc Tukey HSD we identify in detail the fields of study between which there are significant differences. The results of the Tukey HSD test present three homogeneous subsets, classifying them according to the average of employability. In the first subset, the fields of study with the lowest average are emergency management). These fields have a lower affinity for employment opportunities, which reflects the unfavorable position of these fields in relation to market demands. In the second subset, fields such as legal science, economics and business, education and teaching, and social sciences and humanities are grouped, which although they do not have a negative perception do not offer the same level of security in relation to employability compared to technical fields. In subset three, engineering and architecture, and especially computer science and information technology, are grouped in fields that present the highest average and statistically significant difference. These findings represent a higher compatibility through the competencies acquired during the study and the demands of the labor market, mainly the technology and engineering sector.
The findings of this test report that the field of study is a critical factor in terms of employability, where students who are oriented towards technology and engineering are more easily integrated into the labor market after graduation. This fact shows the need for strategic reform of the academic offer to achieve a match between the offer of higher education and the demands of the labor market.
Table 12. Tukey HSD test.
| Field of study | N | Subset for alpha = 0.05 |
|---|
| 1 | 2 | 3 |
|---|
| Tukey HSDa,b
| Agriculture and Environmental Sciences | 27 | 2.6833 | | |
| Arts and design | 23 | 2.8667 | 2.8667 | |
| Other | 12 | 2.9333 | 2.9333 | |
| Legal science | 39 | 3.0327 | 3.0327 | |
| Economics and business | 50 | 3.2152 | 3.2152 | |
| Education and teaching | 29 | 3.2762 | 3.2762 | |
| Social sciences and humanities | 45 | 3.3238 | 3.3238 | |
| Engineering and architecture | 31 | | 3.7704 | 3.7704 |
| Computer science and information technology | 44 | | | 4.3944 |
| Sig. | | 0.004 | 0.001 | 0.0490 |
| Means for groups in homogeneous subsets are displayed. |
| a. Uses Harmonic Mean Sample Size = 6.482. |
| b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed. |
Table 13 presents the hypotheses and sub-hypotheses raised in the research, along with the tests used to validate the hypotheses.
Table 13. Verification of hypothesis.
| Alternative Hypothesis | Statistical test used | Result |
|---|
H1: Higher education has a statistically significant impact on the employability level of
graduates in Kosovo | Simple linear regression | Accepted |
| H1a | Study programs in higher education have a statistically significant impact on
the employability level of graduates. | Pearson’s correlation &
Simple linear regression | Accepted |
| H1b | Collaboration between the university and businesses has a statistically
significant impact on the employability level of graduates. | Pearson’s correlation &
Simple linear regression | Accepted |
| H1c | Teaching quality has a statistically significant impact on the employability level
of graduates. | Pearson’s correlation &
Simple linear regression | Accepted |
| H1d | Flexibility of Academic Programs has a statistically significant impact on the employability level
of graduates. | Pearson’s correlation &
Simple linear regression | Accepted |
| H1e | Higher Education Infrastructure and Resources have a statistically significant impact on the
employability level of graduates. | Pearson’s correlation &
Simple linear regression | Accepted |
H2: There is a statistically significant difference in the level of employability
of graduates depending on the type of higher education institution (public or
private) | T-test, | Accepted |
H3: There is a statistically significant difference in the level of employability of
graduates depending on the graduate profile characteristics | ANOVA & Post Hoc test | Accepted |
| H3a | There is a statistically significant difference in the employability of graduates
based on their level of education | ANOVA | Rejected |
| H3b | There is a statistically significant difference in the employability of graduates
based on their grade point average. | ANOVA | Rejected |
| H3c | There is a statistically significant difference in the employability of graduates
based on their field of study | ANOVA & Post Hoc test | Accepted |
H4: External drivers of employability have a statistically significant impact on
the employability of graduates | Simple linear regression | |
| H4a | Labor market demand has a statistically significant impact on the employability
of graduates | Simple linear regression | Accepted |
| H4b | Role of technology and innovation has a statistically significant impact on the
employability of graduates | Simple linear regression | Rejected |
Conclusions
The results show discrepancy between the offer of higher education and the demands of the labor market, where despite the fact that some dimensions present positive responses, gaps are evident in terms of the integration of practical skills, the cooperation of higher education institutions with industry, the inclusion of technology and innovation in study programs, leading to a lack of coherence between the objectives of HEIs and the demands of the labor market. This leads to graduates who are not ready to face the challenges posed by the workplace. This paper also presents a significant difference between public and private HE institutions in Kosovo in the context of their preparation for the labor market. Private institutions were found to perform better compared to public institutions in all descriptive dimensions of HE such as study programs (SPH), teaching quality (TQ), university-industry collaboration (CUB), flexibility of academic programs (FAP), infrastructure and resources (HEIR) and labor market demands. This conclusion is supported by statistical inferential tests, demonstrating statistically significant differences.
The latent structures of the variables (SPH, CUB, TQ, FAP, HEIR, LMD) were also confirmed by factor analysis, where according to consistent results these variables represent important factors in the model of labor market preparation. The values of KMO (.889) and Bartlett’s Test, where p<0.05 are indicators of high suitability of the data for factor analysis. The impact of these factors on employability is also presented through Pearson correlation and simple linear regression, where the variables LMD (R2=.390) and FAP (R2=.367) resulted as the factors with the greatest impact on youth employability (EY) followed by TQ and HEIT. The impact of technology and innovation (RTI) resulted as non-significant despite its importance in literature, a result which represents a gap in the inclusion of this variable in study programs. This result was the same for private and public institutions, showing the need for intervention to increase compatibility with technological developments in the market.
According to the results of the t-tests, a statistically significant difference was identified between private and public institutions regarding all variables of the model, where the most pronounced difference resulted in infrastructure and resources, where public institutions do not provide technical support and modern teaching environments. This situation will not only reflect on the quality of learning but will also be an indicator of a decrease in the potential for the development of practical skills, further increasing the gap between the labor market and higher education. Also, the results of One-way ANOVA showed that the field of study FS is an important factor in employability, where graduates of engineering, architecture, computer science and information technology have an advantage in employability in relation to fields such as agriculture, arts, legal sciences and social fields. This fact represents the compliance of these fields with the demands of the labor market, as a result requiring a redesign of the academic offer in accordance with current technological and economic trends. This research is an indicator that higher education in Kosovo is characterized by two relatively opposed realities, where the orientation of private institutions is towards competition and the market, while public ones are led by bureaucracy and lack of reform. This state of public institutions demonstrates the need for radical reform of educational policies, structures, and practices to eliminate the gap that exists between academic preparation and labor market demands.
