Determining factors of labour market outcomes for recently graduated,underrepresented college students

Abstract

This study examined the factors that determined labour market outcomes for recently graduated, underrepresented college students. Chile’s largest higher education institution, which has a significant number of first-generation students from more deprived social sectors, was considered. A quantitative methodology was applied using logistic and multinomial regression models. Occupational status and income level were chosen as the dependent variables and five dimensions of independent variables were considered: sociodemographic attributes, human capital, academic characteristics, personality traits, and work environment. The results indicated that males, graduates who worked during their studies, heads of households, graduates from technical-professional high schools, those who completed their higher education studies in a timely manner, those who worked for larger private companies, and those who worked in a different geographical region to the one in which they studied had better labour market outcomes. Suggestions for institutional practices to help underrepresented students have successful career transitions are discussed.

Keywords

Labour market outcomes occupational status income level underrepresented college students

Introduction

Over the course of the last few decades, access to tertiary education has increased considerably throughout the world. In a process that has involved a transition from elite to mass higher education, new groups – previously lacking the opportunity to access higher education – are able to continue their studies after finishing high school. This has allowed institutions to enrol first-generation students and those from more socially-deprived sectors and more diverse cultural backgrounds, challenging the way institutions prepare their students (Altbach et al., 2010; Schofer & Meyer, 2005).

From 1970 to 2010, the gross enrolment rate in tertiary education increased from 9.7% to 38.0% (UNESCO Institute for Statistics, 2020), and the increase in underrepresented students’ enrolment rates has led to educational institutions developing measures to help them succeed in college. In this context, the following questions about the integration of non-traditional students are raised. Are non-traditional students socially and academically capable of doing well in higher education? Which policies should be developed to support the integration of students from underrepresented groups? (Bailey et al., 2010; Bettinger et al., 2013).

To this end, most initiatives and studies have been concerned with addressing student access to, and success in, higher education. Among these, pre-college outreach programs for high school students, as well as tutoring, counselling, and peer-to-peer mentoring strategies for underrepresented higher education students should be highlighted (Hagedorn et al., 1999; Harrison & Waller, 2017). Similarly, other institutions have focused on dealing mainly with the achievement gaps for these students through remedial courses (Calcagno & Long, 2009; Illich et al., 2004).

The debate about integrating new students into higher education, however, has lacked deeper reflection on how to support the career development of socially underrepresented graduates. While helping students be successful during post-high school education is a highly relevant matter, it is also important to examine the way in which these new students join the labour market, since it is known that they face greater challenges when developing their careers (Garriott, 2020; Gibbons & Woodside, 2014; Paisey et al., 2020).

In addition, most of the research on career paths has lacked a specific focus on diverse populations (Akkermans & Kubasch, 2017). The study of the premises and consequences of career development today is a relevant topic that has contributed significantly by proposing practical recommendations to improve integration into the labour market (Akkermans & Kubasch, 2017). Nevertheless, the majority of these studies have been conducted in developed countries and selective institutions, which makes the study of this topic on socially underrepresented students a relevant avenue to explore.

In this scenario, this study explored the decisive elements of occupational status and income level in recently graduated college students. These two indicators were chosen as they are a good representation of successful graduate transition to the labour market (Essilfie, 2020; Jackson, 2014). To this end, graduates from a non-selective, Chilean institution that offered mostly technical degrees were examined using a quantitative approach through logistic and multinomial regression models.

The study of these phenomena in the Chilean context complies with the need to find out more about the career paths of diverse populations (Akkermans & Kubasch, 2017). Recently, Chile has broadened the scope of its higher education system by incorporating new students who are now starting to develop their careers (Organisation for Economic Co-operation and Development) (OECD, 2017). During Pinochet’s dictatorship in the 1970s and 1980s, Chile implemented a neo-liberal socioeconomic model, whose main pillars included a new higher education system that transferred the cost of education almost completely to students, as well as liberalizing the educational offering by promoting the emergence of new private sector suppliers (Beyer, 2000). Based on these reforms, enrolment rates in the following decades increased significantly. However, there are still relevant quality and equity problems, for example, that students’ socioeconomic backgrounds are still a relevant predictor of access to, and success in, higher education (Ministerio de Educación) (MINEDUC, 2017; OECD, 2017).

Regarding the new types of students accessing higher education in Chile, they are mostly first family members to access higher education, come from lower socioeconomic groups, and have received poorer-quality secondary education. Moreover, most of the studies concerning them have focused on college access and success (see Bordón et al., 2015; Venegas-Muggli, 2020). Therefore, there is a need to complement previous studies by exploring the challenges that new students accessing higher education face when beginning their careers from a perspective that considers multiple dimensions that might affect labour market outcomes.

In this context, this study aimed to explore how five important dimensions determined graduates’ occupational status and income level: these dimensions were sociodemographic attributes, human capital, academic characteristics, personality traits, and work environment. These five dimensions represent different areas from the literature that have been highlighted as relevant to explain labour-market integration, which has been defined as a multi-causal phenomenon (Choi, 2015; Ng & Feldman, 2014). Given this, and considering the exploratory nature of this study in a little researched area, all of these dimensions are considered equally, without defining specific hypotheses regarding which variables should have greater relevance in explaining occupational status and income level.

Determining factors of labour market outcomes among recently graduated students

Considering the variety of factors involved in career path development, some authors have suggested different typologies of features associated with labour market outcomes. Choi (2015), for example, argued that there were five dimensions to career development: personality, demographics, motivation, work environment, and home environment. Similarly, in their meta-analysis of objective and subjective career success predictors, Ng et al. (2005) argued that there were four categories of predictor: human capital, organizational sponsorship, sociodemographic status, and stable individual differences.

These existing categorizations, however, are insufficient to explain labour market outcomes for recently graduated students, since they refer to a more general employee profile. Therefore, for this research, a new typology of labour market determining factor outcomes was proposed, combining elements from previous literature on the topic. Specifically, the following five factors were defined as the key areas linked to career path development in recently graduates: sociodemographic attributes, human capital, academic characteristics, personality traits, and work environment.

According to Ng et al. (2005), sociodemographic attributes are understood as characteristics that reflect an individual’s demographic and social background, such as gender, age, marital status, socioeconomic level, and race. When reviewing the literature on this topic, there were several studies that examined the importance of people’s background to their career paths. Among them, the studies that examined a gender gap that favours male employees should be highlighted (Frear et al., 2019; Martin, 2017). Similarly, other studies have shown that age, marital status, and socioeconomic status also affect employment outcomes (Jackson, 2014; Paisey et al., 2020).

The concept of human capital is defined as a set of personal conditions that might affect individual career paths, such as education, knowledge, work experience, and training (González Romá et al., 2016). Several studies have shown how higher levels of human capital are linked to better job opportunities, since many job offers require specific levels and types of education (Fugate et al., 2004).

The third dimension involves academic characteristics, which are particularly relevant when studying the career paths of recent graduates. These refer to both objective graduate attributes, such as program duration and field of study, as well as how graduates perform during their formal studies. In addition, this dimension includes high school attributes. As far as college attributes are concerned, in a study of community college graduates, Faber and Slantcheva-Durst (2020) found that the field of study and the number of students who received federal loans significantly predicted graduate earnings. In relation to student performance, Essilfie (2020) showed that the overall grade point average was associated positively with earnings in a sample of Canadian graduates.

Personality traits are broadly understood to be referring to diverse aspects of people’s personalities, such as emotional adjustment, motivation, self-management, extraversion, and openness, all of which affect career success (Choi, 2015; Ng & Feldman, 2014). This implies that not only do people’s structural conditions affect their integration into the labour market but also how they are disposed to develop in these conditions.

The last proposed dimension links career paths with the work environment. This refers to specific attributes of employee workplaces that might affect labour market outcomes, either directly or by moderating other variables. Some of these studies have shown how company training or mentoring programs and other types of assistance for employees facilitated career success (Choi, 2015; Ng et al., 2005; Yu, 2012). Similarly, other studies have considered how the specific attributes of certain companies – such as size, field of operation, and type of contracts offered – are linked to labour outcomes (Guarnaccia et al., 2018; Rajendran et al., 2020).

Socially underrepresented students and career development

Garriott’s (2020) critical cultural wealth model (CCWM) for understanding first-generation and economically marginalized college (FGEM) students’ academic and career development is a relevant and useful framework to support the analysis of the determining factors of socially underrepresented students’ labour market outcomes. Garriott (2020) argued that the unique experiences of FGEM students required a framework that captures their specific strengths and challenges. To this effect, the CCWM – instead of focusing on how career choices are influenced by perceived abilities and interests – places special emphasis on the ability to successfully negotiate external influences and make choices that reflect people’s values and principles.

This model suggests considering four dimensions when studying FGEM students’ academic and career paths. First, Garriott (2020) highlighted the relevance of structural and institutional conditions. These are defined as any institutionalized policy or practice that leads to a specific form of oppression, such as exploitation and/or marginalization. The second dimension is the social-emotional crossroads, which is understood as students’ level of stress at higher education institutions in several domains, such as campus cultural fit and family support for their decision to go to college. Third, Garriott emphasized self-authorship, which refers to one’s ability to make self-reflective career decisions, taking context into consideration and bearing in mind one’s capacity for agency and problem-solving. The last dimension is cultural wealth, defined as the assets, strengths, and capital of marginalized groups, which are used positively to obtain better academic and career outcomes.

The CCWM provides a relevant framework for understanding the particularities of career path development for socially underrepresented students. Although this framework provides different, but overlapping, dimensions of analysis, these will not be the main focus in this study as factors to define specific variables linked with labour market outcomes. Rather, this model will complement the typology of this phenomenon’s determining factors as specified in the previous section, considering it as a perspective that overlaps all previously defined dimensions (sociodemographic attributes, personality traits, human capital, academic characteristics, and work environment). Therefore, the CCWM will be considered when interpreting the relationships among the five factors and labour market outcomes

Method

Procedure

To examine the determining factors of labour market outcomes for recently graduated college students, we considered a representative sample of graduates who had finished their studies at the INACAP Technical Training Center (TTC) and the INACAP Professional Institute (PI) in 2019. INACAP (the National Training Institute) is an integrated Chilean higher education facility, made up of these two institutions (the PI and the TTC) and the INACAP Technological University of Chile. This study only considered students from the IP and the TTC, since these are focused on vocational education and therefore enrol a higher proportion of socially underrepresented students. INACAP is Chile’s largest higher education institution, with approximately 130,000 students spread over 27 campuses throughout the country. It is a non-profit, private, non-selective institution, which means it receives a significant number of underrepresented social groups, particularly first-generation students from low and middle-income families.

Participants

The sample was made up of 5,650 graduates from 55 different study programs taken from the population of graduates who finished their studies in 2019. To select the respondents, a stratified random sampling procedure was used. Strata were defined, considering both the graduates’ study program and campus, in order to have representative samples from every study program on each campus. The survey was carried out by a public opinion research agency (Ekhos) using computer-assisted telephone interviewing (CATI). All respondents were asked to give their informed consent before answering.

The average age of graduates was 25.8 years, 67% were male, and 68% had worked while they were studying. Regarding the fields of knowledge involved, 64% had degrees in the fields of engineering or technology and 36% in other subjects. To examine the determining factors of occupational status, we only considered graduates who were active in the labour market (either employed or unemployed and looking for a job) at the time of the survey. This sample included 5,092 graduates, of whom 86% were employed. To analyse the determining factors of income level, only graduates who were employed were considered. This sample contained 4,082 graduates.

Variables

This study considered the following two dependent variables measuring labour market outcomes:

Occupational status

This variable had three categories: employed, unemployed, and inactive (unemployed and not looking for a job). These categories were created from graduates’ responses to seven specific questions about their employment activities in the week prior to the survey. Inactive graduates were not considered in the analysis, thus, this variable was considered as a dummy indicator (employed/unemployed).

Income level

This was operationalized as three categories: high, medium, and low income. These three categories were created by coding responses to the following question: What was the range of your monthly net salary or total income (considering all the jobs you do)? The categories were defined using groups of equal sizes.

Regarding the independent variables, several indicators were measured based on the previously defined five key areas linked to the development of career paths of recently graduated students:

Sociodemographic attributes

This dimension considered graduates’ gender, age, and socioeconomic level, as well as a variable that indicated whether they were the head of their household.

Human capital

This dimension considered three indicators: whether the graduate was a first-generation higher education student, the number of years of schooling of their mother, and a dummy variable that indicated if the graduate had worked or not during their studies.

Academic characteristics

The following academic attributes were considered: type of high school (public/private), type of higher education (technical-professional/scientific-humanistic), type of institution (TTC/PI), type of student (daytime/evening), and field of study (engineering and technology/other), as well as an indicator that measured whether the graduate completed their studies in a timely manner or not and their final grade for their degree as a measure of academic achievement.

Personality traits

There were no available variables associated with this topic. Instead of this, to be able to incorporate some indicators on student perceptions, two indices about satisfaction and perceived abilities were included. First was an index of institutional satisfaction (8 items; 5-point Likert scale; e.g., “Lecturers with experience and knowledge in each discipline”; α = .82). The second index measured the perceived abilities of INACAP graduates (4 items; 1 = strongly disagree to 5 = strongly agree; e.g., “They are organized and plan in order to achieve their aims”; α = .80.

Work environment

The following characteristics of employed graduates were examined: type of contract (indefinite/fixed term), type of employer (public/private), company size (small/medium/large), type of workday (full-time/part-time), and whether they worked in the same geographical region as where they completed their studies.

The study also included the graduates’ campus as a control variable.

Data analysis

The study used a quantitative design to examine the determining factors of both occupational status and income level. First, multiple logistic regression models were used to analyse the determining factors of occupational status (employed/unemployed), with all the indicators described in the previous section – with the exception of those measuring the work environment – employed as independent variables. In the second stage, multinomial regression models were used to examine the determining factors of income level (low/medium/high), considering the indicators of the five dimensions theoretically defined as key areas in recent graduates’ career paths.

Additionally, we also estimated interaction effects in order to examine how the effects of specific independent variables varied according to different levels of a second independent variable. The results of these interactions are presented using the estimation of fitted probabilities.

Results

Occupational status

Table 1 presents the results from the estimated models predicting occupational status. Two models were fitted considering the different variable specifications. Model 1 included all previously described variables considered as potential factors associated with employment outcomes, with the exception of the indicators measuring the work environment. Beta coefficients, their standard errors, and beta exponentials are reported.

Table 1.

Logistic regression for predicting occupational status.

	Model 1		Model 2
	b (se)	exp (b)	b (se)	exp (b)
Variables
Gender (female)	−0.44***	0.64	−0.41***	0.66
Gender (female)	(0.08)	0.64	(−0.12)	0.66
Age	−0.05***	0.95	–	–
Age	(0.01)	0.95	–	–
Head of household	1.06***	2.88	0.926***	2.52
Head of household	(0.41)	2.88	(0.12)	2.52
Socioeconomic level (low)
Middle–Low	−0.12	0.89	−0.08	0.93
Middle–Low	(0.13)	0.89	(0.14)	0.93
Middle	0.03	1.03	0.001	1.00
Middle	(0.21)	1.03	(0.19)	1.00
Middle–High	0.31	1.37	0.27	1.30
Middle–High	(0.37)	1.37	(0.25)	1.30
High	0.53	1.70	0.54	1.71
High	(0.94)	1.70	(0.55)	1.71
Worked during studies	0.90***	2.46	0.71***	2.03
Worked during studies	(0.27)	2.46	(0.10)	2.03
Mothers’ years of education	0.004	0.02	–	–
Mothers’ years of education	(0.02)	0.02	–	–
First college generation	0.14	1.15	0.14	1.16
First college generation	(0.14)	1.15	(0.10)	1.16
Secondary school type (public)
Private-subsidized	0.00	1.00	0.10	1.11
Private-subsidized	(0.13)	1.00	(0.12)	1.11
Delegated administration	−0.25	0.78	−0.14	0.87
Delegated administration	(0.16)	0.78	(0.20)	0.87
Private	−0.65	0.52	−0.48	0.62
Private	(0.27)	0.52	(0.50)	0.62
Secondary education type (technical/professional)
Scientific-humanistic	−0.31	0.74	−0.32*	0.73
Scientific-humanistic	(0.12)	0.74	(0.15)	0.73
Mixed	−0.14	0.87	−0.06	0.94
Mixed	(0.12)	0.87	(0.13)	0.94
Institution (TTC)	0.13	1.14	0.18	1.20
Institution (TTC)	(0.12)	1.14	(0.10)	1.20
Field of study (other)	0.05	1.06	0.12	1.12
Field of study (other)	(0.14)	1.06	(0.12)	1.12
Final mark	0.42**	1.52	–	–
Final mark	(0.21)	1.52	–	–
Higher education timely completion	0.06	1.07	0.21*	1.23
Higher education timely completion	(0.13)	1.07	(0.11)	1.23
Type of student (daytime)	−0.25*	0.78	−0.18	0.83
Type of student (daytime)	(0.10)	0.78	(0.12)	0.83
Institutional satisfaction	0.75	2.13	–	–
Institutional satisfaction	(0.84)	2.13	–	–
Perceived abilities	0.26	0.46	–	–
Perceived abilities	(0.46)	0.46	–	–
Constant	−1.09	0.34	0.52	1.69
Nagelkerke R²	.09		.08
Hosmer-Lemeshow χ²	6.63		8.63
Brier score	.11		.11
N	3970		4399

TTC: technical training center.

*p < .05, **p < .01, ***p < .001.

This first model included several significant associations between the independent variables and the odds of being employed. Nevertheless, this model had multicollinearity violations when the variance inflation factor (VIF) was estimated for each of the included variables. As stated by Kline (1998), a VIF threshold of 5 should be considered when examining regression models. In the case of Model 1, the VIF analysis showed five variables with values > 5 (age, mother’s years of education, final grade, institutional satisfaction, and perceived abilities). This meant that the other variables in the model were highly correlated to these five variables, something that was to be expected, since this model included variables that are known to be correlated with age, indicators of human capital, and academic performance, such as socioeconomic level and being the head of a household.

Given these results, a second model was fitted that excluded these five variables. This generated a new model with no multicollinearity problems, and a good fit. Although Nagelkerke’s R-square was low (0.09), the Hosmer-Lemeshow goodness-of-fit test indicated that the model was well calibrated, since there was no evidence to reject the null hypothesis that the model is not acceptable. Additionally, when the Brier score was estimated, a value of .11 was obtained, indicating that the model accurately predicted employment probability. The coefficients of this model were therefore examined to analyse the association between the defined independent variables and occupational status.

In relation to sociodemographic attributes, Model 2 from Table 1 shows that there is a significant association at the 1% level between gender and occupational status. Specifically, when controlling for all the other variables included in the model, it was estimated that the odds of being employed were 33.6% lower for women than for men. Similarly, also estimated at the 1% level, the odds of being employed were 152% higher for graduates who were heads of households. Finally, it is observed that occupational status and socioeconomic level were not significantly associated.

With regard to human capital, graduates who worked during their studies were more likely to get a job. On the other hand, being first-generation college students did not affect graduates’ chances of getting a job.

Regarding academic characteristics, those who graduated from technical-professional high schools had significantly better chances of being employed than those who graduated from scientific-humanistic high schools. In addition, it was estimated – at the 10% level – that those who completed their higher education studies in a timely manner were 23% more likely to be employed. It also was observed that neither high school type, higher education institution, field of study, nor type of student were associated significantly with occupational status.

One last analysis on occupational status involved examining for possible interaction effects to explain the factors associated with this variable. For this purpose, several interaction terms were included in a third fitted model to examine whether the effects of specific independent variables on the probability of being employed varied according to different levels of a second independent variable. Of all the included interaction terms, only two were statistically significant: gender/working during studies and gender/field of study.

To facilitate the interpretation of these interactions, the fitted probabilities of being employed were estimated. These analyses used regression coefficients from this third fitted model to predict the probability of being employed for the combination of all categories of the interaction terms’ two independent variables, controlling for the rest of the model’s variables. To do so, the values of all model variables were fixed at a specific value (the mode or mean of distribution), except for those whose associations were being examined.

Table 2 shows that the effect of gender on being employed varies if this association is separately analysed according to employment status during college studies. While the difference in the probability of men and women being employed is 15.7% for graduates who worked during their studies, this is 11.1% when only graduates who did not work during their studies are considered. Similarly, in Table 3, it can be seen that the favourable difference in job attainment for men is significantly higher for graduates in the field of engineering and technology than for graduates in other fields.

Table 2.

Fitted probabilities of being employed (1).

Gender	Worked during studies
Gender	Yes	No
Female	68.0%	57.9%
Male	83.7%	69.0%
Difference	15.7%	11.1%

Table 3.

Fitted probabilities of being employed (2).

Gender	Field of study
Gender	Engineering and technology	Other
Female	66.0%	76.6%
Male	83.7%	79.9%
Difference	17.7%	3.3%

Income level

In relation to the analysis of income level determining factors, multinomial regression models were applied. When there is a dependent variable for three categories, this method fits two logistic regression models, considering two of the three possible combinations of these three categories. For these specific analyses, two different types of model were fitted to predict income level: (a) medium versus low income level, and (b) high versus low income level.

Table 4 presents two models that allow for factors associated with medium-income level instead of low-income level employment to be examined. Model 1 includes all the defined independent variables. This model, however, had multicollinearity violations, as five variables had a VIF score > 5 (age, mothers’ years of education, final mark, institutional satisfaction, and perceived abilities). Therefore, Model 2 was fitted excluding these variables. This new model was accurate, based on the results of the Hosmer-Lemeshow goodness-of-fit and the Brier score, meaning it can be used to examine income-level determining factors.

Table 4.

Logistic regression for predicting income level (medium/low).

	Model 1		Model 2
	b (se)	exp (b)	b (se)	exp (b)
Variables
Gender (female)	−0.77***	0.46	−0.67***	0.51
Gender (female)	(0.14)	0.46	(0.13)	0.51
Age	0.07***	1.07	–	–
Age	(0.02)	1.07	–	–
Head of household	0.51***	1.71	0.68***	1.96
Head of household	(0.13)	1.71	(0.12)	1.96
Socio-economic level (low)
Middle–Low	0.35**	1.42	0.22	1.24
Middle–Low	(0.16)	1.42	(0.15)	1.24
Middle	0.06	1.06	−0.04	0.96
Middle	(0.21)	1.06	(0.19)	0.96
Middle–High	0.67**	1.96	0.59*	1.80
Middle–High	(0.29)	1.96	(0.27)	1.80
High	0.26	1.30	0.44	1.56
High	(0.74)	1.30	(0.75)	1.56
Mothers’ years of education	−0.04**	0.96	–	–
Mothers’ years of education	(0.02)	0.96	–	–
First college generation	−0.18	0.84	−0.02	0.99
First college generation	(0.13)	0.84	(0.12)	0.99
Worked during studies	0.03	1.03	0.08	1.09
Worked during studies	(0.13)	1.03	(0.12)	1.09
Secondary school type (public)
Private-subsidized	0.24*	1.27	0.20	1.22
Private-subsidized	(0.14)	1.27	(0.13)	1.22
Delegated administration	−0.30	0.74	−0.30	0.74
Delegated administration	(0.25)	0.74	(0.23)	0.74
Private	0.58	1.78	0.12	1.13
Private	(0.70)	1.78	(0.68)	1.13
Secondary education type (technical/professional)
Scientific-humanistic	−0.20	0.82	−0.18	0.82
Scientific-humanistic	(0.16)	0.82	(0.15)	0.82
Mixed	−0.26*	0.77	−0.20	0.82
Mixed	(0.15)	0.77	(0.14)	0.82
Institution (TTC)	−0.65***	0.52	−0.69***	0.50
Institution (TTC)	(0.13)		(0.12)
Field of study (Other)	−0.36**	0.70	−0.44**	0.64
Field of study (Other)	(0.14)		(0.13)
Final mark	0.11	1.12	–	–
Final mark	(0.15)	1.12	–	–
Higher education timely completion	0.18	1.20	0.20	1.22
Higher education timely completion	(0.14)	1.20	(0.12)	1.22
Type of student (daytime)	−0.18	0.84	−0.42**	0.66
Type of student (daytime)	(0.14)	0.84	(0.13)	0.66
Institutional satisfaction	−0.49	0.61	–	–
Institutional satisfaction	(0.47)	0.61	–	–
Perceived abilities	0.99**	2.70	–	–
Perceived abilities	(0.43)	2.70	–	–
Type of contract (indefinitive)	0.26**	1.29	0.25*	1.29
Type of contract (indefinitive)	(0.12)	1.29	(0.12)	1.29
Company size (big)
Small	−1.50***	0.22	−1.48***	0.23
Small	(0.16)	0.22	(0.15)	0.23
Medium	−0.52***	0.60	−0.54***	0.58
Medium	(0.13)	0.60	(0.12)	0.58
Works in the same region he/she studied	−0.21	0.81	−0.27	0.77
Works in the same region he/she studied	(0.19)	0.81	(0.18)	0.77
Type of workday (full-time)	1.74***	5.72	1.80***	6.05
Type of workday (full-time)	(0.17)	5.72	(0.17)	6.05
Type of employer (private)	0.55**	1.73	0.54**	1.71
Type of employer (private)	(0.18)	1.73	(0.17)	1.71
Constant	−2.17*	0.12	0.19	1.20
Constant	(1.10)	0.12	0.54	1.20
Nagelkerke R²	.24		.231
Hosmer-Lemeshow χ²	7.16		12.22
Brier score	.17		.168
N	2080		2287

TTC: technical training center.

*p < .05, **p < .01, ***p < .001.

Regarding sociodemographic attributes, Model 2 in Table 4 shows that the odds of having a medium-income level job rather than a low-income level one were 48.8% lower for women than for men. In the same way, it was estimated that the odds of having a better-paid job were 96% higher for heads of households. Additionally, the results showed that graduates’ socioeconomic level did not significantly determine their income level.

Regarding human capital indicators, neither being first-generation college students nor employment condition during studies significantly affected recent graduates’ chances of having a medium-income level job instead of a low-income one.

As far as academic characteristics were concerned, three variables associated with income level were found to be statistically significantly: type of student, type of higher education institution, and field of study. It was estimated that evening students, those who graduated from a professional institute rather than a technical training centre, and graduates from the field of engineering and technology had higher odds of getting a better-paid job.

Finally, regarding the work environment dimension, four variables were found to affect graduates’ income level. As expected, it was estimated that full-time employees with permanent contracts had better-paid jobs. In addition, it was observed that bigger companies offer better salaries. For instance, it can be seen that the odds of having medium-income level rather than low-income level employment are 77.4% lower for graduates who work for a small company (1–10 employees) than for a larger one (+200 employees). Last, the odds of having better-paid employment were also higher for those who worked in private companies.

Regarding interaction effects, a third model was fitted, in which the interaction between gender and company size only was found to be statistically significant. Table 5 shows this interaction through fitted probabilities analyses, with the gender gap favouring men being much larger in smaller companies. While the difference in the probability of having a medium-income level job instead of a low-income one between men and women was 32.4% in companies with a maximum of 10 employees, it was only 3.8% in companies employing 200 people or more.

Table 5.

Fitted probabilities of having a medium income level.

Gender	Company size (number of employees)
Gender	Small (1–10)	Medium (11–199)	Big (200+)
Female	45.0%	79.8%	88.1%
Male	77.5%	87.5%	91.9%
Difference	32.4%	7.7%	3.8%

Table 6 shows the second part of the analysis of income level determining factors by presenting two logistical regression models that estimated the probability of having a high income level job rather than a low income one. The first model considered all variables and Model 2 excluded the same five variables that were omitted in previous models due to multicollinearity. This second model presented a good fit, which meant it can be examined in order to understand the determining factors of labour market outcomes.

Table 6.

Logistic regression for predicting income level (high/low).

	Model 1		Model 2
	b (se)	exp (b)	b (se)	exp (b)
Variables
Gender (female)	−1.09***	0.34	−0.78**	0.46
Gender (female)	(0.26)	0.34	(0.24)	0.46
Age	0.18***	1.20	–	–
Age	(0.03)	1.20	–	–
Head of household	1.28***	3.59	1.83***	6.22
Head of household	(0.21)	3.59	(0.19)	6.22
Socio-economic level (low)
Middle–Low	0.25	1.28	−0.02	0.98
Middle–Low	(0.29)	1.28	(0.25)	0.98
Middle	−0.04	0.96	−0.15	0.86
Middle	(0.34)	0.96	(0.31)	0.86
Middle–High	0.71	2.03	0.58	1.79
Middle–High	(0.51)	2.03	(0.45)	1.79
High	−0.42	0.66	0.07	1.07
High	(0.82)	0.66	(0.71)	1.07
Worked during studies	0.12	1.13	0.43*	1.53
Worked during studies	(0.24)	1.13	(0.22)	1.53
Mothers’ years of education	−0.03	0.97	–	–
Mothers’ years of education	(0.03)	0.97	–	–
First college generation	−0.40*	0.67	−0.11	0.89
First college generation	(0.23)	0.67	(0.18)	0.89
Secondary school type (Public)
Private-subsidized	0.47	1.60	0.27	1.31
Private-subsidized	(0.24)	1.60	(0.22)	1.31
Delegated administration	0.39	1.48	0.26	1.30
Delegated administration	(0.39)	1.48	(0.35)	1.30
Private	2.00**	7.42	1.67*	5.32
Private	(0.86)	7.42	(0.78)	5.32
Secondary education type (technical/professional)
Scientific-humanistic	−0.62	0.54	−0.56*	0.60
Scientific-humanistic	(0.27	0.54	(0.26)	0.60
Mixed	−0.64**	0.53	−0.52*	0.60
Mixed	(0.26)	0.53	(0.23)	0.60
Institution (TTC)	−1.18***	0.31	−1.04***	0.36
Institution (TTC)	(0.21)	0.31	(0.19)	0.36
Field of study (Other)	−1.08***	0.34	−1.07***	0.34
Field of study (Other)	(0.28)	0.34	(0.25)	0.34
Final mark	0.24	1.27	–	–
Final mark	(0.26)	1.27	–	–
Higher education timely completion	−0.05	0.95	0.01	1.01
Higher education timely completion	(0.22)	0.95	(0.18)	1.01
Perceived abilities	1.20*	3.31	–	–
Perceived abilities	(0.68)	3.31	–	–
Institutional satisfaction	−0.47	0.62	–	–
Institutional satisfaction	(0.70)	0.62	–	–
Type of student (daytime)	−0.44**	0.64	−0.93***	0.39
Type of student (daytime)	(0.20)	0.64	(0.19)	0.39
Type of contract (indefinitive)	0.48**	1.62	0.43**	1.53
Type of contract (indefinitive)	(0.23)	1.62	(0.20)	1.53
Company size (big)
Small	−2.35***	0.10	−2.55***	0.07
Small	(0.30)	0.10	(0.29)	0.07
Medium	−1.36***	0.26	−1.33***	0.26
Medium	(0.21)	0.26	(0.20)	0.26
Works in the same region he/she studied	−0.89***	0.41	−0.95***	0.39
Works in the same region he/she studied	(0.28)	0.41	(0.26)	0.39
Type of workday (full-time)	1.61***	4.99	1.70***	5.49
Type of workday (full-time)	(0.30)	4.99	(0.29)	5.49
Type of employer (private)	0.851**	2.34	0.80***	2.22
Type of employer (private)	(0.29)	2.34	(0.28)	2.22
Constant	−5.38	0.005	0.69	1.99
Constant	1.79	0.005	0.75	1.99
Nagelkerke R²	.54		.50
Hosmer-Lemeshow χ²	8.75		10.42
Brier score	.28		.27
N	1320		1454

TTC: technical training center.

*p < .05, **p < .01, ***p < .001.

In relation to sociodemographic attributes, gender and being the head of household were significantly associated with income level. First, at the 1% level, it was estimated that the odds of having a high-income job instead of a low-income one were 54.4% lower for women than for men. Similarly, the heads of households were six times more likely to be in the highest income level category.

Regarding the variables measuring human capital, and unlike the previous model, a statistically significantly association was found. Specifically, it was observed that the odds of having a high-income level job were 53.1% higher for the graduates who worked during their studies.

With regards to academic characteristics, four significant associations were estimated. Graduates who attended technical-professional high schools, higher education evening students, those who graduated from the professional institute, and those whose field of study was engineering and technology had higher odds of getting a better-paid job.

For work environment indicators, five significant income level determining factors were identified. Specifically, those who worked in bigger-size companies, had permanent contracts, worked in a different region to the one they studied in, had a full-time job, and worked in private companies had better odds of having high-income level employment.

Finally, regarding the interaction effects for this second model for predicting income level, only the interaction between gender and type of institution was statistically significant. See Table 7. The differences observed between men and women concerning the probability of having a high-income job varied according to the institution they attended. Although in both institutions, men were more likely to find a better-paid job, this difference was much greater among graduates from the professional institute.

Table 7.

Fitted probabilities of having a high income level.

Gender	Institution
Gender	TTC	PI
Female	45.1%	51.6%
Male	55.7%	81.0%
Difference	−10.6%	−29.4%

PI: professional institute; TTC: technical training center.

Discussion

This study aimed to examine the determining factors of both occupational status and income level in recent graduates of a technical, non-selective, Chilean higher education institution. Given this context, these associations were examined from theoretical perspectives that placed a special emphasis on the career development of socially underrepresented students.

Specifically related to occupational status, it was found that men, heads of households, and those who worked during their studies had a greater chance of being employed. Additionally, those who graduated from technical-professional high schools and who completed their higher education studies in a timely manner had significantly better chances of getting a job after graduating. Finally, the effect of gender on employment status favoured men when those who worked during their studies and who got a degree in the field of engineering and technology were considered.

Concerning income level, men and heads of households had higher odds of having better-paid employment. In addition, better employment outcomes regarding salary were found among evening students, those who graduated from the professional institute and the engineering and technology field, and those who worked during their studies. Regarding work environment variables, indefinite and full-time employees, those who worked for larger and private companies, and those who worked in a different geographical region to the one they studied in had better salaries. Last, interaction effects indicated that gender gaps in salary were higher for those working for small companies and who graduated from the professional institute.

A relevant finding that should be highlighted is the low predictive power of socioeconomic status variables on both indicators of labour market outcomes. This implies that, in the case of technical graduates in Chile, socioeconomic background does not play a relevant role in this aspect of career development. This could be explained by what has been called the declining effect of social origin, which states that the impact of social origin declines over time, given educational expansion (Breen & Jonsson, 2007).

In sum, these results showed that both occupational status and income level were determined by several variables belonging to diverse dimensions such as sociodemographic characteristics and academic attributes. Thus, this evidence reinforces the idea that when career development is being studied, a multi-dimensional perspective must be considered (Choi, 2015; Ng & Feldman, 2014).

When linking these results to the career development literature of socially underrepresented students, several theoretical implications can be considered regarding Garriott’s (2020) critical cultural wealth model (CCWM). CCWM proposes four dimensions to understand underrepresented students’ academic and career paths: structural and institutional conditions, social-emotional crossroads, self-authorship, and cultural wealth. Regarding structural and institutional conditions, it was observed that being female was the greatest structural limitation that students faced, especially for those who graduated in the field of engineering and technology and who worked in small-size companies. Also, it is relevant to highlight that, contrary to Garriott’s expectations, graduates’ socioeconomic background did not limit their career development.

As far as the social-emotional crossroads were concerned, the results also differed from Garriott’s (2020) approach. In Garriott’s study, having greater domestic responsibilities negatively affected academic careers. In our study, students with more responsibilities during their studies (those who worked during their studies and who were heads of households) got better job results. Given this, our results showed that, in the case of Chilean technical students, greater responsibilities – rather than being a mechanism that increases stress – positively influenced professional career development.

As to self-authorship, the results identified two types of decisions made by students that were associated with better job placement. The first was that the chance of job success after graduating from technical careers increased if the student took a technical course at high school. At the same time, it was observed that seeking employment in geographical areas other than the ones where they graduated was also a type of self-reflective decision that potentially encouraged graduates to consider their context and capabilities.

Finally, in relation to cultural wealth, the study did not measure how certain assets of marginalized groups could be associated with better career outcomes. Future studies of this type should try to measure more directly graduates’ perceptions of their resources that favoured their labour integration processes.

Different practical implications also can be inferred from these results, specifically regarding initiatives for students during their last years of higher education. First, when the gender gap differs in relation to the field of study and the size of the company, working on programs that promote female participation through networks of alumni women is recommended, with networking and making contacts encouraged through talks or mentoring initiatives. At the same time, these findings lead us to recommend that students work during their studies, since this seems to favour employment. Therefore, improving support for working students and providing flexibility so that they have successful academic trajectories is recommended. Likewise, students obtaining jobs in their area of study should be supported. Finally, technical higher education institutions are also recommended to encourage graduates to seek job opportunities in geographical areas different to the ones in which they studied.

Regarding study limitations, the most important one was the inability to measure personality traits directly, such as emotional adjustment and self-management, although this study did measure institutional satisfaction and perceived skills. Future studies should consider the inclusion of these types of variables. Future studies also could conduct qualitative studies to generate rich, in-depth information to further understand and explicate this study’s findings. For example, why the gender wage gap is greater in smaller companies could be better understood through focus groups with men and women who worked in companies of different sizes.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Juan I Venegas-Muggli

Claudia Cifuentes-Donald

References

Akkermans

Kubasch

(2017). Trending topics in careers: A review and future research agenda. Career Development International, 22(6), 56–627. https://doi.org/10.1108/CDI-08-2017-0143

Altbach

P. G.

Reisberg

Rumbley

L. E.

(2010). Tracking a global academic revolution. Change: The Magazine of Higher Learning, 42, 30–39. https://doi.org/10.1080/00091381003590845

Bailey

Jeong

D. W.

Cho

S. W.

(2010). Referral, enrollment, and completion in developmental education sequences in community colleges. Economics of Education Review, 29, 255–270. https://dx-doi-org.web.bisu.edu.cn/10.1016/j.econedurev.2009.09.002

Bettinger

E. P.

Boatman

Long

B. T.

(2013). Student supports: Developmental education and other academic programs. The Future of Children, 23, 93–115. https://doi.org/10.1353/foc.2013.0003

Beyer, H. (2000). Entre la Autonom_ıa y la Intervenci_on: Las Reformas de la Educaci_on en Chile [Between autonomy and intervention: Educational reforms in Chile]. In F. Larra_ın & R. Vergara, (Eds.), La Transformaci_on Econ_omica de Chile [Chile’s economic transformation] (pp. 644–709). Centro de Estudios Públicos.

Bordón

Canals

Rojas

(2015). Retención en los programas e instituciones de educación superior: Nueva evidencia Para Chile [Retention in higher education programs and institutions: New evidence for Chile]. Estudios de Política Educativa, 2, 176–214. http://biblioteca.digital.gob.cl/bitstream/handle/123456789/261/NOTA%20DE%20INV.%20EPE%20N2.pdf?sequence=1&isAllowed=y

Breen

Jonsson

(2007). Explaining change in social fluidity: Educational equalisation and educational expansion in twentieth-century Sweden. American Journal of Sociology, 112, 1775–1810. https://doi.org/10.1086/508790

Calcagno

J. C.

Long

B. T.

(2009). Evaluating the impact of remedial education in Florida community colleges: A quasi-experimental regression discontinuity design. https://files.eric.ed.gov/fulltext/ED506374.pdf

Choi

(2015). Career success factors of women in the public sector: An empirical analysis of the Korean central government. Asian Social Science, 11(9), 300–311. https://doi.org/10.5539/ass.v11n9p300

10.

Essilfie

(2020). Grades and labour market earnings in Canada: New evidence from the 2018 National Graduates Survey (Doctoral dissertation. Lethbridge: University of Lethbridge. https://opus.uleth.ca/bitstream/handle/10133/5733/ESSILFIE_JOSEPHINE_MA_2020.pdf?sequence=3

11.

Faber

Slantcheva-Durst

(2020). The impact of community college attributes on the earnings of their students. Community College Journal of Research and Practice. https://doi.org/10.1080/10668926.2020.1798302

12.

Frear

K. A.

Paustian‐Underdahl

S. C.

Heggestad

E. D.

Walker

L. S.

(2019). Gender and career success: A typology and analysis of dual paradigms. Journal of Organizational Behavior, 40, 400–416. https://doi.org/10.1002/job.2338

13.

Fugate

Kinicki

A. J.

Ashforth

B. E.

(2004). Employability: A psycho-social construct, its dimensions, and applications. Journal of Vocational Behavior, 65, 14–38. https://doi.org/10.1016/j.jvb.2003.10.005

14.

Garriott

P. O.

(2020). A critical cultural wealth model of first-generation and economically marginalized college students’ academic and career development. Journal of Career Development, 47, 80–95. https://dx-doi-org.web.bisu.edu.cn/10.1177/0894845319826266

15.

Gibbons

M. M.

Woodside

(2014). Addressing the needs of first‐generation college students: Lessons learned from adults from low‐education families. Journal of College Counseling, 17, 21–36. https://doi.org/10.1002

16.

González-Romá

Gamboa

J. P.

Peiró

J. M.

(2018). University graduates’ employability, employment status, and job quality. Journal of Career Development, 45, 132–149. https://doi.org/10.1177/0894845316671607

17.

Guarnaccia

Scrima

Civilleri

Salerno

(2018). The role of occupational self-efficacy in mediating the effect of job insecurity on work engagement, satisfaction and general health. Current Psychology, 37, 488–497. https://doi.org/10.1007/s12144-016-9525-0

18.

Hagedorn

L. S.

Siadat

M. V.

Fogel

S. F.

Nora

Pascarella

E. T.

(1999). Success in college mathematics: Comparisons between remedial and nonremedial first-year college students. Research in Higher Education, 40, 261–284. https://doi.org/10.1023/A:1018794916011

19.

Harrison

Waller

(2017). Success and impact in widening participation policy: What works and how do we know? Higher Education Policy, 30(2), 141–160. https://doi.org/10.1057/s41307-016-0020-x

20.

Illich

P. A.

Hagan

McCallister

(2004). Performance in college-level courses among students concurrently enrolled in remedial courses: Policy implications. Community College Journal of Research and Practice, 28, 435–453. https://doi.org/10.1080/10668920490444463

21.

Jackson

(2014). Factors influencing job attainment in recent bachelor graduates: Evidence from Australia. Higher Education, 68(1), 135–153. https://doi.org/10.1007/s10734-013-9696-7

22.

Kline, R. B. (1998). Principles and Practice of Structural Equation Modeling. New York: The Guilford Press.https://doi.org/10.1177/0002764217734273

23.

MINEDUC. (2017). Revisión de las políticas educativas en Chile desde 2004 a 2016. Reporte Nacional de Chile [Revision of education policies in Chile from 2004 to 2016. Chile National Report].Centro de Estudios Ministerio de Educación de Chile.

24.

T. W.

Eby

L. T.

Sorensen

K. L.

Feldman

D. C.

(2005). Predictors of objective and subjective career success: A meta‐analysis. Personnel Psychology, 58, 367–408. https://doi.org/10.1111/j.1744-6570.2005.00515.x

25.

T. W.

Feldman

D. C.

(2014). Subjective career success: A meta-analytic review. Journal of Vocational Behavior, 85(2), 169–179. https://doi.org/10.1016/j.jvb.2014.06.001

26.

OECD (2017). Education in Chile: Reviews of national policies for education. OECD Publishing.

27.

Paisey

Tarbert

(2020). Deprivation, social class and social mobility at big four and non-big four firms. Accounting and Business Research, 50, 61–109. https://doi.org/10.1080/00014788.2019.1647127

28.

Rajendran

E. S.

Sears

Ayub

(2020). Determinants of migrant career success: A study of recent skilled migrants in Australia. International Migration, 58, 30–51. https://doi.org/10.1111/imig.12586

29.

Schofer

Meyer

J. W.

(2005). The worldwide expansion of higher education in the twentieth century. American Sociological Review, 70(6), 898–920. https://doi.org/10.1177/000312240507000602

30.

UNESCO Institute for Statistics. (2020, July 9). Data for sustainable development. http://uis.unesco.org/

31.

Venegas-Muggli

J. I.

(2020). Secondary school context and (timely) college completion in chile’s education system: A cohort study. Journal of College Student Retention: Research, Theory & Practice. Advance online publication. https://doi.org/10.1177%2F1521025120910477

32.

(2012). Career success and its predictors: Comparing between Canadian and Chinese. International Journal of Business and Management, 7(14), 88. https://doi.org/10.5539/ijbm.v7n14p88