1 s2.0 S1747938X18302744 Main

Educational Research Review 29 (2020) 100305
Contents lists available at ScienceDirect
Educational Research Review

journal homepage: www.elsevier.com/locate/edurev
Socio-economic status and academic performance in higher

T
education: A systematic review☆
Carlos Felipe Rodríguez-Hernándeza,∗, Eduardo Cascallara, Eva Kyndtb,c
a
Center for Occupational & Organizational Psychology and Professional Learning, KU Leuven, Belgium, Dekenstraat 2, PB3772, 3000, Leuven,
Belgium
b
Department of Training and Education Sciences, University of Antwerp, Belgium, Sint-jacobstraat 2, 2000, Antwerp, Belgium
c
Centre for the New Workforce, Swinburne University of Technology, Melbourne, Australia
A R T IC LE I N F O ABS TRA CT
Keywords: Previous educational research has extensively investigated the relationship between socio-eco-
Socio-economic status nomic status (SES) and academic performance. In higher education, however, this relationship
Academic performance still deserves a comprehensive examination given both practical and conceptual reasons. To at-
Higher education tend to this need, a mixed-methods systematic literature review of 42 studies has been carried
Meta-analysis
out. In the first part, a summative content analysis examines how SES and academic performance
are measured. In the second part, a meta-analysis estimates the effect size of the relationship
between SES and academic performance in higher education. Findings suggest that SES is
measured through education, occupation, income, household resources, and neighborhood re-
sources, while academic performance in higher education is measured through achievement,
competencies, and persistence. Furthermore, the meta-analysis reveals a positive yet weak re-
lationship between SES and academic performance in higher education. Prior academic
achievement, university experience, and working status are more strongly related to academic
performance than SES.
1. Introduction
Over the past years, the student population applying to and entering university has become more diverse in terms of social,
cultural and economic capital, age, nationality (Morlaix & Suchaut, 2014), prior education, and academic achievement (Anderton,
Evans, & Chivers, 2016). Moreover, in many countries, social changes have also contributed to changes in higher education systems;
thus, although there is still a long way to go, participation from students from low social and economic backgrounds in higher
education is increasing (Hansen & Mastekaasa, 2006). In order to achieve a better understanding of these changes in higher education
systems, several researchers have explored the relationship between socio-economic status (SES) and academic performance, finding
a weak to moderate relationship (e.g., Richardson, Abraham, & Bond, 2012; Sackett et al., 2012; Westrick, Le, Robbins, Radunzel, &
Schmidt, 2015).
Given the increasing participation of students from low SES backgrounds and the weak to moderate relation between SES and
academic performance, one could wonder whether we still need to study the relationship between SES and academic performance in
☆
This research was supported in part by the Colombian Department of Science, Technology and Innovation (COLCIENCIAS) under the grant 756
of 2016.
∗
Corresponding author.
E-mail addresses: rodriguezcf@gmail.com (C.F. Rodríguez-Hernández), cascallar@msn.com (E. Cascallar), eva.kyndt@uantwerpen.be (E. Kyndt).
https://doi.org/10.1016/j.edurev.2019.100305
Received 11 June 2018; Received in revised form 17 September 2019; Accepted 20 November 2019
Available online 23 November 2019
1747-938X/ © 2019 Elsevier Ltd. All rights reserved.
C.F. Rodríguez-Hernández, et al. Educational Research Review 29 (2020) 100305
higher education. However, both practical and conceptual reasons exist that warrant continued attention. Firstly, the common trend
in the literature has been to use SES as a covariate instead of establishing, in a more comprehensive way, its influence on students'
experiences and outcomes (McKenzie & Schweitzer, 2001; Walpole, 2003). By doing the latter, the higher education sector can gain a
deeper understanding on how SES and academic performance are related, which might be necessary in order to deal appropriately
with the increased diversity in the student population. Secondly, recent meta-analyses in higher education (Richardson et al., 2012;
Schneider & Preckel, 2017; Westrick et al., 2015) have focused primarily on calculating the effect size of the relationship between SES
and academic performance, possibly ignoring the influence of several other students' and institutional characteristics on this re-
lationship. Thirdly, based on analytic schemes such as the Astin's input-environment-output model (Astin, 1993), it could be argued
that the relationship between SES and academic performance in higher education changes when additional variables are taken into
account. As such, it seems inadequate to explore students' academic performance as a single-factor phenomenon (De Clercq, Galand,
Dupont, & Frenay, 2013).
Although studies reviewing the relationship between SES and academic performance in primary and secondary schools are
available (e.g., Sirin, 2005), it cannot be assumed that these results are generalizable to the context of higher education. By carrying
out a mixed-methods systematic literature review, this study seeks to fill the gap that exists in the extensive literature devoted to
examining the relationship between SES and academic performance in higher education. This article begins by presenting several
issues pertaining to the definition of SES and academic performance, and then describes previous research in the literature regarding
their relationship. Next, the methodology of this study is outlined in detail, explaining how both summative content analysis and
meta-analysis are carried out. Subsequently, the answers to the research questions addressed in this research are presented. The
article concludes with the discussion and implications of the key findings of this systematic literature review.
2. Theoretical framework
2.1. Conceptualizing and measuring socio-economic status (SES)
The understanding of students’ socio-economic conditions became a major concern for educational researchers when low aca-
demic performance at school was observed in students whose parents had low income, low levels of education, and were employed at
low-status jobs (Cowan et al., 2012). Although SES can be considered as one of the most commonly used variables in educational
research (Sirin, 2005), it has been conceptualized in different ways in the literature. For instance, Chapin (as cited in White, 1982)
defined SES in 1928 as: “the position that an individual or family occupies with reference to the prevailing average of standards of
cultural possessions, effective income, material possessions, and participation in group activity in the community” (p. 99). Mueller
and Parcel (1981) defined SES as the position of an individual, family, or group on a hierarchy based on economic, power, and
prestige dimensions. More recently, SES has been defined as the amount of economic, social, and cultural resources available to one
student (Cowan et al., 2012; De Clercq, Galand, & Frenay, 2017).
The different dimensions of SES have been operationalized using either single indicators, multiple indicators analyzed separately,
or several indicators combined in a composite score (Australian Bureau of Statistics, 2011; Cowan et al., 2012; Shavers, 2007).
Moreover, the indicators of SES can be observed at several levels, namely, the individual, family, or area levels (Australian Bureau of
Statistics, 2011; Krieger, Williams, & Moss, 1997). At the individual level, education, occupation, and income have been used as
indicators for SES in previous educational research (Cowan et al., 2012; Sackett, Kuncel, Arneson, Cooper, & Waters, 2009; Van Ewijk
& Sleegers, 2010). Education, occupation, and income can consistently capture students' socio-economic conditions regardless of the
time in which they are observed (Erola, Jalonen, & Lehti, 2016). In addition, these measurements are easy to interpret and com-
municate (Cowan et al., 2012). At the family level, household resources have been suggested as the fourth indicator for SES (Sirin,
2005). Household resources refer to possessions such as cars, books, computers, and musical instruments (De Clercq et al., 2017;
Pedrosa, Dachs, Maia, Andrade, & Carvalho, 2007). Finally, at the area level, neighborhood resources have been reported as the fifth
indicator for SES (Australian Bureau of Statistics, 2011; Cowan et al., 2012; Shavers, 2007). Interestingly, financial and social
resources that do not come exclusively from the family can also be related to students’ academic performance (Cowan et al., 2012).
Such is the case, for example, with neighborhood characteristics and resources like the degree of urbanization (Hansen & Mastekaasa,
2006), and the number of parks and libraries in the area where students live (Cowan et al., 2012).
How education, occupation, income, and household resources interact in the measurement of SES, however, is rather complex. In
particular, education has been the most commonly used indicator to assess SES (Australian Bureau of Statistics, 2011; Shavers, 2007)
because of its relationship with other aspects of socio-economic status (Erola et al., 2016; Galobardes, Shaw, Lawlor, Lynch, & Smith,
2006). In fact, higher levels of education are related to the subsequent benefits they offer for a person's life and wellbeing, such as a
better job, working conditions, and higher-income (Shavers, 2007). Similarly, occupation is also commonly used as an indicator of
SES mainly due to its relationship with education and income (Erola et al., 2016; Ganzeboom, De Graaf, & Treiman, 1992). In this
case, income represents the amount of social and economic resources a student can have (Australian Bureau of Statistics, 2011;
Galobardes et al., 2006; Sirin, 2005), whereas household resources, therefore, can also indicate whether a student's home situation is
adequate for learning (Van Ewijk & Sleegers, 2010).
In summary, SES is a broad concept that encompasses two primary dimensions: prestige and resources (Krieger et al., 1997). The
first dimension determines the hierarchical position of an individual in a society (Mueller & Parcel, 1981), while the second di-
mension determines the economic, social, and cultural resources which an individual has access to (Cowan et al., 2012; De Clercq
et al., 2017). In addition, education, occupation, income, and household resources have been widely used as measurements to assess
SES.
2
2.2. Measuring academic performance in higher education
When searching for the definition of students' academic performance in higher education, the lack of consensus in the educational
literature is evident. Terms such as performance, achievement, and success are used interchangeably among educational researchers
without any specific reason (e.g., Casillas et al., 2012; Rochford, Connolly, & Drennan, 2009; Tracey, Allen, & Robbins, 2012). Also,
an operationalization (e.g., Grade Point Average) rather than a conceptual definition is mostly reported when defining academic
performance in higher education. Despite this lack of consensus, academic achievement, competencies, and persistence have been
used as separate, although interrelated, measurements to assess students' academic performance in higher education. Simply looking
at academic achievement does not necessarily encompass or represent students’ acquisition of competencies nor their persistence
(York, Gibson, & Rankin, 2015).
Academic achievement can be defined as the attainment of either medium- or long-term educational goals (Yusuf, 2002). In this
respect, Li, Chen, and Duanmu (2010) have pointed out that prior academic achievement is strongly related to students' academic
performance at university. As a matter of fact, a considerable number of studies have reported the explanatory role of prior academic
achievement in academic performance at university (e.g. Betts, Elder, Hartley, & Blurton, 2008; Byrne & Flood, 2008; Casillas et al.,
2012; McKenzie & Schweitzer, 2001; Pike & Saupe, 2002; Roberts, 2007). Furthermore, a competency is a “performance capacity to
do as well as to know which is judged by some level or standard of performance” (Shavelson, 2010, p. 44). In particular, higher
education aims at developing both specific and generic competencies (Sadler, 2013). Undoubtedly, a deeper understanding of aca-
demic performance in higher education requires the assessment of both generic and specific student competencies (Blömeke, Zlatkin-
Troitschanskaia, Kuhn, & Fege, 2013). Consequently, the assessment of competencies has arisen in many countries at different stages
of the higher education learning process (Zlatkin-Troitschanskaia, Shavelson, & Kuhn, 2015). Finally, Tinto's (1993) theory of de-
parture indicates that students persist when they are integrated into both the academic and social systems of the university. Per-
sistence can be understood as the students' academic progression towards degree completion regardless of institutional transfers,
academic programs, or institutional contexts (York et al., 2015). The dropout rate has usually been suggested as an indicator of
persistence in higher education (Hilton, 1982; Tinto, 1975, 1993).
2.3. The relationship between SES and academic performance
One of the most crucial turning points in educational research during the 20th century was the publication of a 1966 report by
Coleman and colleagues entitled Equality of Educational Opportunity (EEO). The report suggested that high school characteristics
were unrelated to academic performance in the USA, but students' socio-economic conditions were. Consequently, many educational
researchers have carried out studies aimed at understanding Coleman's main findings.
On the one hand, several authors have postulated theoretical frameworks, such as Astin's (1984, 1999) student involvement
theory or Bourdieu's (1986) social capital theory, which could explain how SES and academic performance are related. Regarding
social capital theory (Bourdieu, 1986), Dika and Singh (2002) critically reviewed the literature (published between 1986 and 2001)
that relates social capital to educational outcomes and identified several problems with the conceptualization and measurement of
social capital as a predictor of academic performance. In particular, sources of social capital are often confused with the resources and
opportunities coming from it; thus, there is no clear distinction between possession and activation of social capital (Dika & Singh,
2002). Additionally, the selection of cross-sectional data has made it difficult to determine how social resources and educational
outcomes are related (Dika & Singh, 2002). Therefore, it cannot be entirely accepted that social capital explains how SES is related to
academic performance (Jæger, 2011; Sullivan, 2001).
On the other hand, several meta-analytic studies have been conducted in different educational settings. Focusing on the ele-
mentary and secondary levels, the main objectives of these meta-analyses (e.g., Sirin, 2005; Van Ewijk & Sleegers, 2010; White, 1982)
have been: (1) to determine the effect size of the relationship between SES and academic performance; and (2) to identify which
factors could moderate the relationship between SES and academic performance. Regarding the first objective, the findings show a
positive, albeit moderate, relationship, as indicated by average correlations of .343 (S.D. = .204; White, 1982), .299 (S.D. = .169;
Sirin, 2005) and .32 (S.E. = .016; Van Ewijk & Sleegers, 2010). With respect to the second objective, the results suggest that
methodological factors such as the unit of analysis (Sirin, 2005; White, 1982), the definition of SES (White, 1982), the source of the
SES data, the range of the SES, and the type of SES-performance measure (Sirin, 2005), moderate the relationship between SES and
academic performance.
Delving into higher education, one of the objectives of these meta-analyses (e.g., Richardson et al., 2012; Schneider & Preckel,
2017; Westrick et al., 2015) has been to explore the effect size of the relationship between SES and academic performance.
Richardson et al. (2012) found a small correlation between SES and academic performance (r = .11, 95% CI [.08, .15]). Similarly,
Westrick et al. (2015) reported that SES is weakly related to first-year GPA (r = .24, 95% CI [.24, .25]) and second-year retention at
university (r = .10, 95% CI [.09, .11]). Finally, SES was ranked in the 68th place among 105 variables associated with academic
performance in higher education (Schneider & Preckel, 2017). A concluding remark of this body of research is that there is a weak to
moderate relationship between SES and academic performance in higher education. However, a more comprehensive exploration of
this relationship is still missing in the educational literature.
2.4. Mediators of the relationship between SES and academic performance
Several theories have been proposed to explain how students grow and change during their university studies (Long, 2012). A
3
well-known example within this body of literature is Astin's (1984, 1999) theory of involvement. Students' involvement refers to the
extent to which a student invests energy into his or her university experience —the greater the student's involvement, the greater the
students' learning and development (Astin, 1984, 1999). One analytic scheme drawn from Astin's theory of involvement is the input-
environment-output (IEO) model (Astin, 1993). The I-E-O model suggests that students' educational outcomes are defined by the
students' inputs (e.g., their demographic characteristics, prior academic achievement), the environmental elements (e.g., university
organizations, peer relations), and the interaction among students' inputs and environmental elements (Astin, 1993). Most of the
predictive studies in higher education lack of a theoretical foundation, being more empirically-based than theory-driven. Never-
theless, the I-E-O model seems to be a suitable approach to analyze the literature in higher education regarding the prediction of
academic performance for two reasons. First, the I-E-O model allows for the investigation of the direct and indirect influence (via
environmental characteristics) of students' inputs on students' educational outcomes. Second, the I-E-O model recognizes the long-
itudinal nature of the student retention process and provides a framework for the investigation of it (Kelly, 1996).
Starting from the I-E-O model, predictive studies in higher education can be classified into two types. The first type of predictive
studies, which is most likely the most frequent one in higher education, has examined academic performance as an “input-output”
process, a common approach in the educational field. A review of the literature on the input-output analysis of the schools was
carried out by Glasman and Biniaminov (1981). The authors concluded that inputs can be categorized into either student or school
type, while the outputs can be classified as either cognitive or non-cognitive. Furthermore, a causal model including both the direct
and indirect effects of inputs on outputs was also proposed by Glasman and Biniaminov (1981). The second type of predictive studies
in higher education has investigated academic performance as an “input-environment-output” process. A similar analytic scheme at
the school level is the context, input, process, and output model proposed by Scheerens (1990). In brief, context here refers to the
school environment as well as the policy measures at a higher administrative level, while input relates to the available resources,
teacher qualifications, and student characteristics. Process includes curriculum, school organization, and school climate, while output
is generally defined in terms of students’ achievement.
Furthermore, the I-E-O model can frame the analysis of the mediators of the relationship between SES and academic performance
in higher education. First, several “input-output” studies in higher education (e.g., Crawford, 2014; Stratton & Wetzel, 2011;
Warburton, Bugarin, & Nuñez, 2001) have shown that (1) prior academic achievement is strongly related to academic performance in
university and (2) prior academic achievement might diminish the strength of the SES-academic performance relationship. Although
it has been amply documented that SES determines prior academic achievement, Marks (2017) has suggested that the influence of
prior academic achievement on students’ outcomes is not solely explained by their SES at previous stages of life. Therefore, the first
mediator investigated in this systematic literature review is prior academic achievement.
Second, the “input-process-output” studies in higher education have suggested that university experience (Gerken & Volkwein,
2000; Smith, 2016; Walpole, 2003) does influence the relationship between SES and academic performance. University experience
refers to how a student connects to the academic environment of the university (Astin, 1999). More specifically, perception of the
learning environment, peer support, and institutional commitment could define university experience (Astin, 1999; De Clercq et al.,
2013). Thus, the second mediator investigated in this systematic literature review is university experience.
Third, although Astin (1999) suggested that holding a part-time job on campus could have a beneficial influence on students'
retention, the explanatory role of working status has usually been investigated separately from the university experience. In this
respect, the general trend in the literature has been to analyze the influence of worked hours on students’ academic performance (e.g.,
Nonis & Hudson, 2006; Rochford et al., 2009; Stinebrickner & Stinebrickner, 2003). Nevertheless, the reasons why students decide to
work when attending university are diverse; so, it can be argued that not only would low SES students work during their studies
(Yanbarisova, 2015). Hence, the third mediator investigated in this systematic literature review is working status.
3. Present study
Past educational research has contributed to the understanding of the relationship between SES and academic performance.
However, recent meta-analyses in higher education (Richardson et al., 2012; Schneider & Preckel, 2017; Westrick et al., 2015) are
limited, for two main reasons. Firstly, both SES and academic performance had been operationalized in these studies using only one
indicator, which narrowed the understanding regarding how additional measures to assess such complex terms are related. Secondly,
these meta-analyses did not explore how SES is first related to several student characteristics and, subsequently, is related to aca-
demic performance. This systematic review focuses on investigating the relationship between socio-economic status and academic
performance in higher education in a more comprehensive manner. The first objective of this study is to analyze the different
measures of SES and academic performance in higher education. Thus, the first research question addressed in this study is: How are
SES and academic performance in higher education measured? The second objective of this study is to determine the mediating role of
several factors on the relationship between SES and academic performance in higher education. Therefore, the second pair of research
questions of this study are: (a) What is the relationship between SES and academic performance in higher education? And (b) Is the
relationship between SES and academic performance in higher education mediated by (i) prior academic achievement, (ii) university ex-
perience, and/or (iii) working status? To answer these research questions, a mixed-methods research synthesis (Heyvaert, Maes, &
Onghena, 2013) of the selected studies has been carried out. A mixed-methods approach makes it possible to integrate both quali-
tative analyses (e.g., for this study, summative content analysis) and quantitative analyses (e.g., for this study, meta-analysis) of the
results of the studies in order to obtain conclusions about the current state of the art of the literature (Heyvaert et al., 2013).
4
Table 1
Overview of the literature search hits.
Query Search terms ERIC (Ovid) PsycArticles Scopus SSCI
1 ‘Higher Education’ AND ‘Academic Achievement’ 13,668 429 988 345 15,430
Update May 2017 333 17 110 80 540
2 ‘Higher Education’ AND ‘Academic Performance’ 9535 287 752 553 11,127
Update May 2017 264 9 186 145 604
3 ‘Higher Education’ AND ‘Academic Outcomes’ 0 1708 83 66 1857
Update May 2017 395a 78 20 15 508
Sub-total initial search 23,203 2424 1823 964 28,414
Sub-total update May 2017 992 104 316 240 1652
Total 24,195 2528 2139 1204 30,066
a
This number of hits was obtained using the multi-field function and selecting references between 2000 and 2017.
4. Method
This systematic review proceeded in three phases. In the first phase, an extensive literature search within several scientific
databases was conducted. Next, the relevant literature retrieved from these databases was selected for inclusion according to several
criteria. In the second phase, the quality of the selected studies was critically appraised, and finally, the primary studies were
analyzed following the guidelines of Aveyard (2014) and the performance of both a summative content analysis and meta-analysis.
4.1. Literature search and literature selection
In this systematic review, four databases were consulted: ERIC, Scopus, SSCI, and PsycArticles, using combinations of the fol-
lowing search terms: “higher education”, “academic performance”, “academic achievement”, and “academic outcomes”. The initial
search yielded 28,414 non-unique studies, as shown in Table 1. The selection of the literature was based on eight criteria for
inclusion. As reported by Gamoran and Long (2007), the citations of Coleman's report increased again by the end of the ‘90s and
achieved an average of 55 citations per year from 2000, which represents a new interest for studying academic performance since
then. Thus, studies published after 2000 (Criterion 1) were considered for subsequent analysis. Second, only empirical studies ex-
ploring academic performance in higher education were included in this analysis (Criterion 2). Studies focusing on specific subgroups
of students (i.e., students with disabilities; online learners) were excluded (Criterion 3). As this systematic literature review focused
on variables related to SES, studies concerning learning styles were not included (Criterion 4). Similarly, studies regarding age
differences (Criterion 5), gender differences (Criterion 6), and ethnic differences (Criterion 7) were not included in this analysis.
Finally, only studies that explicitly report the relationship between SES and academic performance were selected for further analysis
(Criterion 8).
The selection process of the initial 28,414 studies was carried out in six steps. In step 1, duplicate studies were eliminated using
the EndNote software, leaving 24,246 studies. In step 2, studies published before 2000 were excluded, leaving 9928 studies. Next, in
step 3, studies without a date, a wrong date, or those that were not written in English or Spanish were eliminated, leaving 9658
studies. As expected, a summative content analysis requires a full understanding of the reviewed studies. Therefore, in step 4, the title
and the abstract of the remaining studies were screened using the first seven criteria for inclusion presented above. As a result of this
fourth step, 208 studies were retained for further analysis. In step 5, a total of 202 studies were scanned/read diagonally as six full
texts articles could not be retrieved. This fifth step, using the same inclusion criteria, led to a further reduction in the number of
studies to a total of 100. In the final step, the full texts of these 100 studies were read in-depth, and 69 additional studies were
excluded using Criterion 8 presented above. As a result of the selection process, 31 studies remained. As any systematic review is a
challenging and time-consuming task, 18 months passed between the first literature search and the writing of the first draft of this
article. Hence, an update of the literature search was carried out on May 2017. In total, 1652 new hits were retrieved, and after using
the previously mentioned selection criteria, six additional studies were selected. Finally, a back-tracing process of the 37 resulting
studies was performed to identify and select additional relevant studies. The outcome of this process was the selection of ten further
studies, and as a result of the selection process, a total of 47 primary studies were considered for critical appraisal.
4.2. Critical appraisal
The quality of the selected primary studies (n = 47) was evaluated using the checklists of the National Institute for Health and
Clinical Excellence (2009). The main criteria for the quality appraisal were (a) a clear statement of the aims of the research, (b) an
appropriate research design, (c) a well-described and appropriate sampling strategy, data collection, and analysis method; and (d) a
clear description of the research findings. Each study was rated as either high, medium, or low quality. Following this critical
appraisal, five studies were excluded due to their low quality, leaving 42 studies for analysis. Appendix A shows the results of the
critical appraisal process.
5
Fig. 1. Conceptual diagram for the analysis of the mediators.
4.3. Analysis of the literature
The analysis of the literature was conducted under the guidelines proposed by Aveyard (2014). First, the main characteristics of
the selected studies were summarized (see Appendix B). Then, every study was reread to identify and codify the relevant information.
Consequently, all the studies were coded and assigned the following themes: operationalizations of SES (Theme 1); oper-
ationalizations of academic performance (Theme 2); relationship between SES and academic performance (Theme 3); prior academic
achievement (Theme 4), university experience (Theme 5), and working status (Theme 6). The first research question was answered by
summative content analysis. The second research questions were answered using both summative content analysis and meta-analysis.
4.3.1. Summative content analysis

A summative content analysis goes beyond quantifying the occurrence of words and content within texts; it aims at analyzing and
interpreting that content (Hsieh & Shannon, 2005). To address the first research question, a summative content analysis of the first
two themes listed above was carried out. Therefore, the operationalizations of SES (Theme 1) were categorized as parental educa-
tional level, parental occupation, income, household resources, or neighborhood resources. Similarly, the operationalizations of
academic performance (Theme 2) were categorized as achievement, competencies, or persistence.
To answer the first part of the second research question, the relationship between SES and academic performance in higher
education (Theme 3) was classified into three categories: significant and positive; significant and negative; and not significant. Those
categories were chosen as they are the three possible basic ways in which two variables are correlated. Subsequently, to answer the
second part of the second research question, the mediating roles of prior academic achievement (Theme 4), university experience
(Theme 5), and working status (Theme 6) were explored following the rationale that will be explained subsequently. The interest was
to determine the effect size and the significance level of the relationships indicated with numbers in Fig. 1. That is, the relationship
between SES and academic performance (1), once the mediator was also introduced into the same explanatory model; and, next, the
relationship between the mediator and academic performance (2). As such, a larger and significant relationship between the mediator
and academic performance would suggest mediation. It is worth noting, however, that not all the analyzed references provided
information on the relationship between the SES and the mediator (3). Consequently, this third relationship could not be included in
the analysis, so it was not possible to distinguish between complete and partial mediation.
4.3.2. Statistical meta-analysis

Hunter and Schmidt (1990) define a meta-analysis as the quantitative summary and analysis of different effect sizes retrieved
from several various studies. To further extend the answer to the second research question, a meta-analysis was carried out to
calculate the average effect size of (a) the relationship between SES and academic performance and (b) the relationship between the
investigated mediators and academic performance.
Metric from expressing effect sizes. The effect size (ES) selected for this study was Pearson's correlation coefficient (r). Most of the
included studies reported a standardized regression coefficient, which was converted into an r-value using the guidelines proposed by
Peterson and Brown (2005). Although alternative methods for transforming standardized regression coefficients into either partial or
semi-partial correlations exist (Aloe & Becker, 2012; Aloe & Thompson, 2013; Fernández-Castilla et al., 2019), there was not enough
information in the primary studies to carry out such conversions. In four studies (Birch & Miller, 2006; Delaney, Harmon, &
Redmond, 2011; Guimarães & Sampaio, 2013; Win & Miller, 2005) the unstandardized regression coefficient was first converted into
a standardized coefficient using the standard deviation of both the predicted and the predictor variable as suggested by Bowman
(2012). In six studies (Arulampalam, Naylor, & Smith, 2004; Hansen & Mastekaasa, 2006; Smith, 2016; Walpole, 2003; Yanbarisova,
2015; Yao, Zhimin, & Peng, 2015) the reported odds ratio were converted into an r-value following the procedure proposed by
Borenstein, Hedges, Higgins, and Rothstein (2009). In one study (Waqas, Abbasi, & Idrees, 2013) the reported t-value was first
converted into a standard mean difference (d-value) and, then, this d-value was converted into an r-value, also according to the
procedure suggested by Borenstein et al. (2009).
In total, 13 studies were left out of the meta-analysis as will be explained next. The standardized regression coefficient in four
6
articles (Loehr, Almarode, Tai, & Sadler, 2012; Morlaix & Suchaut, 2014; Pedrosa et al., 2007; Rodríguez Albor, Ariza, & Ramos,
2014) could not be calculated due to insufficient information. It is essential to notice that the inclusion of unstandardized regression
coefficients could dramatically increase the average effect size. Similarly, in two of the analyzed studies (Frischenschlager, Haidinger,
& Mitteraurer, 2005; Stratton & Wetzel, 2011), it was not possible to convert the reported odds ratio due to a lack of information. In
five of the articles (Anderton et al., 2016; Gerken & Volkwein, 2000; Ifenthaler & Widanapathirana, 2014; Nguyen, 2016; Zheng,
Saunders, Shelley, Mack, & Whalen, 2002) the effect size was not reported due to it not being significant. Finally, in two of the articles
(Bahamón & Reyes Ruiz, 2014; Triventi, 2014), the relationship between SES and academic performance was described but not
reported through a specific effect size.
Statistical independence. One crucial methodological aspect when performing a meta-analysis is to ensure that the effect sizes are
independent among them. To achieve such independence, the average effect size was calculated in studies reporting more than one
effect size. In this way, the sample on which it was based contributed only with one effect size to the analysis (Sirin, 2005). One
article (De Clercq et al., 2013) reported two different effect sizes coming from two different data sets. Both effect sizes were included
in the meta-analysis.
Combining effect sizes across studies. Another major important consideration when conducting a meta-analysis is to decide how to
transform the different effect sizes retrieved from the examined studies so that a meaningful and valid aggregation can be made. In
this meta-analysis, the effect sizes were converted using Fisher's transformation as recommended by Hedges and Olkin (1985).
Fisher's transformation is a variance stabilizing transformation so that the r-value becomes independent from the population p-value
(Aloe & Becker, 2012; Bowman, 2012).
Homogeneity analysis. The homogeneity among the effect sizes was analyzed using Hedge's Q test for homogeneity. This test is
based on chi-square statistics with k-1 degrees of freedom, where k is the number of investigated effect sizes. A significant result
suggests that effect sizes across the studies are heterogeneous, so further exploration of the existence of possible mediators should be
conducted.
Publication bias. A common fact in scientific literature is publication bias. Publication bias means that only statistically significant
results or results which support the expected relationship are published (Rothstein, Sutton, & Borenstein, 2005). In this study,
publication bias was assessed using the funnel plot. To evaluate the funnel plot symmetry, Egger's regression test (Egger, Smith,
Schneider, & Minder, 1997) was performed, and the results indicated that there was no publication bias in the studies summarized
through this meta-analysis (z = .79, p = .43).
Analysis of the mediators. Similar to the summative content analysis, the existence of mediators was analyzed by contrasting the
effect size of the relationships indicated with numbers in Fig. 1. The average effect size between SES and academic performance (1)
was compared to the average effect size between the presumed mediator and academic performance (2). As such, a larger and
significant average effect size between the mediator and academic performance would suggest mediation.
5. Results
The results of this systematic literature review are presented in accordance with the research questions. Sections 5.1 and 5.2 are
based on the results derived from the summative content analysis, while Section 5.3 is based on the results of both the summative
content analysis and the meta-analysis.
5.1. Measuring SES
Table 2 presents the classification of the operationalizations of SES into five major measurements: parental educational level,
parental occupation, income, household resources, and neighborhood resources. Table 2 also provides information regarding the type
of measure (i.e., single or composite score) and the type of scale (i.e., categorical or continuous) of each one of the examined
operationalizations. It is important to note that a study can appear more than once in Table 2, as it might have reported several
indicators for the assessment of SES. The classification of the operationalizations of SES is described in detail in the following
sections.
5.1.1. Parental educational level

This category (n = 25) comprises operationalizations that assessed parental educational level as a single indicator (n = 21) or
within a composite score (n = 4). With regard to “single indicator” (n = 21), the use of categorical scales (n = 19) was identified to
assess parental educational level in three possible ways: (a) the highest level of parental education (n = 4); (b) a dichotomous
variable which indicated whether the parents had attended college (n = 3); and using several categories—ranging from no level of
education to a university degree —to measure (c) both parents' level of education (n = 9); (d) only the mother's level (n = 2); and (e)
only the father's level. Additionally, continuous scales (n = 2) were utilized to measure the length of the parents' education in two
different ways. Firstly, Rothstein (2004) used the average number of years of education of the students' parents. Secondly, Delaney
et al. (2011) converted the qualifications reported by the parents into years of education by estimating the number of years which are
required to obtain those degrees. Consequently, this variable ranges from 8 (time necessary to complete primary school) to 19 (time
necessary to complete a Ph.D.).
Regarding “within a composite score” (n = 4), parental educational level was assessed through both categorical (n = 3) and
continuous scales. On the categorical scales (n = 3), De Clercq et al. (2017), and Rodríguez Albor et al. (2014), reported the use of the
highest level of parental education (n = 2). In addition, Gouvias, Katsis, and Limakopoulou (2012) measured parental education
7
Table 2
Operationalizations of SES.
Measurement Type of measurement Scale Operationalization Reference(s)
Parental Educational level As a single indicator Categorical (n = 19) Highest level of parental education (n = 4) Bruinsma and Jansen (2007); Ifenthaler and Widanapathirana (2014);
(n = 25) (n = 21) Loehr et al. (2012); Tai, Sadler, and Loehr (2005)
A dichotomous variable to determine whether parents Bonsaksen (2016); Nguyen (2016); Waqas et al. (2013)
have attended college (n = 3)
3 categories Rodríguez Ayan and Ruiz Díaz(2011)
C.F. Rodríguez-Hernández, et al.
4 categories Stratton and Wetzel (2011)

4 categories De Clercq et al. (2013)
4 categories (years of education) Triventi (2014)
5 categories Beyene and Yimam (2016)
5 categories Frischenschlager, Haidinger, and Mitterauer (2005)
5 categories Gerken and Volkwein (2000)
5 categories Harb and El-Shaarawi (2007)
7 categories Guimarães and Sampaio (2013)
Only father's educational level (8 categories) Morlaix and Suchaut (2014)
Only mother's educational level (3 categories) Wolniak and Engberg (2010)
Only mother's educational level (5 categories) Black et al. (2015)
Continuous (n = 2) Years of education (average) Rothstein (2004)
Years of education: going from 8 to 19 Delaney et al. (2011)
Within a composite score Categorical (n = 3) Highest level of parental education (n = 2) De Clercq et al. (2017); Rodríguez Albor et al.(2014)
(n = 4) ISCED (International Standard Classification of Gouvias et al. (2012)
Education)
Continuous Occupational prestige and socioeconomic scores (Nakao Walpole (2003)
8
& Treas, 1994).
Parental Occupation (n = 10) As a single indicator Categorical (n = 7) Dichotomous variable to indicate whether parents work Guimarães and Sampaio (2013)
(n = 7) Dichotomous variable to indicate whether parents' Bonsaksen (2016)
occupation(s) requires higher education
4 categories Yao et al. (2015)
5 categories Arulampalam et al. (2004)
8 categories Morlaix and Suchaut (2014)
10 categories Hansen and Mastekaasa (2006)
The Standard Occupational Classification scale 2000 Smith (2016)
Within a composite score Categorical (n = 2) 7 categories Gouvias et al. (2012)
(n = 3) 12 categories Rodríguez Albor et al. (2014)
Continuous Occupational prestige and socioeconomic scores (Nakao Walpole (2003)
& Treas, 1994).
(continued on next page)
Table 2 (continued)
Measurement Type of measurement Scale Operationalization Reference(s)
Income (n = 20) As a single indicator Categorical (n = 16) 3 categories (high, middle and low) (n = 3) Frischenschlager et al. (2005); Wolniak and Engberg (2010); Yao et al.
(n = 17) (2015)
5 categories Black et al. (2015)
5 categories (quintiles) Bozick (2007)
6 categories Yanbarisova (2015)

7 categories Guimarães and Sampaio (2013)
7 categories Waqas et al. (2013)
Financial aid (receiving a scholarship or a loan) (n = 4) Black et al. (2015); Morlaix and Suchaut (2014); Stratton and Wetzel
(2011); Triventi (2014)
To receive subsidized lunches (n = 2) Black et al. (2015); Rothstein (2004)
University tuition fees (quartiles) Triventi (2014)
Continuous Parents' relative income going from 0 to 1 Hansen and Mastekaasa (2006)
Within a composite score Continuous (n = 3) Monthly income Rodríguez Albor et al.(2014)
(n = 3) Occupational prestige and socioeconomic scores (Nakao Walpole (2003)
& Treas, 1994)
University tuition fees Rodríguez Albor et al. (2014)
Household resources (n = 8) As a single indicator Categorical (n = 4) Computer at home and internet connection (n = 3) Gouvias et al. (2012); Guimarães and Sampaio (2013); Pedrosa et al.
(n = 6) (2007)
Possession of books related to schoolwork Gouvias et al. (2012)
Continuous (n = 2) Household crowding (n = 2) Harb and El-Shaarawi (2007); Gouvias et al. (2012)
Within a composite score Continuous (n = 2) Number of possessions (car, books, musical instruments, De Clercq et al. (2017), Pedrosa et al. (2007)
(n = 2) computer) (n = 2)
9
Neighborhood resources (n = 9) As a single indicator Categorical (n = 3) Students' stratum (n = 2) Bahamón and Reyes Ruiz (2014); Rodríguez Albor et al. (2014)
(n = 3) Degree of urbanization Hansen and Mastekaasa (2006)
As a composite score Continuous (n = 6) Index of Economic Resources (IER) (n = 2) Birch and Miller (2006); Win and Miller (2005)
(n = 6) Index of Education and Occupation (IEO) Win and Miller (2005)
Index of Multiple Deprivation (IMD) Thiele et al. (2016)
Index of Relative Socioeconomic Advantage and Puddey and Mercer (2014)
Disadvantage (IRSAD)
School decile Shulruf et al. (2008)
using the International Standard Classification of Education (ISCED). The ISCED is a classification of educational programs based on
levels and areas of knowledge proposed in 1997 by the United Nations Educational, Scientific, and Cultural Organization (UNESCO),
and then revised in 2011. The main goal behind this classification is the need to compare internationally different educational
systems (Schneider, 2013). More specifically, there are nine ISCED levels ranging from early childhood education to a doctoral or
equivalent level. On the continuous scales, Walpole (2003) recoded the parental educational level into a continuous SES variable
based on the occupational prestige and socioeconomic scores as proposed by Nakao and Treas (1994). In particular, Nakao and Treas
(1994) suggested the following regression equation to calculate a socio-economic index (SEI): SEI = 9.24 + 0.64 (Education) + 0.31
(Income).
5.1.2. Parental occupation

This category (n = 10) includes operationalizations of parental occupation as a single indicator (n = 7) or within a composite
score (n = 3). Regarding “single indicator” (n = 7), the use of categorical scales (n = 7) was identified to assess parental occupation in
three possible ways: (a) as a dichotomous variable to indicate whether the parents worked; (b) as a dichotomous variable to indicate
whether the parents’ occupation(s) required a higher educational degree; (c) as several scales ranging from unskilled worker/un-
employed to managerial/professional (n = 4). In addition, Smith (2016) reported the use of the 2000 Standard Occupational Clas-
sification (SOC). The 2000 SOC is a classification system of the occupational structure of the United States, proposed by The Bureau of
Labor Statistics. The 2000 SOC allows for comparisons to be made among paid occupations, based on the type of work, required skills,
education, and training. More specifically, the 2000 SOC includes occupations from the public, private, and military sectors (Bureau
of Labor Statistics, 2006).
Regarding “within a composite score” (n = 3), Rodríguez Albor et al. (2014) used nominal categories which described parental
occupation. In addition, Gouvias et al. (2012) reported the use of seven categories which ranged from unemployed to professionals,
managers, and business owners. Finally, Walpole (2003) utilized the occupational prestige and socioeconomic scores, proposed by
Nakao and Treas (1994), to recode parental occupation into a continuous SES variable.
5.1.3. Income
This category (n = 20) sorts between operationalizations that measured income as a single indicator (n = 17) or within a com-
posite score (n = 3). Regarding “single indicator” (n = 17), the use of categorical scales (n = 16) was identified to measure income
with: (a) several categories to discriminate between low and high income (n = 9); (b) dichotomous scales to determine whether
students received financial aid (n = 4) or received subsidized lunches (n = 2); and (c) quartiles to classify university tuition fees.
Additionally, Hansen and Mastekaasa (2006) proposed a continuous scale between 0 and 1 to determine the relative position of a
student's family within the distribution of all family incomes (i.e., a score of 0.52 indicates that 52% of the families have less income).
Concerning “within a composite score” (n = 3), continuous scales were used by Rodríguez Albor et al. (2014) to combine tuition
fees and monthly income into a composite score. Similar to parental educational level and occupation, parental income was also re-
coded by Walpole (2003) into a continuous SES variable, based on the occupational prestige and socioeconomic scores proposed by
Nakao and Treas (1994).
5.1.4. Household resources

This category (n = 8) contains operationalizations of household resources as a single indicator (n = 6) or within a composite score
(n = 2). Regarding “single indicator” (n = 6), categorical scales (n = 4) were selected to determine whether students have a computer
and internet connection (n = 3); and books related to schoolwork. In addition, continuous scales (n = 2) were used to calculate
household crowding. Harb and El-Shaarawi (2007) defined household crowding as the proportion of the number of family members
to the number of rooms in the house. Concerning “within a composite score” (n = 2), continuous scales were used to include the
number of possessions as an indicator of household resources.
5.1.5. Neighborhood resources

This category (n = 9) presents operationalizations of socio-economic conditions at the area level. The use of both a single in-
dicator (n = 3) and a composite score (n = 6) was identified in the examined primary studies. Among “single indicator” (n = 3),
categorical scales (n = 3) were used to assess the students' stratum (n = 2) and degree of urbanization. In particular, the students’
stratum refers to the six categories used by the Colombian government to classify households based on their physical characteristics
and surroundings. The main reason behind this classification is to establish the price of public services hierarchically in each area
(The World Bank, 2012).
Regarding “composite score” (n = 6), the connection of students' postal codes to several indices aimed at classifying areas based on
socio-economic advantages and disadvantages was reported in the analyzed primary studies, with these indices being the Socio-
Economic Indices for Areas (SEIFA) first developed by the Australian Bureau of Statistics. In particular, the Index of Economic
Resources (Birch & Miller, 2006; Win & Miller, 2005); the Index of Education and Occupation (Win & Miller, 2005); and the Index of
Relative Socioeconomic Advantage and Disadvantage (Puddey & Mercer, 2014) were identified in the examined references. Ad-
ditionally, Thiele, Singleton, Pope, and Stanistreet (2016) used indices suggested by the Higher Education Funding Council for
England (HEFCE). More specifically, the Index of Multiple Deprivation (IMD) was utilized to rank area deprivation in five quintiles
(where the first quintile included the most deprived areas, and the fifth quintile included the least deprived areas). Finally, Shulruf,
Hattie, and Tumen (2008) reported the use of the school decile, which is a system from New Zealand aimed at classifying schools into
ten categories according to their students’ socio-economic conditions. Such a classification makes it possible to determine, for
10
Fig. 2. Authors' model for measuring SES.
instance, whether schools are receiving students coming from low-SES areas.
5.1.6. Model for measuring SES

Fig. 2 shows the model to assess SES resulting from the summative content analysis carried out in section 5.1. Education, oc-
cupation, income, household resources, and neighborhood resources are the five major measurements for assessing SES. Moreover,
such measurements can be assessed at different possible levels, namely, the individual, family, or area levels.
5.2. Measuring academic performance in higher education
Table 3a and Table 3b present the classification of the operationalizations to assess academic performance in higher education
into three measurements: achievement, competencies, and persistence, both before and during university. Tables 3a and 3b also
display information regarding the level of measurement (i.e., categorical or continuous) as well as the scale used in each of the
examined references. Again, the studies can appear more than once in Tables 3a and 3b, as they might have used several indicators to
assess academic performance. The classification of the operationalizations of academic performance in higher education is explained
in the upcoming sections.
5.2.1. Academic performance at university

This category includes operationalizations which assessed academic performance at university (n = 55). Operationalizations
conforming to this category were grouped into three subcategories: academic achievement (n = 43); competencies (n = 6), and
persistence (n = 6). Regarding the subcategory academic achievement (n = 43), the operationalizations identified in the analyzed
studies were Grade Point Average (GPA; n = 32); single grades (n = 4); number of credits (n = 6) and the relative achievement index.
More specifically, three different types of GPA were identified in the examined references: the average at the end of the first year
(n = 16), the current average when the study was conducted (n = 8), and the final average upon graduation (n = 8). Although both
categorical (n = 8) and continuous (n = 24) scales were used to assess GPA, the most frequently used scales for measuring GPA were
continuous from 0 to 4 (n = 15) and from 0 to 100 (n = 5). In addition, the use of a single grade was identified as the grades in
subject areas (n = 3) and the grades on a first-year test. In the case of the number of credits (n = 6), Rodríguez Ayan and Ruiz Díaz
(2011) proposed an original indicator —the degree progression index—which is the relationship between the actual number of
credits, and the expected number of credits that a student must take. Moreover, Triventi (2014) suggested that ECTS credits are a
concise way to measure students' academic achievement because they represent not only how many exams a student has completed
satisfactorily but also their importance. Finally, the relative achievement index was proposed by Pedrosa et al. (2007) to compare
students’ performance at the entrance and exit of all courses. Differences in both the grading system and the number of students
enrolled in the courses were the main reasons for developing this index, instead of using the actual numerical value of the grades.
In the subcategory competencies (n = 6), Bahamón and Reyes Ruiz (2014) and Rodríguez Albor et al. (2014) used the SABER PRO
test, which is a standardized test taken by the students at the end of the university level in Colombia to determine both their general
competencies (e.g., quantitative reasoning, critical reading) and specific competencies within each field of study. Similarly, students'
proficiency in English was assessed through several standardized tests such as the College English Test (CET) and the internal Post-
Entrance Literacy Assessment (PELA). In addition, Morlaix and Suchaut (2014) reported the use of the Diplôme Approfondi de Langue
Française (DALF) test to evaluate the first-year students’ written comprehension of French. Finally, Puddey and Mercer (2014) re-
ported the use of the Graduate Medical School Assessment Test (GAMSAT) to assess the level of preparation for undertaking studies in
medicine and/or attending medical school at the graduate level at Australian, British, and Irish universities.
Regarding the subcategory persistence (n = 6), all the analyzed operationalizations were related to completion status. As such, the
operationalizations grouped in this subcategory were dropout (n = 2), final degree classification (n = 2), attending graduate school,
and graduation rate. In particular, the UK undergraduate degree classification was sorted into this subcategory.
11
Table 3a
Operationalizations of academic performance at university.
Measurement Operationalization Type of scale Scale Reference(s)
Achievement (n = 43) GPA (n = 32) First year GPA (n = 16) Categorical (n = 4) 3 categories (high, medium, Shulruf et al. (2008); Yao et al. (2015)
low) (n = 2)
4 categories De Clercq et al. (2017)

5 categories Hansen and Mastekaasa (2006)
Continuous From 0 to 4 (n = 6) Anderton et al. (2016); Black et al. (2015); Rothstein (2004); Sackett
(n = 12) et al. (2009); Wolniak and Engberg (2010); Zheng et al. (2002)
From 1 to 10 Bruinsma and Jansen (2007)
From 0 to 20 Morlaix and Suchaut (2014)
From 0 to 100 (n = 4) Birch and Miller (2006); De Clercq et al. (2013); De Clercq et al.
(2017); Win and Miller (2005)
GPA (n = 8) Categorical (n = 2) 3 categories Yao et al. (2015)
3 categories Yanbarisova (2015)
Continuous (n = 6) From 0 to 4 (n = 5) Black et al. (2015); Bonsaksen (2016); Harb and El-Shaarawi (2007);
Sackett et al. (2009); Wang et al. (2010)
From 0 to 12 Rodríguez Ayan and Ruiz Díaz(2011)
Final GPA (n = 8) Categorical (n = 2) 3 categories Yao et al. (2015)
5 categories Waqas et al. (2013)
Continuous (n = 6) From 0 to 4 (n = 4) Black et al. (2015); Gerken and Volkwein (2000); Sackett et al.
(2009); Zheng et al. (2002)
From 0 to 7 Puddey and Mercer (2014)
12
From 0 to 100 Thiele et al. (2016)
Single grade (n = 4) Test at the end of the first year Categorical High or low performance Frischenschlager et al. (2005)
Introductory college chemistry Continuous (n = 3) From 0 to 12 Rodríguez Ayan and Ruiz Díaz(2011)
Introductory college chemistry From 0 to 100 Tai et al. (2005)
Introductory college biology From 0 to 100 Loehr et al. (2012)
Number of credits Course credits (n = 2) Continuous (n = 4) Black et al. (2015); Gerken and Volkwein (2000)
(n = 6) Study unit outcomes (n = 2) Ifenthaler and Widanapathirana (2014); Puddey and Mercer (2014)
Degree progression index Continuous From 0 to 1 Rodríguez Ayan and Ruiz Díaz(2011)
ECTS credits Continuous Triventi (2014)
Relative achievement Index to compare students' performance at Continuous From −1 to 1 Pedrosa et al. (2007)
the entrance and exit of all courses
Competencies (n = 6) Standardized test SABER PRO (Colombia) (n = 2) Continuous (n = 2) From 1 to 300 Bahamón and Reyes Ruiz (2014); Rodríguez Albor et al. (2014)
(n = 6) CET (College English Test) (China) Continuous The highest possible score is Yao et al. (2015)
710
DALF (France) Continuous From 0 to 100 Morlaix and Suchaut (2014)
GAMSAT (Australia) Continuous From 0 to 100 Puddey and Mercer (2014)
PELA (Australia) Continuous From 0 to 10 Anderton et al. (2016)
Persistence (n = 6) Completion status Dropout (n = 2) Categorical (n = 2) Yes/No Arulampalam et al. (2004); Gerken and Volkwein (2000)
(n = 6) Final degree classification (n = 2) Categorical (n = 2) British undergraduate degree Smith (2016); Thiele et al. (2016)
classification
Attending graduate school Categorical Yes/No Walpole (2003)
Graduation rate Continuous Stratton and Wetzel (2011)
Table 3b
Operationalizations of academic performance before university.
Measurement Operationalization Type of scale Scale Reference
Achievement (n = 20) HSGPA (n = 14) Categorical (n = 4) 2 categories Smith (2016)


4 categories Triventi (2014)
4 categories Morlaix and Suchaut (2014)
Continuous From 0 to 4 (n = 4) Gerken and Volkwein (2000); Rothstein (2004); Wolniak and Engberg (2010);
(n = 10) Zheng et al. (2002)
From 1 to 4 De Clercq et al. (2017)
From 1 to 5 (being 1 the best, and 5 Frischenschlager et al. (2005)
the worst)
From 1 to 6 Hansen and Mastekaasa (2006)
From 1 to 10 Bruinsma and Jansen (2007)
From 1 to 20 Gouvias et al. (2012)
UCAS tariff points Thiele et al. (2016)
Single grade (n = 3) Continuous (n = 3) HS grades in science, mathematics, Loehr et al. (2012); Tai et al. (2005)
and English courses (n = 2)
HS grades in biology, chemistry and Arulampalam et al. (2004)
physics courses
Subject area exams Leaving Certificate Continuous From 0 to 600 Delaney et al. (2011)
(n = 3) (Ireland)
13
NCEA (New Zealand) Continuous From 0 to 80 credits (each NCEA Shulruf et al. (2008)
level)
SABER 11 (Colombia) Continuous Before 2014: From 0 to 400 Bahamón and Reyes Ruiz (2014)
As of 2014: From 0 to 500
Competencies (n = 15) Aptitude exams SAT (USA) (n = 9) Continuous (n = 9) Before 2016: from 600 to 2400 Black et al. (2015); Gerken and Volkwein (2000); Loehr et al. (2012);
(n = 12) As of 2016: from 400 to 1600 Rothstein (2004); Sackett et al. (2009); Stratton and Wetzel (2011); Tai et al.
(2005); Walpole (2003); Wolniak and Engberg (2010)
ACT (USA) (n = 3) Continuous (n = 3) From 1 to 36 Stratton and Wetzel (2011); Wolniak and Engberg (2010); Zheng et al. (2002)
Subject area exams Vestibular (Brazil) Continuous (n = 3) Guimarães and Sampaio (2013)
(n = 3) Not specified (Greece) Gouvias et al. (2012)
Not specified (Vietnam) Nguyen (2016)
Persistence (n = 8) Admission ranks High school rank (n = 2) Continuous (n = 2) Black et al. (2015); Zheng et al. (2002)
(n = 5) Tertiary Entrance Rank Continuous (n = 3) From 0 to 99.95 Birch and Miller (2006); Win and Miller (2005)
(TER) (n = 2)
Australian Tertiary Anderton et al. (2016)
Admission Rank (ATAR)
Grade retention Years repeated in high Categorical Yes/No De Clercq et al. (2013)
(n = 3) school
Number of class repetitions Continuous (n = 2) Frischenschlager et al. (2005)
Years repeated in high Morlaix and Suchaut (2014)
school
5.2.2. Academic performance before university

This category includes operationalizations which assessed academic performance before attending university (n = 43). In this
case, the analyzed operationalizations were also grouped in three subcategories: prior academic achievement (n = 20), competencies
(n = 15), and persistence (n = 8).
In the subcategory prior academic achievement (n = 20), academic performance before university was measured using HSGPA
(n = 14), single grades (n = 3), and subject area exams (n = 3). HSGPA was calculated through both categorical (n = 4) and con-
tinuous (n = 10) scales. In particular, the continuous scale ranging from 0 to 4 (n = 4), was the most frequently reported in the
analyzed studies. Regarding single grades, grades in specific subjects such as science, mathematics, and English (n = 2) alongside
biology, chemistry, and physics, were reported in the examined primary studies. Furthermore, subject area exams such as the NCEA
qualification system in New Zealand (Shulruf et al., 2008), the Leaving Certificate in Ireland (Delaney et al., 2011), and the SABER 11
test in Colombia (Bahamón & Reyes Ruiz, 2014) were identified in the examined studies.
In the subcategory competencies (n = 15), several entrance exams to university were identified in the analyzed studies. These
exams were the SAT (formerly known as the Scholastic Assessment Test; n = 9) and the ACT (originally American College Testing;
n = 3) in the USA. These entrance exams are of importance because they indicate students’ academic preparation for university.
Remarkably, in three of the investigated references (Gouvias et al., 2012; Guimarães & Sampaio, 2013; Nguyen, 2016), subject area
exams were used as entrance exams to the university.
Within the subcategory persistence (n = 8), the operationalizations of academic performance before university were grouped
between admission ranks (n = 5) and grade retention (n = 3). In relation to admission ranks, the Tertiary Entrance Rank (TER) and
the Australian Tertiary Admission Rank (ATAR) were identified in the analyzed studies. ATAR ranks students’ previous academic
achievement in high school and is mostly used as an admission criterion for university in Australia (Li & Dockery, 2014). It is
important to note that the ATAR replaced the TER in 2010. Regarding grade retention (n = 3), the operationalizations identified in
the analyzed studies were the years repeated in high school (n = 2) and the number of class repetitions.
5.2.3. Model for measuring academic performance in higher education

Fig. 3 proposes a model to measure academic performance in higher education drawn from the summative content analysis
conducted in section 5.2. Achievement, competencies, and persistence are the three measurements used to assess academic perfor-
mance in higher education. Furthermore, those measurements can be assessed both during and before university.
5.3. The relationship between SES and academic performance in higher education
Table 4 provides information on the relationship between SES and academic performance in higher education, as well as the
mediators identified in the analyzed primary studies. Such information was analyzed through both summative content analysis and
meta-analysis. This section first presents the results for the SES-academic performance relationship. Next, the results for the in-
vestigated mediators are provided.
The summative content analysis demonstrated that there are three types of relationships between SES and academic performance
in higher education: namely, positive (n = 25), negative (n = 6), and no significant (n = 12) relationships. A positive relationship
indicated that the better one's socio-economic conditions, the better one's academic performance in higher education. However, a
closer review of the negative relationship (n = 6) revealed enlightening information. Pedrosa et al. (2007) indicated that students who
came from public schools had a better academic performance than their counterparts coming from private schools. Those students
with less favorable socio-economic conditions were able to develop a certain “educational resilience”, which was described by
Pedrosa et al. (2007) as the process of transforming early disadvantages in life into better academic performance in higher education.
In addition, students who either (a) were scholarship holders (Morlaix & Suchaut, 2014), (b) lived in a crowded house (Harb & El-
Fig. 3. Authors’ model for measuring academic performance in higher education.
14
Table 4
The relationship between SES and academic performance in higher education.
Reference N SES-Academic Average Prior academic PAA-AP Average University UE-AP Average Working status WS-AP Average
performance effect size achievement relationship effect size experience (UE) relationship effect size (WS) measure relationship effect size
relationship (PAA) measure measure
Anderton et al. (2016) 414 Not significant ATAR score Positive

Arulampalam et al. 51,810 Negative −0.01 HSGPA Negative −0.02
(2004)
Bahamón and Reyes Ruiz 68 Positive

(2014)
Beyene and Yimam 925 Positive 0.14 Entrance exam Positive 0.05 Academic Positive 0.12
(2016) score experience (number
of assessments)
Birch and Miller (2006) 1803 Negative −0.04+ TER score Positive 0.58
Black et al. (2015) 23,792 Negative −0.32+
Bonsaksen (2016) 123 Not significant 0.02 Prior experience Positive 0.26 Employed or not Not −0.04
in higher significant
education
Bozick (2007) 10,164 Positive 0.01 Employed or not Negative −0.13
Bruinsma and Jansen 62 Not significant 0.06 HSGPA Positive 0.56 Academic Negative −0.04 Employed or not Not 0.07
(2007) experience significant
(classroom climate,
quantity and
quality of the
instruction)
15
De Clercq et al. (2013) 111 Positive 0.26 Failure in Negative −0.21
secondary school
De Clercq et al. (2013) 206 Not significant 0.15 Failure in Negative −0.19 Social experience Not 0.15
secondary school (peer support) significant
De Clercq et al. (2017) 2178 Positive 0.12 HSGPA Positive 0.37
Delaney et al. (2011) 1867 Not significant −0.02 Entrance exam Positive 0.31
score
Frischenschlager et al. 245 Not significant HSGPA Positive
(2005)
Gerken and Volkwein – Not significant HSGPA Positive Academic Positive
(2000) experience
Gouvias et al. (2012) 874 Positive 0.06 HSGPA Positive 0.79
Guimarães and Sampaio 54,877 Positive 0.09
(2013)
Hansen and Mastekaasa 56,792 Positive 0.02 HSGPA Positive 0.36
(2006)
Harb and El-Shaarawi 296 Negative −0.05 Science HS Positive 0.11 Institutional Positive 0.17 Employed or not Negative −0.3
(2007) diploma experience
(positive attitude
towards university)
Ifenthaler and 146,001 Not significant
Widanapathirana
(2014)
Loehr et al. (2012) 2667 Positive HSGPA Positive
(continued on next page)
Table 4 (continued)
Reference N SES-Academic Average Prior academic PAA-AP Average University UE-AP Average Working status WS-AP Average
performance effect size achievement relationship effect size experience (UE) relationship effect size (WS) measure relationship effect size
relationship (PAA) measure measure
Morlaix and Suchaut 543 Negative HSGPA, no Positive

(2014) failure in
secondary school
Nguyen (2016) 616 Not significant

Pedrosa et al. (2007) 6701 Negative HS school courses Positive
Puddey and Mercer 219 Positive 0.17+ GPA at entry of Positive 0.25
(2014) graduate
program,
GAMSAT score
Rodríguez Albor et al. 14,829 Positive Employed or not Negative
(2014)
Rodríguez Ayan and Ruiz 312 Not significant −0.02 Employed or not Negative −0.22
Díaz(2011)
Rothstein (2004) 14,102 Positive 0.07 HSGPA, SAT Positive 0.71
Sackett et al. (2009) 17,630 Positive 0.09 SAT Positive 0.37
(n = 17,244)
Shulruf et al. (2008) 1880 Positive 0.01+
Smith (2016) 23,793 Positive 0.03 HSGPA Positive 0.34
Stratton and Wetzel 5823 Positive HSGPA Positive
(2011)
Tai et al. (2005) 1333 Positive 0.12 HSGPA, SAT Positive 0.17
16
Thiele et al. (2016) 3730 Positive 0.05+ UCAS points Positive 0.003
Triventi (2014) 1834 Positive
Walpole (2003) 6470 Positive 0.01 GPA at entry of Positive 0.09 Academic Positive 0.13
graduate (n = 1177) experience (n = 1177)
program (working on
research)
Wang et al. (2010) 323 Positive 0.06 Institutional Positive 0.12 Reasons for Positive 0.06
experience (sense of working and
belonging and characteristics of
school integration) work
Waqas et al. (2013) 267 Positive 0.05 Employed or not Negative −0.14
Win and Miller (2005) 1803 Positive 0.03+ TER score Positive 0.1
Wolniak and Engberg 3750 Positive HSGPA, SAT Positive 0.26
(2010)
Yanbarisova (2015) 1988 Not significant 0.03 Characteristics of Negative −0.01
work
Yao et al. (2015) 2989 Positive 0.16
Zheng et al. (2002) 1166 Not significant HSGPA Positive
(+) Effect sizes calculated from SES indicators at area level. Not considered in the analysis.
Fig. 4. Forest plot of the relationship between SES and academic performance in higher education.
Shaarawi, 2007), or (c) came from high schools with a high proportion of free/reduced lunches (Black, Lincove, Cullinane, & Veron,
2015), also had low academic performance in university. Finally, Arulampalam et al. (2004) suggested that students who had higher
scores in biology, chemistry, and physics in high school also had a lower likelihood of dropping out during university.
The meta-analysis revealed that the average effect size of the relationship between SES and academic performance in higher
education was weak and significant (ES = .06, Se = .013, CI = [.03; .08], p < .001). The Q test for homogeneity for this average
effect size was significant (Q = 460.30, df = 22, p < .001). Fig. 4 displays the forest plot with the average effect size of the re-
lationship between SES and academic performance in higher education.
5.3.1. Prior academic achievement

The mediating role of prior academic achievement was investigated based on the information presented in Table 4 and following
the rationale explained in section 4.3.1. Therefore, it was possible to compare the relationship between SES and academic perfor-
mance with the relationship between prior academic achievement and academic performance.
The summative content analysis suggested that HSGPA (n = 15), entrance exams (n = 5), admission ranks (n = 3), failure in
secondary school (n = 2), and type of high school diploma were the potential mediators of the relationship between SES and aca-
demic performance in higher education. Regarding HSGPA (n = 15), high school grades were more strongly related to academic
performance than SES. In such cases, the relationship between SES and academic performance was (a) positive (n = 9), (b) negative
(n = 2), and (c) not significant (n = 4). Concerning entrance exams (n = 5), it was found that SAT scores (n = 4) and Leaving
Certificate scores were more strongly related to academic performance than SES. Correspondingly, the reported relationship between
SES and academic performance was (a) positive (n = 4) and (b) not significant. In addition, failure in secondary school (n = 2) was
more strongly related to academic performance than SES. In such cases, the relationship between SES and academic performance was
(a) negative and (b) not significant. Finally, Harb and El-Shaarawi (2007) indicated that receiving a scientific diploma in high school
was more strongly related to GPA than living in a crowded household.
The meta-analysis demonstrated that the average effect size of the relationship between prior academic achievement and aca-
demic performance in higher education was positive and significant (ES = .29, Se = .07, CI = [.15; .42], p < .001). The Q test for
homogeneity for this average effect size was significant (Q = 12088.23, df = 20, p < .001).
17
5.3.2. University experience

The mediating role of university experience was investigated based on the information presented in Table 4 and following the
rationale explained in section 4.3.1. Thus, the relationship between SES and academic performance was compared to the relationship
between university experience and academic performance.
The summative content analysis revealed that academic experience (n = 3) and institutional experience (n = 2) were the po-
tential mediators of the relationship between SES and academic performance in higher education. Regarding academic experience
(n = 3), Bruinsma and Jansen (2007) indicated that students who had higher grades were also more satisfied with the teacher's
ability to explain the topic, the teacher's ability to use the resources, and the teacher's openness to questions. Moreover, the higher the
study load, the lower the grades in the course would be, while the higher the number of contact hours, the higher the grades. In this
case, Bruinsma and Jansen (2007) also found no significant relationship between SES and academic performance in higher education.
Similarly, Walpole (2003) suggested that low-SES students who had worked on their professors' research during university were more
likely to enroll in graduate programs following their completion of their undergraduate degree at university. In this case, the re-
lationship between SES and academic performance was significant but weaker. Finally, Gerken and Volkwein (2000) identified that
students' academic conscientiousness was more strongly related to their degree completion and GPA than their parental educational
level. Regarding institutional experience (n = 2), Harb and El-Shaarawi (2007) found that students who had a positive attitude towards
university also performed better academically. This influence was greater than the influence of SES on academic performance. Also,
Wang, Kong, Shan, and Vong (2010) reported that students' sense of belonging was more strongly related to students' GPA than their
father's education and family income.
The meta-analysis revealed that the average effect size of the relationship between university experience and academic perfor-
mance in higher education was positive and significant (ES = .13, Se = .02, CI = [.09; .16], p < .001). The Q test for homogeneity
for this average effect size was not significant (Q = 2.4094, df = 5, p = .79).
5.3.3. Working status

The mediating role of working status was investigated based on the information presented in Table 4 and following the rationale
explained in section 4.3.1. Hence, a comparison was made between the relationship between SES and academic performance and the
relationship between working status and academic performance.
The summative content analysis indicated that whether a student was employed or not (n = 5), the characteristics of their work
(n = 2) and their reasons for working were the potential mediators of the relationship between SES and academic performance in
higher education. Regarding being employed or not (n = 5), the influence of employment on academic performance was stronger than
the influence of SES. In such cases, the relationship between SES and academic performance was (a) positive (n = 3), (b) negative,
and (c) not significant. With respect to the characteristics of work, Yanbarisova (2015) found that students who were working full-time
outside their academic fields displayed worse academic performance than their counterparts. In this case, the relationship between
SES and academic performance was not significant. In addition, Wang et al. (2010) indicated that jobs that provided students the
opportunity to learn new things had a greater influence on their academic performance than their father's education and occupation,
and their family income. Regarding the reasons for working, Wang et al. (2010) also indicated that when the reason for working was to
acquire working experience, the part-time jobs were more strongly related to students' GPA than their father's education and oc-
cupation, and their family income.
The meta-analysis demonstrated that the average effect size of the relationship between working status and academic perfor-
mance in higher education was negative and significant (ES = −.10, Se = .05, CI = [-.19; −.01], p < .001). The Q test for
homogeneity for this average effect size was significant (Q = 51.7643, df = 7, p < .001).
In summary, the summative content analysis suggested that the investigated mediators were more strongly related to academic
performance than SES. Furthermore, the meta-analysis showed that the average effect sizes of the mediators were significant and
larger than the effect size of the SES-academic performance relationship. The mediator with the largest average effect size was prior
academic achievement, followed by university experience, and working status.
6. Discussion
The objectives of this systematic literature review were (1) to analyze how SES and academic performance in higher education are
measured; (2) to determine whether the relationship between SES and academic performance in higher education is mediated by a)
prior academic achievement; b) university experience; and c) working status.
6.1. Conclusions and implications for practice
6.1.1. Measuring SES

The first conclusion of this study is that five major measurements should be considered when assessing SES: education, occu-
pation, income, household resources, and neighborhood resources. The findings of this systematic literature review also suggest
specific ways to operationalize each one of the measurements of SES.
First, it was found that education is traditionally assessed through categorical variables that indicate the achieved academic
degree, ranging from no education up to a doctoral degree. In this respect, the use of the International Standard Classification of
Education (ISCED) is highly recommended. The ISCED has established a unique scale which allows for comparisons to be made
among different international contexts.
18
Second, the results of this study showed that occupation is predominantly assessed using categorical scales. In this regard, a well-
established classification to measure occupation is the Standard Occupation Classification (SOC). Such a classification is based on the
type of work, skills, and level of education. Even though the SOC was first proposed in the USA, national variants also exist in
countries within Europe and Asia. Therefore, the SOC could be used not only for international comparisons but also as a classification
system for countries which have not developed their own classification system.
Third, the findings of this study indicated that income is typically measured with intervals to categorize the amount of money
being earned. However, a more advisable way to operationalize income is through a multiple of the minimum wage paid in each
country (i.e., equivalent to one minimum wage, twice minimum wage, etc.). As such, the minimum wage could be related to the type
of work and level of education of a wage-earner. Besides, the use of the minimum wage as a measurement would allow for the
comparison of socio-economic conditions across several different countries.
Fourth, although the results from this study revealed that measures of household resources are related to the possessions available
at home, it seems adequate to distinguish between material resources and cultural resources (Gouvias et al., 2012). Material resources
are merely the items students have at home (i.e., a personal computer, internet connection, an individual room), while cultural
resources are items which might represent an intellectual added value to the students (i.e., books related to schoolwork or musical
instruments).
Finally, this study revealed that neighborhood resources can be operationalized through indexes which rank areas according to
their socio-economic advantages and disadvantages. Well-established examples of these indexes are the Socio-Economic Indexes for
Areas (SEIFA) proposed by the Australian Bureau of Statistics and the Index of Multiple Deprivation (IMD) proposed by the UK
government. Even though these indexes were developed for specific national contexts, the methodology underlying their creation can
be replicated to develop indexes within each country. However, it does not seem convenient to use these indices directly to measure
students’ SES. What is recommended, instead, is to use area-based indicators such as SEIFA to achieve a better understanding of the
social and economic conditions where students live (Australian Bureau of Statistics, 2011). This suggestion is consistent with previous
meta-analytic studies focusing on primary and secondary education (e.g., Sirin, 2005; White, 1982), where it has been recommended
to avoid using aggregated indicators to assess SES at the individual level, as they can overestimate the effect of the relationship
between SES and academic performance.
6.1.2. Measuring academic performance in higher education

The second conclusion of this study is that academic performance in higher education should be assessed considering three major
measurements: academic achievement, competencies, and persistence. This study also identified several ways to operationalize such
measures both during and prior to university.
Regarding academic achievement, the results from this study corroborate the assertion that first-year GPA is the most used
operationalization of academic achievement at university, as first-year GPA is considered a strong predictor of subsequent academic
outcomes at university (Cliffordson, 2008; Gerken & Volkwein, 2000). When the interest is to compare GPA across different fields and
institutions, it is necessary to take into account that the grading system does change between fields and institutions of higher
education; furthermore, there could also be variations in the assessment process regardless of the use of the same grading scale
(Hansen & Mastekaasa, 2006). In addition, the findings of this study also reveal that HSGPA is the most common operationalization of
prior academic achievement before university. HGSPA depends on the curriculum followed in each institution (Westrick et al., 2015),
the quality and strictness of the scoring system, as well as the student population in each institution. The selection of HSGPA as a
predictor of academic performance increases the explained variance of GPA at university (Zheng et al., 2002), and furthermore,
HSGPA seems to have a larger predictive validity than entrance exams, regardless of the grading system and the academic program
(Cliffordson, 2008).
With reference to competencies, the findings of this systematic literature review indicate that competencies have been oper-
ationalized through the outcomes of standardized testing both during and prior to university. However, the difference across levels
might lie on the purpose of such tests. In the case of academic performance at university, standardized tests aim to evaluate the
acquisition of both general and specific competencies pertaining to each study area. In the case of academic performance before
enrolling in university, standardized tests are designed as entrance exams to estimate the students’ academic preparation for their
university studies.
Nevertheless, the use of achievement tests as entrance exams to university was also identified in the examined primary studies.
This finding is somewhat problematic for several reasons. Firstly, achievement tests are designed to measure past accomplishment in
learning instead of measuring the capacity for future accomplishment (Sidhu, 2005). Secondly, whereas a great deal is known about
the predictive validity of aptitude tests (e.g., the SAT I), the predictive validity of achievement tests (e.g., the SAT II) is still unclear in
the literature (Cliffordson, 2008). Thirdly, Zwick (2012) has suggested that achievement tests might indicate the degree to which
wealthier students have access to either better information for the test (content hypothesis) or better preparation for the test
(coaching hypothesis). Therefore, achievement tests seem to be more related to students’ SES than aptitude tests.
Finally, this study suggests that persistence in university can be measured in terms of students' degree completion. In addition,
persistence before university can be assessed as the students' academic rank and grade repetition. An important conceptual distinction
between persistence and retention should be acknowledged, as persistence is an individual phenomenon, while retention is an
institutional one; therefore, these terms should not be used interchangeably (Reason, 2009).
6.1.3. The relationship between SES and academic performance in higher education
The third conclusion of this study is that the relationship between SES and academic performance in higher education is weak.
19
This result is coherent with previous meta-analyses in higher education, which have also reported a weak effect size (e.g., Richardson
et al., 2012). A critical interpretation of the findings of this study could beg the question of the actual importance of SES as a predictor
of academic performance in higher education. To begin with, it could be the case that the influence of socio-economic conditions on
students’ performance is lower in higher education than in previous levels of education (the influence of SES on prior academic
achievement has been well-established in educational research). One can also argue that regardless of the hindrances that low-SES
students face when entering university, those who are admitted share a similar educational experience to their wealthier counterparts
(Smith, 2016). Therefore, the higher education system could have the same influence on any student despite his or her socio-
economic conditions.
Furthermore, the findings of this study indicate that the relationship between SES and academic performance in higher education
is weak when other factors are considered. This fact urges the educational research field to select more robust analytical techniques
when investigating academic performance in higher education. A mere bivariate analysis no longer suffices. In addition, regardless of
the multiple theories in higher education which suggest that SES and academic performance are positively related, strong empirical
evidence supporting these theoretical claims is still missing in the educational literature (Marks, 2017).
However, a weak relationship between SES and academic performance in higher education does not imply that low-SES students
should be ignored or that increasing their participation in university should be dismissed. How to efficiently attend to low-SES
students’ educational needs remains a challenge for the higher education system, and in this respect, the findings of this systematic
literature review might be transferred in one of three ways to properly deal with that challenge.
Firstly, it was identified that prior academic achievement is more strongly related to academic performance in university than
SES. While this finding is not surprising, it does support the need to reinforce the past performance of low-SES students through, for
instance, academic preparation courses before university. An example of this type of program is the enabling programs proposed by
the Australian government. Enabling programs are designed to provide disadvantaged students with specific competencies (e.g.,
literacy, numeracy, communication, and critical thinking skills) so that they can be prepared for their university studies (Pitman
et al., 2016). Moreover, enabling programs are effective pathways to higher education for almost half of the enrolled disadvantaged
students (Hodges et al., 2013).
Secondly, the findings from this study also revealed that the influence of university experience on academic performance is larger
than the influence of SES. The defining factors of the university experience such as classroom climate, quantity and quality of the
instruction (Bruinsma & Jansen, 2007), sense of belonging and school integration (Wang et al., 2010), and peer support (De Clercq
et al., 2013) would help low-SES students to adapt to their new academic settings at university. Interestingly, Devlin, Kift, Nelson,
Smith, and McKay (2012) have proposed a set of teaching guidelines in order to foster low-SES students’ academic performance. Far
from being prescriptive, these guidelines can be understood as key practical advice for teachers whose students come from low socio-
economic settings.
Thirdly, there was also evidence of the likely mediating role of working status in the SES-academic performance relationship. This
finding supports the assumption that working during university studies might have a negative influence on students’ performance.
However, students who work part-time within their academic fields might perform better academically than students who work
outside their academic areas (Wang et al., 2010; Yanbarisova, 2015). Thus, working should allow low-SES students not only to
overcome their financial needs but also to extend their academic experience by increasing their body of knowledge while attending
university. A concrete example of this type of job for low-SES students is serving as an undergraduate teaching assistant (UTA).
6.2. Limitations
Although the results of this systematic literature review provide insights into the relationship between SES and academic per-
formance in higher education, several limitations of the present study should be acknowledged. First, the number of studies which
explore the relationship between SES and academic performance is quite low. Hence, all studies that have aimed to predict academic
performance in higher education were considered. However, a precise definition of the relationship between SES and academic
performance was often lacking in the analyzed studies. Second, several additional variables which might also interact with SES were
not always reported in the reviewed studies. Thus, variables such as students' cognitive factors could not be considered for the
summative content analysis. Third, information about the students' academic programs was not always reported in the analyzed
studies. Therefore, whether the relationship between SES and academic performance in higher education depends on students’
academic programs could not be determined. Fourth, there was no information on the relationship between SES and the meta-
analyzed mediators. Therefore, a complete analysis of the selected mediators could not be carried out. Finally, the Q test for
homogeneity was not significant for the average effect size of university experience. This result might be suggesting that the resulting
average effect size is less generalizable.
6.3. Implications for future research
Starting from the findings and limitations of this systematic literature review discussed earlier, future research in higher education
could benefit from focusing on several topics, listed as follows. Firstly, it is interesting to note that using dimensional reduction
techniques such as Principal Component Analysis (PCA) leads to a composite score to assess SES. However, composite scores also
represent a limitation to fully capturing the underlying variance of SES indicators. Therefore, the matter of how to construct com-
posite scores to assess SES is an issue which requires further exploration.
Secondly, previous research at the elementary and secondary educational levels (e.g., Sirin, 2005; White, 1982) has demonstrated
20
the moderating role of methodological aspects, such as the type of SES-achievement measure, on the relationship between SES and
academic performance. However, exploring such a moderating role still remains an unfinished task for the higher education field.
Thirdly, the criticism that standardized tests merely measure students’ socio-economic conditions and do not predict their future
academic performance in higher education (Mattern, Shawn, & Williams, 2008) persists among some educational researchers.
Therefore, further research could contribute to clarifying the relationship between standardized test outcomes, SES, and academic
performance at university.
Fourthly, when considering working status to predict academic performance, it seems essential to include additional char-
acteristics such as the type of work (part-time or full-time), correspondence with the academic field, and reasons for working. In fact,
including working status merely as a dichotomous variable in any predictive analysis of academic performance could be misleading
(Wang et al., 2010). Therefore, to gain a better understanding of the influence of working status, both quantitative and qualitative
research methods are highly recommended (Yanbarisova, 2015).
Fifthly, after completion of their undergraduate degree, low- SES students are more likely to join the workforce instead of
pursuing a postgraduate degree, as seems to be the case with high- SES students (Walpole, 2003). However, an additional question
worth exploring is what happens with the relationship between SES and the academic performance of low- SES students who do
continue to the postgraduate educational level.
Finally, recent educational research (Musso, Kyndt, Cascallar, & Dochy, 2012, 2013; Cascallar, Musso, Kyndt, & Dochy, 2015;
Kyndt, Musso, Cascallar, & Dochy, 2015; Musso & Cascallar, 2009) has utilized predictive systems based on neural network ap-
proaches to study academic performance in primary, secondary, and higher education. The improvement of the validity, the increase
in the accuracy of the predictions and classifications, and the possibility to determine the predictive weight contributions of each of
the variables in the models are the principal advantages of developing predictive systems based on neural networks. Therefore, we
would like to encourage the use of neural networks in order to gain a more exhaustive understanding of the relationship between SES
and academic performance in higher education.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.org/10.1016/j.edurev.2019.100305.
References
Studies marked with an asterisk (∗) were included in the summative content analysis.
Studies marked with a cross (+) were included in the meta-analysis.
Aloe, A. M., & Becker, B. J. (2012). An effect size for regression predictors in meta-analysis. Journal of Educational and Behavioral Statistics, 37, 278–297. https://doi.
org/10.3102/1076998610396901.
Aloe, A. M., & Thompson, C. G. (2013). The synthesis of partial effect sizes. Journal of the Society for Social Work and Research, 4, 390–405. https://doi.org/10.5243/
jsswr.2013.24.
(∗)Anderton, R., Evans, T., & Chivers, P. (2016). Predicting academic success of health science students for first year anatomy and physiology. International Journal of
Higher Education, 5, 250–260. https://doi.org/10.5430/ijhe.v5n1p250.
(∗+)Arulampalam, W., Naylor, R., & Smith, J. (2004). Factors affecting the probability of first year medical student dropout in the UK: A logistic analysis for the intake
cohorts of 1980–92. Medical Education, 38, 492–503. https://doi.org/10.1046/j.1365-2929.2004.01815.x.
Astin, A. W. (1984). Student involvement: A developmental theory for higher education. Journal of College Student Personnel, 25, 297–308.
Astin, A. W. (1993). What matters in college? Four years revisited. San Francisco, USA: Jossey-Bass.
Astin, A. W. (1999). Student involvement: A developmental theory for higher education. Journal of College Student Development, 40, 518–529.
Australian Bureau of Statistics (2011). Measures of socioeconomic status. ABS Publication No. 1244.0.55.001 . Retrieved from: http://www.ausstats.abs.gov.au/
Ausstats/subscriber.nsf/0/367D3800605DB064CA2578B60013445C/$File/1244055001_2011.pdf.
Aveyard, H. (2014). Doing a literature review in health and social care: A practical guide. London, United Kingdom: McGraw-Hill Educationhttps://doi.org/10.7748/nr.18.
4.45.s2.
(∗)Bahamón, M. J., & Reyes Ruiz, L. (2014). Characterization of the intellectual capacity, sociodemographic, and academic factors of the students with high and low
performance in the SABER PRO test – 2012 [Caracterización de la capacidad intelectual, factores sociodemográficos y académicos de estudiantes con alto y bajo
desempeño en los exámenes SABER PRO-AÑO 2012]. Avances en Psicología Latinoamericana, 32, 459–476. https://doi.org/10.12804/apl32.03.2014.01.
Betts, L. R., Elder, T. J., Hartley, J., & Blurton, A. (2008). Predicting university performance in Psychology: The role of previous performance and discipline-specific
knowledge. Educational Studies, 34, 543–556. https://doi.org/10.1080/03055690802288528.
(∗+)Beyene, K. M., & Yimam, J. A. (2016). Multilevel analysis for identifying factors influencing academic achievement of students in higher education institution:
The case of Wollo University. Journal of Education and Practice, 7, 17–23.
(∗+)Birch, E. R., & Miller, P. W. (2006). Student outcomes at university in Australia: A quantile regression approach. Australian Economic Papers, 45, 1–17. https://doi.
org/10.1111/j.1467-8454.2006.00274.x.
(∗)Black, S. E., Lincove, J., Cullinane, J., & Veron, R. (2015). Can you leave high school behind?. Economics of Education Review, 46, 52–63. https://doi.org/10.3386/
w19842.
Blömeke, S., Zlatkin-Troitschanskaia, O., Kuhn, C., & Fege, J. (2013). Modeling and measuring competencies in higher education. In S. Blömeke, O. Zlatkin-
Troitschanskaia, C. Kuhn, & J. Fege (Eds.). Professional and vet learning: Vol. 1. Modeling and measuring competencies in higher education (pp. 1–10). Dordrecht, The
Netherlands: Sense Publishers. https://doi.org/10.1007/978-94-6091-867-4.
(∗+)Bonsaksen, T. (2016). Predictors of academic performance and education programme satisfaction in occupational therapy students. British Journal of Occupational
Therapy, 79, 361–367. https://doi.org/10.1177/0308022615627174.
Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2009). Introduction to meta-analysis. Chichester, UK: John Wiley.
Bourdieu, P. (1986). The forms of capital. In J. Richardson (Ed.). Handbook of theory and research for the sociology of education (pp. 241–258). New York, NY:
Greenwood.
Bowman, N. A. (2012). Effect sizes and statistical methods for meta-analysis in higher education. Research in Higher Education, 53, 375–382. https://doi.org/10.1007/
s11162-011-9232-5.
(∗+)Bozick, R. (2007). Making it through the first year of college: The role of students' economic resources, employment, and living arrangements. Sociology of
Education, 80, 261–285. https://doi.org/10.1177/003804070708000304.
(∗+)Bruinsma, M., & Jansen, E. P. W. A. (2007). Educational productivity in higher education: An examination of part of the Walberg educational productivity model.
21
School Effectiveness and School Improvement, 18, 45–65. https://doi.org/10.1080/09243450600797711.

Bureau of Labor Statistics (2006). Standard occupational classification-revision for 2010. Retrieved from https://www.bls.gov/soc/soc_may06.pdf.
Byrne, M., & Flood, B. (2008). Examining the relationships among background variables and academic performance of first year accounting students at an Irish
University. Journal of Accounting Education, 26, 202–212. https://doi.org/10.1016/j.jaccedu.2009.02.001.
Cascallar, E., Musso, M. F., Kyndt, E., & Dochy, F. (2015). Modelling for understanding and for prediction/classification the power of neural networks in research.
Frontline Learning Research, 2, 67–81. https://doi.org/10.14786/flr.v2i5.135.
Casillas, A., Robbins, S., Allen, J., Kuo, Y. L., Hanson, M. A., & Schmeiser, C. (2012). Predicting early academic failure in high school from prior academic achievement,
psychosocial characteristics, and behavior. Journal of Educational Psychology, 104, 407–420. https://doi.org/10.1037/a0027180.
Cliffordson, C. (2008). Differential prediction of study success across academic programs in the Swedish context: The validity of grades and tests as selection in-
struments for higher education. Educational Assessment, 13, 56–75. https://doi.org/10.1080/10627190801968240.
Coleman, J., Campbell, E., Hobson, C., McPartland, F., Mood, A., & Weinfeld, F. (1966). Equality of educational opportunity. Washington DC: U.S. Government.
Cowan, C. D., Hauser, R., Kominski, R., Levin, H., Lucas, S., Morgan, S., et al. (2012). Improving the measurement of socioeconomic status for the national assessment of
educational progress: A theoretical foundation. Retrieved from National Center for Education Statistics website https://nces.ed.gov/nationsreportcard/pdf/
researchcenter/Socioeconomic_Factors.pdf.
Crawford, C. (2014). Socio-economic differences in university outcomes in the UK: Drop-out, degree completion and degree class. Institute for Fiscal Studies Working Paper
No. W14/31 . Retrieved from Institute for Fiscal Studies website https://www.ifs.org.uk/uploads/publications/wps/WP201431.pdf.
(∗+)De Clercq, M., Galand, B., Dupont, S., & Frenay, M. (2013). Achievement among first-year university students: An integrated and contextualized approach.
European Journal of Psychology of Education, 28, 641–662. https://doi.org/10.1007/s10212-012-0133-6.
(∗+)De Clercq, M., Galand, B., & Frenay, M. (2017). Transition from high school to university: A person-centered approach to academic achievement. European Journal
of Psychology of Education, 32, 39–59. https://doi.org/10.1007/s10212-016-0298-5.
(∗+)Delaney, L., Harmon, C., & Redmond, C. (2011). Parental education, grade attainment and earnings expectations among university students. Economics of
Education Review, 30, 1136–1152. https://doi.org/10.1016/j.econedurev.2011.04.004.
Devlin, M., Kift, S., Nelson, K., Smith, E., & McKay, J. (2012). Effective teaching and support of students from low socioeconomic status backgrounds: Resources for Australian
higher education. Sydney, Australia: Department of Industry, Innovation, Science, Research, and Tertiary Education.
Dika, S. L., & Singh, K. (2002). Applications of social capital in educational literature: A critical synthesis. Review of Educational Research, 72, 31–60. https://doi.org/
10.3102/00346543072001031.
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315, 629–634. https://doi.org/10.1136/
bmj.315.7109.629.
Erola, J., Jalonen, S., & Lehti, H. (2016). Parental education, class and income over early life course and children's achievement. Research in Social Stratification and
Mobility, 44, 33–43. https://doi.org/10.1016/j.rssm.2016.01.003.
Fernández-Castilla, B., Aloe, A. M., Declercq, L., Jamshidi, L., Onghena, P., Beretvas, S. N., et al. (2019). Concealed correlations meta-analysis: A new method for
synthesizing standardized regression coefficients. Behavior Research Methods, 51, 316–331. https://doi.org/10.3758/s13428-018-1123-7.
(∗)Frischenschlager, O., Haidinger, G., & Mitterauer, L. (2005). Factors associated with academic success at Vienna Medical School: Prospective survey. Croatian
Medical Journal, 46, 58–65.
Galobardes, B., Shaw, M., Lawlor, D. A., Lynch, J. W., & Smith, G. D. (2006). Indicators of socioeconomic position (part 1). Journal of Epidemiology & Community Health,
60, 7–12. https://doi.org/10.1136/jech.2004.023531.
Gamoran, A., & Long, D. A. (2007). Equality of educational opportunity. A 40-year retrospective. In R. Teese, S. Lamb, & M. Duru-Bellat (Eds.). International studies in
educational inequality, theory and policy (pp. 23–47). Dordrecht, The Netherlands: Springer. https://doi.org/10.1007/978-1-4020-5916-2_2.
Ganzeboom, H. B., De Graaf, P. M., & Treiman, D. J. (1992). A standard international socio-economic index of occupational status. Social Science Research, 21, 1–56.
https://doi.org/10.1016/0049-089x(92)90017-b.
(∗)Gerken, J. T., & Volkwein, J. F. (2000, May). Pre-college characteristics and freshman year experiences as predictors of 8-year college outcomes. Cincinnati, OH: Paper
presented at the annual meeting of the Association for Institutional Research.
Glasman, N. S., & Biniaminov, I. (1981). Input-output analyses of schools. Review of Educational Research, 51, 509–539. https://doi.org/10.3102/
00346543051004509.
(∗+)Gouvias, D., Katsis, A., & Limakopoulou, A. (2012). School achievement and family background in Greece: A new exploration of an omnipresent relationship.
International Studies in Sociology of Education, 22, 125–145. https://doi.org/10.1080/09620214.2012.700186.
(∗+)Guimarães, J., & Sampaio, B. (2013). Family background and students' achievement on a university entrance exam in Brazil. Education Economics, 21, 38–59.
https://doi.org/10.1080/09645292.2010.545528.
(∗+)Hansen, M. N., & Mastekaasa, A. (2006). Social origins and academic performance at university. European Sociological Review, 22, 277–291. https://doi.org/10.
1093/esr/jci057.
(∗+)Harb, N., & El-Shaarawi, A. (2007). Factors affecting business students' performance: The case of students in United Arab Emirates. The Journal of Education for
Business, 82, 282–290. https://doi.org/10.3200/joeb.82.5.282-290.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. London: Academic Press.
Heyvaert, M., Maes, B., & Onghena, P. (2013). Mixed methods research synthesis: Definition, framework, and potential. Quality and Quantity, 47, 659–676. https://doi.
org/10.1007/s11135-011-9538-6.
Hilton, T. (1982). Persistence in higher education: A empirical study(College Board report 82-5) . Retrieved from College Board website http://research.collegeboard.org/
sites/default/files/publications/2012/7/researchreport-1982-5-persistence-higher-education.pdf.
Hodges, B., Bedford, T., Hartley, J., Klinger, C., Murray, N., O'Rourke, J., et al. (2013). Enabling retention: Processes and strategies for improving student retention in
university-based enabling programs: Final report 2013Sydney, Australia: Australian Government Office for Learning and Teaching.
Hsieh, H. F., & Shannon, S. E. (2005). Three approaches to qualitative content analysis. Qualitative Health Research, 15, 1277–1288. https://doi.org/10.1177/
1049732305276687.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis. Correcting error and bias in research findings. California: Sage Publications.
(∗)Ifenthaler, D., & Widanapathirana, C. (2014). Development and validation of a learning analytics framework: Two case studies using support vector machines.
Technology, Knowledge and Learning, 19, 221–240. https://doi.org/10.1007/s10758-014-9226-4.
Jæger, M. M. (2011). Does cultural capital really affect academic achievement? New evidence from combined sibling and panel data. Sociology of Education, 84,
281–298. https://doi.org/10.1177/0038040711417010.
Kelly, L. J. (1996, May). Implementing Astin's IEO model in the study of student retention: A multivariate time dependent approach. Paper presented at the 36th annual
forum of the association for institutional research, Albuquerque, NM.
Krieger, N., Williams, D. R., & Moss, N. E. (1997). Measuring social class in US public health research: Concepts, methodologies, and guidelines. Annual Review of Public
Health, 18, 341–378. https://doi.org/10.1146/annurev.publhealth.18.1.341.
Kyndt, E., Musso, M., Cascallar, E. C., & Dochy, F. (2015). Predicting academic performance: The role of cognition, motivation and learning approaches. A neural
network analysis. In V. Donche, S. De Maeyer, D. Gijbels, & H. van den Bergh (Eds.). Methodological challenges in research on student learning (pp. 55–76). Antwerp,
Belgium: Garant.
Li, G., Chen, W., & Duanmu, J. L. (2010). Determinants of international students' academic performance: A comparison between Chinese and other international
students. Journal of Studies in International Education, 14, 389–405. https://doi.org/10.1177/1028315309331490.
Li, I. W., & Dockery, A. M. (2014). Socio-economic status of schools and university academic performance: Implications for Australia's higher education expansion. National
centre for student equity in higher education (NCSEHE)Perth: Curtin University.
(∗)Loehr, J. F., Almarode, J. T., Tai, R. H., & Sadler, P. M. (2012). High school and college biology: A multi-level model of the effects of high school courses on
introductory course performance. Journal of Biological Education, 46, 165–172. https://doi.org/10.1080/00219266.2011.617767.
22
Long, D. (2012). Theories and models of student development. In L. J. Hinchliffe, & M. A. Wong (Eds.). Environments for student growth and development: Librarians and
student affairs in collaboration (pp. 41–55). Chicago: Association of College & Research Libraries.
Marks, G. N. (2017). Is SES really that important for educational outcomes in Australia? A review and some recent evidence. Australian Educational Researcher, 44,
191–211. https://doi.org/10.1007/s13384-016-0219-2.
Mattern, K., Shawn, E., & Williams, F. (2008). Examining the relationship between the SAT®, high school measures of academic performance, and socioeconomic status:
Turning our attention to the unit of analysis (College Board RN-36) . Retrieved from: https://files.eric.ed.gov/fulltext/ED562606.pdf.
McKenzie, K., & Schweitzer, R. (2001). Who succeeds at university? Factors predicting academic performance in first year Australian university students. Higher
Education Research and Development, 20, 21–33. https://doi.org/10.1080/07924360120043621.
(∗)Morlaix, S., & Suchaut, B. (2014). The social, educational and cognitive factors of success in the first year of university: A case study. International Review of
Education, 60, 841–862. https://doi.org/10.1007/s11159-014-9459-4.
Mueller, C. W., & Parcel, T. L. (1981). Measures of socioeconomic status: Alternatives and recommendations. Child Development, 52, 13–30. https://doi.org/10.2307/
1129211.
Musso, M. F., & Cascallar, E. C. (2009). Predictive systems using artificial neural networks: An introduction to concepts and applications in education and social
sciences. In M. C. Richaud, & J. E. Moreno (Vol. Eds.), Research in behavioral sciences: Vol. 1, (pp. 433–459). Buenos Aires, Argentina: CIIPME/CONICET.
Musso, M., Kyndt, E., Cascallar, E. C., & Dochy, F. (2012). Predicting mathematical performance: The effects of cognitive processes and self-regulation factors.
Educational Research International, 1–13. https://doi.org/10.1155/2012/250719 2012.
Musso, M. F., Kyndt, E., Cascallar, E. C., & Dochy, F. (2013). Predicting general academic performance and identifying the differential contribution of participating
variables using artificial neural networks. Frontline Learning Research, 1, 42–71. https://doi.org/10.14786/flr.v1i1.13.
Nakao, K., & Treas, J. (1994). Updating occupational prestige and socioeconomic scores: How the new measures measure up. Sociological Methodology, 24, 1–72.
https://doi.org/10.2307/270978.
National Institute for Health and Clinical Excellence (2009). Methods for the development of NICE public health guidance (2nd ed.). . Retrieved from http://www.nice.org.
uk/media/2FB/53/PHMethodsManual110509.pdf.
(∗)Nguyen, T. M. (2016). Learning approaches, demographic factors to predict academic outcomes. International Journal of Educational Management, 30, 653–667.
https://doi.org/10.1108/ijem-06-2014-0085.
Nonis, S. A., & Hudson, G. I. (2006). Academic performance of college students: Influence of time spent studying and working. The Journal of Education for Business, 81,
151–159. https://doi.org/10.3200/joeb.81.3.151-159.
(∗)Pedrosa, R. H., Dachs, J. N. W., Maia, R. P., Andrade, C. Y., & Carvalho, B. S. (2007). Academic performance, students' background and affirmative action at a
Brazilian University. Higher Education Management and Policy, 19, 1–20. https://doi.org/10.1787/hemp-v19-art18-en.
Peterson, R. A., & Brown, S. P. (2005). On the use of beta coefficients in meta-analysis. Journal of Applied Psychology, 90, 175–181. https://doi.org/10.1037/0021-
9010.90.1.175.
Pike, G. R., & Saupe, J. L. (2002). Does high school matter? An analysis of three methods of predicting first-year grades. Research in Higher Education, 43, 187–207.
Pitman, T., Trinidad, S., Devlin, M., Harvey, A., Brett, M., & McKay, J. (2016). Pathways to higher education: The efficacy of enabling and sub-bachelor pathways for
disadvantaged students. National centre for student equity in higher education (NCSEHE). Perth: Curtin University.
(∗+)Puddey, I. B., & Mercer, A. (2014). Predicting academic outcomes in an Australian graduate entry medical programme. BMC Medical Education, 14, 1–12. https://
doi.org/10.1186/1472-6920-14-31.
Reason, R. D. (2009). An examination of persistence research through the lens of a comprehensive conceptual framework. Journal of College Student Development, 50,
659–682. https://doi.org/10.1353/csd.0.0098.
Richardson, M., Abraham, C., & Bond, R. (2012). Psychological correlates of university students' academic performance: A systematic review and meta-analysis.
Psychological Bulletin, 138, 353–387. https://doi.org/10.1037/a0026838.
Roberts, A. S. (2007). Predictors of future performance in architectural design education. Educational Psychology, 27, 447–463. https://doi.org/10.1080/
01443410601104361.
Rochford, C., Connolly, M., & Drennan, J. (2009). Paid part-time employment and academic performance of undergraduate nursing students. Nurse Education Today,
29, 601–606. https://doi.org/10.1016/j.nedt.2009.01.004.
(∗)Rodríguez Albor, G., Ariza Dau, M., & Ramos Ruíz, J. L. (2014). Calidad institucional y rendimiento académico: El caso de las universidades del Caribe colombiano.
[Institutional quality and academic performance: The case of Colombian Caribbean universities]. Perfiles Educativos, 36, 10–29. https://doi.org/10.1016/s0185-
2698(14)70607-5.
(∗+)Rodríguez Ayan, M. N., & Ruiz Díaz, M. A. (2011). Performance's indicators of college students: Grades versus accumulated credits [Indicadores de rendimiento
de estudiantes universitarios: Calificaciones versus créditos acumulados]. doi:10-4438/1988-592X-RE-2011-355-033 Revista de Educación, 355, 467–492.
(∗+)Rothstein, J. M. (2004). College performance predictions and the SAT. Journal of Econometrics, 121, 297–317. https://doi.org/10.1016/j.jeconom.2003.10.003.
Rothstein, H. R., Sutton, A. J., & Borenstein, M. (2005). Publication bias in meta-analysis: Prevention, assessment and adjustments. Chichester, UK: John Wiley.
(∗+)Sackett, P. R., Kuncel, N. R., Arneson, J. J., Cooper, S. R., & Waters, S. D. (2009). Does socioeconomic status explain the relationship between admissions tests and
post-secondary academic performance? Psychological Bulletin, 135, 1–22. https://doi.org/10.1037/a0013978.
Sackett, P. R., Kuncel, N. R., Beatty, A. S., Rigdon, J. L., Shen, W., & Kiger, T. B. (2012). The role of socioeconomic status in SAT-grade relationships and in college
admissions decisions. Psychological Science, 23, 1000–1007. https://doi.org/10.1177/0956797612438732.
Sadler, D. R. (2013). Making competent judgments of competence. In S. Blömeke, O. Zlatkin-Troitschanskaia, C. Kuhn, & J. Fege (Eds.). Professional and vet learning:
Vol. 1. Modeling and measuring competencies in higher education (pp. 13–27). Dordrecht, The Netherlands: Sense Publishers. https://doi.org/10.1007/978-94-6091-
867-4.
Scheerens, J. (1990). School effectiveness research and the development of process indicators of school functioning. School Effectiveness and School Improvement, 1,
61–80. https://doi.org/10.1080/0924345900010106.
Schneider, S. L. (2013). The international standard classification of education 2011. In G. E. Birkelund (Ed.). Comparative social research: Vol. 30. Class and stratification
analysis (pp. 365–379). Bingley, UK: Emerald Group Publishing Limited. https://doi.org/10.1108/s0195-6310(2013)0000030017.
Schneider, M., & Preckel, F. (2017). Variables associated with achievement in higher education: A systematic review of meta-analyses. Psychological Bulletin, 143,
565–600. https://doi.org/10.1037/bul0000098.
Shavelson, R. J. (2010). On the measurement of competency. Empirical Research in Vocational Education and Training, 2, 41–63.
Shavers, V. L. (2007). Measurement of socioeconomic status in health disparities research. Journal of the National Medical Association, 99, 1013–1023.
(∗)Shulruf, B., Hattie, J., & Tumen, S. (2008). Individual and school factors affecting students' participation and success in higher education. Higher Education, 56,
613–632. https://doi.org/10.1007/s10734-008-9114-8.
Sidhu, K. S. (2005). New approaches to measurement and evaluation. New Delhi: India: Sterling Publishers Private Limited.
Sirin, S. (2005). Socioeconomic status and academic achievement: A meta-analytic review of research. Review of Educational Research, 75, 417–453. https://doi.org/10.
3102/00346543075003417.
(∗+)Smith, E. (2016). Can higher education compensate for society? Modelling the determinants of academic success at university. British Journal of Sociology of
Education, 37, 970–992. https://doi.org/10.1080/01425692.2014.987728.
Stinebrickner, R., & Stinebrickner, T. R. (2003). Working during school and academic performance. Journal of Labor Economics, 21, 473–491. https://doi.org/10.1086/
345565.
(∗)Stratton, L. S., & Wetzel, J. N. (2011, May). The role of socioeconomic status when controlling for academic background in a multinomial logit model of six-year
college outcomes. Paper presented at the 51st annual forum of the association for institutional research, Toronto, ON.
Sullivan, A. (2001). Cultural capital and educational attainment. Sociology, 35, 893–912. https://doi.org/10.1017/s0038038501008938.
(∗+)Tai, R. H., Sadler, P. M., & Loehr, J. F. (2005). Factors influencing success in introductory college chemistry. Journal of Research in Science Teaching, 42, 987–1012.
https://doi.org/10.1002/tea.20082.
23
The World Bank (2012). Reviews of national policies for education: Tertiary education in Colombia 2012. Paris: OECD Publishinghttps://doi.org/10.1787/
9789264180697-en.
(∗+)Thiele, T., Singleton, A., Pope, D., & Stanistreet, D. (2016). Predicting students' academic performance based on school and socio-demographic characteristics.
Studies in Higher Education, 41, 1424–1446. https://doi.org/10.1080/03075079.2014.974528.
Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research. Review of Educational Research, 45, 89–125.
Tinto, V. (1993). Leaving college. Rethinking the causes and cures of student attrition. Chicago, IL: The University of Chicago Press.
Tracey, T. J., Allen, J., & Robbins, S. B. (2012). Moderation of the relation between person–environment congruence and academic success: Environmental constraint,
personal flexibility and method. Journal of Vocational Behavior, 80, 38–49. https://doi.org/10.1016/j.jvb.2011.03.005.
(∗)Triventi, M. (2014). Does working during higher education affect students' academic progression? Economics of Education Review, 41, 1–13. https://doi.org/10.
1016/j.econedurev.2014.03.006.
Van Ewijk, R., & Sleegers, P. (2010). The effect of peer socioeconomic status on student achievement: A meta-analysis. Educational Research Review, 5, 134–150.
https://doi.org/10.1016/j.edurev.2010.02.001.
(∗+)Walpole, M. (2003). Socioeconomic status and college: How SES affects college experiences and outcomes. The Review of Higher Education, 27, 45–73. https://doi.
org/10.1353/rhe.2003.0044.
(∗+)Wang, H., Kong, M., Shan, W., & Vong, S. K. (2010). The effects of doing parttime jobs on college student academic performance and social life in a Chinese
society. Journal of Education and Work, 23, 79–94. https://doi.org/10.1080/13639080903418402.
(∗+)Waqas, M., Abbasi, A., & Idrees, R. (2013). Academic performance of undergraduates in universities of Pakistan. Middle-East Journal of Scientific Research, 16,
55–61.
Warburton, E., Bugarin, R., & Nuñez, A. (2001). Bridging the gap: Academic preparation and postsecondary success of first-generation students (NCES 2001–153).
Washington, DC: National Center for Education Statistics, U.S. Government Printing Office.
Westrick, P. A., Le, H., Robbins, S. B., Radunzel, J. M., & Schmidt, F. L. (2015). College performance and retention: A meta-analysis of the predictive validities of ACT®
scores, high school grades, and SES. Educational Assessment, 20, 23–45. https://doi.org/10.1080/10627197.2015.997614.
White, K. (1982). The relation between socioeconomic status and academic achievement. Psychological Bulletin, 91, 461–481. https://doi.org/10.1037//0033-2909.91.
3.461.
(∗+)Win, R., & Miller, P. W. (2005). The effects of individual and school factors on university students' academic performance. The Australian Economic Review, 38,
1–18. https://doi.org/10.1111/j.1467-8462.2005.00349.x.
(∗+)Wolniak, G. C., & Engberg, M. E. (2010). Academic achievement in the first year of college: Evidence of the pervasive effects of the high school context. Research
in Higher Education, 51, 451–467. https://doi.org/10.1007/s11162-010-9165-4.
(∗+)Yanbarisova, D. M. (2015). The effects of student employment on academic performance in Tatarstan higher education institutions. Russian Education and Society,
57, 459–482. https://doi.org/10.1080/10609393.2015.1096138.
(∗+)Yao, G., Zhimin, L., & Peng, F. (2015). The effect of family capital on the academic performance of college students. A survey at 20 higher education institutions
in Jiangsu province. Chinese Education and Society, 48, 81–91. https://doi.org/10.1080/10611932.2015.1014713.
York, T. T., Gibson, C., & Rankin, S. (2015). Defining and measuring academic success. Practical Assessment, Research and Evaluation, 20, 1–20.
Yusuf, A. (2002). Interrelationships among academic performance, academic achievement and learning outcomes. Journal of Curriculum and Instruction, 1, 87–96.
(∗)Zheng, J. L., Saunders, K. P., Shelley, I. I., Mack, C., & Whalen, D. F. (2002). Predictors of academic success for freshmen residence hall students. Journal of College
Student Development, 43, 267–283.
Zlatkin-Troitschanskaia, O., Shavelson, R. J., & Kuhn, C. (2015). The international state of research on measurement of competency in higher education. Studies in
Higher Education, 40, 393–411. https://doi.org/10.1080/03075079.2015.1004241.
Zwick, R. (2012). The role of admissions test scores, socioeconomic status, and high school grades in predicting college achievement. Pensamiento Educativo. Revista de
Investigación Educacional Latinoamericana, 2, 23–30. https://doi.org/10.7764/pel.49.2.2012.3.
24

1 s2.0 S1747938X18302744 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S1747938X18302744 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S1747938X18302744 Main

Uploaded by

Copyright:

Available Formats

Educational Research Review 29 (2020) 100305

Contents lists available at ScienceDirect

Educational Research Review

Socio-economic status and academic performance in higher

2.1. Conceptualizing and measuring socio-economic status (SES)

2.2. Measuring academic performance in higher education

2.3. The relationship between SES and academic performance

2.4. Mediators of the relationship between SES and academic performance

4.1. Literature search and literature selection

4.2. Critical appraisal

Fig. 1. Conceptual diagram for the analysis of the mediators.

4.3. Analysis of the literature

4.3.1. Summative content analysis

4.3.2. Statistical meta-analysis

5.1. Measuring SES

5.1.1. Parental educational level

4 categories Stratton and Wetzel (2011)

Measurement Type of measurement Scale Operationalization Reference(s)

6 categories Yanbarisova (2015)

5.1.2. Parental occupation

5.1.4. Household resources

5.1.5. Neighborhood resources

Fig. 2. Authors' model for measuring SES.

5.1.6. Model for measuring SES

5.2. Measuring academic performance in higher education

5.2.1. Academic performance at university

4 categories De Clercq et al. (2017)

Achievement (n = 20) HSGPA (n = 14) Categorical (n = 4) 2 categories Smith (2016)

4 categories Stratton and Wetzel (2011)

5.2.2. Academic performance before university

5.2.3. Model for measuring academic performance in higher education

Fig. 3. Authors’ model for measuring academic performance in higher education.

Anderton et al. (2016) 414 Not signiﬁcant ATAR score Positive

Bahamón and Reyes Ruiz 68 Positive

Morlaix and Suchaut 543 Negative HSGPA, no Positive

Nguyen (2016) 616 Not signiﬁcant

5.3.1. Prior academic achievement

5.3.2. University experience

5.3.3. Working status

6.1. Conclusions and implications for practice

6.1.1. Measuring SES

6.1.2. Measuring academic performance in higher education

6.3. Implications for future research

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.edurev.2019.100305.

School Eﬀectiveness and School Improvement, 18, 45–65. https://doi.org/10.1080/09243450600797711.

You might also like