Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1–16 | 23 07- 6267 | DOI: 10.144 26/ js aa .v4i1.141
The contours of inequality: The links between
socio-economic status of students and other variables
at the University of Johannesburg
André van Zyl*
* Academic Development Centre, University of Johannesburg, South Africa. Email: firstname.lastname@example.org
The low level of student success in South Africa is an intractable problem, with levels of success
differing between the various groups that make up South African society. One of the major constraints
influencing student success involves the socio-economic status (SES) of newly entering students. In
the South African context, with its very high levels of SES inequality and other social stratifications,
a better understanding of issues related to SES would allow them to be addressed in targeted ways
that lead to improved student success. This study was conducted at the University of Johannesburg
and used data collected between 2010 and 2015. In this study, the SES of students was determined
by measuring their self-reported Living Standards Measure (LSM) level. The relationships between
the SES level and various socio-demographic variables were then tested using the chi-square test
with standardised residuals. The trends that emerged can assist institutions to gain a more nuanced
understanding of SES and its impact in the South African context. Three clear clusters emerged each
with their own distinguishing attributes and risk profiles.
Higher education, inequality, social stratification, transformation, University of Johannesburg, South
Students in South African higher education find it difficult to succeed. South Africa’s
combination of a low participation rate and a high dropout rate has been called a “low
participation, high attrition” system (CHE, 2013, p. 52). Not only are South African students
and institutions failing to create a situation in which students have a reasonable chance of
success, the net effect of the current situation is that only 5% of African and Coloured young
people are succeeding in higher education (CHE, 2013). This state of affairs is worrying and
has led to a lot of attention being focused on a variety of issues related to student success and
equity of outcomes. The terms Coloured, White, African and Indian in this study refer to self-
identified classifications according to nationally used equity criteria.
2 Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1-16 | 23 07- 62 67 | DOI : 10.14 42 6/jsa a.v4 i1.141
Many students who fail are poor and, as Scott, Yeld and Hendry (2007) point out, the
concept of student under-preparedness is often used to discount these poor students. This
simplistic view is, however, not tenable. Issues such as social capital, schooling and a lack
of career guidance are directly related to poverty and are known to play important roles in
determining student success. The divided and unequal state of the socio-economic status
(SES) distribution in South Africa has a crucial impact and, according to Walton, Bowman
and Osman (2015), crystallises in the student protests about funding on many South African
campuses. This also leads to a questioning of the concept under-preparedness (CHE, 2013,
p. 17) and an acceptance that “a gap can be closed from either side” (from the student/
societal or the institutional sides).
Schreiber, Leuscher-Mamashela and Moja (2014, p. vii) point out that the most
important modern theorist on student academic persistence (Hausmann, Schofield and
Woods, 2007), Vincent Tinto, links pre-entry attributes to student integration. They
further identify the whole idea of integration as especially important in a context with
“fragmented social structures” and “deepened social cleavages”. Tinto (2014) framed his
South African discussions as part of the Quality Enhancement Project by pointing out
that there is a performance gap between relatively rich and relatively poor students in the
USA and that this gap seems to be growing over time. As Tinto (2014, p. 6) stated during
his South African visit: “Providing students access without support is not opportunity.
Without support, academic, social, and financial, too many students do not complete their
programmes of study. It is my view that once an institution admits a student, it becomes
obligated to provide, as best it can, the support needed to translate the opportunity access
provides to success”. Walton, Bowman and Osman (2015) found that finance plays an
enabling role allowing students admitted to university also to succeed.
The link between the financial resources available to a student and his or her ability to
persist has been made by many researchers, including Astin (2005); Berkovitz and O’Quin
(2006); Isaak, Graves and Mayers (2006); Kreysa (2006); and Veenstra (2009). In the South
African context, the link between SES and student persistence has been confirmed by De
Beer (2006); Manik (2014); and Van Zyl, Gravett and De Bruin (2012). Reason (2009)
found that SES was the second most powerful predictor of student success (after previous
academic performance) in the United States. Poor students often have a combination of
factors that puts them at a higher than normal risk of non-completion (Van Rooyen, 2001;
Wessel et al., 2006).
Many authors, including Caison (2005) and Kuh et al. (2007), have found that poorer
students often do not have the necessary skills and support to manage on their own. These
students tend to have a variety of complex risk factors present in their background and
demographic characteristics (Johnston & MacLeod, 2004; Kuh et al., 2007). McLoughlin
(2012) and Williams, Leppel and Waldauer (2005), for example, identify SES as an
important factor in student career choice; lower levels of academic preparedness; general
academic performance; and ability to complete their studies. Lower-SES students are often
first-generation university entrants; have poorer high school education; and have access
to very low levels of financial support and other socio-cultural factors (Jones et al., 2008;
André van Zyl: The contours of inequ ality 3
Wessel et al., 2006). McLoughlin (2012, pp. 12–13) also suggests that low-SES students
experience higher education differently from their richer colleagues. This includes their
perceived ability to make friends and “fit in”, their experiencing pressure more acutely, and
the fact that they experience pressures to access basic necessities. Such students often also
lack the ability to make the necessary social links needed for academic success (Astin, 2005).
Breier (2010) points out that “financial constraints” in some contexts refer to
temporary financial problems with which institutions are often able to assist students. As
a result, internationally, low SES is often identified as a secondary cause for student early
departure and/or dropout. The concept “financial constraints” can, however, often have a
very different meaning depending on the context within which it is used. Breier (2010,
p. 669) uses the words “deprivation” and “extreme poverty” to indicate the deeper level
of financial constraints faced by students in the South African context. When someone is
poor in South Africa, it often means they do not have access to many relatively basic life
requirements. A lack of finances tends to impede their academic success more acutely and
the wide range of serious financial side effects might cause them to drop out at any point
during their academic career. Breier (2010) found that “financial constraints” have a greater
and a more continuous effect on poorer students in South Africa than on their richer
South Africa still suffers from deep economic fragmentation linked to the country’s
history, clearly illustrated in one of the highest Gini coefficients in the world (0.63 in
2011 compared to 0.41 in the USA). This deep level of poverty prevalent in South
African society is illustrated in the publication Poverty trends in South Africa (Statistics South
Africa, 2014). In this document, it was reported that 45% of the South African population
(approximately 23 million people) were classified as “poor” with 20.2% (10.2 million
people) living in “extreme poverty”. Not only is there an exceptionally wide division
between rich and poor (as reflected in the Gini coefficient), that division is still strongly
delineated according to race (Manik, 2014). This is illustrated by the fact that 54% of black
Africans are classified as poor and only 0.8% of Whites are so classified (Statistics South
Africa, 2014). According to Breier (2010), these patterns of poverty continue to reflect the
country’s racially divided past. This has led Letseka, Breier and Visser (2009, p. 25) to apply
the concept of “two nations” living simultaneously in South Africa to the South African
context. When students from the very poor SES groups enter university, they often struggle
to meet the basic financial requirements of university studies; any unforeseen expenses
exacerbate the problems they face.
It is therefore clear that many talented students in South Africa find themselves
constrained by finances and, as a result, unable to translate their potential into actual
performance. Yorke and Longden (2004) also point out that making progress in the area of
student success in a relatively poor country, like South Africa, is a far greater challenge than
in richer countries with more resources available to them. This makes it very important
to unpack the various socio-economic status levels by looking into their constituents. As
Reason (2009) points out, such an understanding would allow institutions the benefit of
being able to target interventions at specific sub-groups.
4 Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1-16 | 23 07- 62 67 | DOI : 10.14 42 6/jsa a.v4 i1.141
Research method and analysis
To address the above questions in the context of one university in the South African
context, this paper presents research conducted using a sample of 21 037 student responses
collected using the Student Profile Questionnaire (SPQ) between 2010 and 2015 at the
University of Johannesburg (UJ). UJ is a merged institution that came into existence in
2005, with four campuses, each with its own history. The demographic characteristics of
the participants in this study closely matched the demographic profile of the institution
(South African National Census 2011 data given in brackets; Statistics South Africa, 2012),
with 82.1% (79.2%) of the sample being black African; 3.5% (2.5%) Indian; 3.6% (8.9%)
Coloured; and 10.8% (8.9%) White. This is broadly representative of the demographic
profile of South Africa. A total of 9 011 (42.8%) male students and 12 026 (57.2%) female
students took part in this study and the four campuses and nine faculties of the university
were proportionally represented in the sample.
To investigate the relationship between SES and other variables, it was decided to use the
Living Standards Measure (LSM) instrument that was developed and refined by the South
African Advertising Research Foundation (SAARF) (Martins, 2006). The LSM measure
was used as a recognised SES measure to investigate links to other student data obtained
from the Student Profile Questionnaire. The original LSM instrument was created in the
1980s; the updated universal Living Standards Measure came into use in, and was refined
from, 2001 (Martins, 2006). The LSM subdivides the population into 10 LSM categories,
which, for the purpose of this study, have been grouped into five groups. The development
and testing of the SPQ is described in Van Zyl (2010) and van Zyl et al. (2012). The
relationships between 21 socio-demographic and academic variables with the five SES
levels created from the LSM scores were investigated. The LSM levels were typified as Low,
Medium Low, Medium High, High and Very High. Classification was done based on the
LSM level divisions as per the SAARF website (www.saarf.co.za).
Both variables in this study were categorical and, as a result, the chi-square test was
selected to investigate the statistical association between them (Agresti & Finlay, 2009). The
chi-square test assumes that no relationship between the variables exists and then tests that
assumption statistically. The Pearson chi-square value was used (Pallant, 2005) to test if a
statistically significant relationship between the two variables existed. While analysing the
cross-tabs, a significance level of p = < 0.05 was used in selecting significant variables and
Cramer’s V was calculated to determine the effect size of the variables.
The results obtained from the analyses for the whole group using LSM and selected
socio-demographic variables are shown in Table 1.
André van Zyl: The contours of inequ ality 5
Table 1: Chi-square results for all variables with LSM level
Variable c2df PCramer’s V
Gender 11.674 40.020 0.024
Population group 2 485.712 12 <0.001 0.199
Campus 236.303 12 <0.001 0.063
Did you visit a campus before coming to
15.876 40.003 0.027
Why are you studying? 49.010 28 0.008 0.024
Which role does your family play in your
54.740 16 <0.001 0.026
How easy will making friends be? 219.359 12 <0.001 0.059
Have you considered changing course? 45.976 8<0.001 0.033
Self-rated English level 799.421 12 <0.001 0.113
How many books were there in the house
in which you grew up?
1 592.084 20 <0.001 0.215
How many books have you read for fun in
the past year?
355.544 12 <0.001 0.117
Rate your English teacher’s English level 417.644 12 <0.001 0.130
For how many hours did you study at
112.952 16 <0.001 0.037
NBT Quantitative Lit. level 436.270 8<0.001 0.175
NBT Academic Lit. Level 486.653 8<0.001 0.188
Distance from campus 359.715 16 <0.001 0.065
Where will you stay? 585.630 20 <0.001 0.083
Are you worried about money stopping
your studies ?
3 868.258 4<0.001 0.429
How are you financed? 433.151 16 <0.001 0.072
Which level of education does the parent
with the highest level have?
288.936 20 <0.001 0.060
First-generation status 422.394 20 <0.001 0.071
Note: Statistically significant pre-entry attributes on the p ≤ 0.001 level shown in bold face
Chi-square results indicate a statistically significant relationship, but do not indicate where
within the variables the relationship resided. By calculating standardised residuals for all
instances where a statistically significant chi-square result was found, it was possible to
see where in the variables the relationship was located (see Table 2). The general rule for
standardised residuals is that an absolute value of 2 or greater (or –2 or less) implied that
there is a 95% chance that the variation had been caused by the one variable’s influence on
the other. Any standardised residual of 3 or more (or –3 or less) moved the level of certainty
up to the 99% level (Hinkle et al., 1988). A positive standardised residual indicated that the
6 Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1-16 | 23 07- 62 67 | DOI : 10.14 42 6/jsa a.v4 i1.141
observed frequency in that cell was higher than would be expected if no relationship was
found. A negative residual indicated that the cell had a lower frequency than would be
expected if no relationship existed.
Table 2: Standardised residuals LSM and socio-demographic factors
LSM level Low Medium
High Ve r y
White −8.1 −14.7 −14.0 −10.3 35.1
Indian −9.8 −4.8 −4.5 12
Coloured 4.6 −3.8 −4.1 2.0
African 2.1 7.5 6.9 5.0 −15.6
Campus 1 (City, degree focus) −3.3 −3.8 6.9
Campus 2 (City, diploma focus) 3.8
Campus 3 (Inner city, diploma) 3.3 −3.8
Campus 4 (City/informal) 2.0 4.5 4.9 −8.0
Because I really want to −2.4 2.2
Shows some interest, not very
3.2 2.0 −3.9
Very tough −3.5 3.8
Somewhat difficult 3.2 3.9 3.4 −8.5
Very easy −5.0 6.7
Considered changing course
No −2.0 3.1
Yes, but I did not change it −3.4
Yes, and I changed it to something
First language −9.1 −8.4 −7.3 17.1
Second language 6.7 6.0 5.5 −12.4
Books in house
None −11.3 −4.6 3.2 23.6
1−4.6 2.6 4.9
André van Zyl: The contours of inequ ality 7
LSM level Low Medium
High Ve r y
2 to 10 −8.4 2.9 5.0 6.1
11 to 20 3.4 −6.9
21 to 50 8.9 −3.7 −3.9 −7.7
More than 50 14.2 −2.1 −4.7 −8.2 −7.6
None −4.2 −4.0 −2.2 15.2
Fewer than 5 −5.0
Fewer than 10 −4.9
First language 9.9 −3.1 −4.5 −7.9
Second language (good) −9.4 3.5 4.9 7.0
Third language (reasonable) −4.6 4.4
Fourth (poor) 2.4
Previous study habits
Fewer than 5 hours per week 2.3 −3.8 3.9
Between 15 and 20 hours per
More than 20 hours per week 2.0 2.1 −6.0
NBT Quantitative Lit.
Basic 4.3 3.8 3.7 −10.8
Intermediate −2.7 5.9
Proficient −4.0 −5.1 −4.5 12.6
NBT Academic Lit.
Basic 5.6 4.3 3.6 −10.8
Proficient −4.6 −5.3 −6.6 14.4
Distance from campus
On campus −2.8 2.6
Within easy walking distance 4.8 4.2 4.2 −9.4
Less than 20 minutes’ drive away −2.4 3.6
Between 20 minutes and one
−3.4 −4.8 −2.4 7.7
More than one hour’s drive 2.1 2.7 2.0 2.7 −7.3
Place of residence
At home −6.4 −6.6 −6.3 12.9
Institutional accommodation −2.2 2.2 4.1 3.9 −5.1
Private accommodation (students
5.8 5.0 3.9 −10.9
8 Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1-16 | 23 07- 62 67 | DOI : 10.14 42 6/jsa a.v4 i1.141
LSM level Low Medium
High Ve r y
(students and non-students) −2.0
Not at home but with family or
Other −2.0 3.5
Worried about money
Ye s 27.0 −5.3 −19.9
No −41.3 2.7 8.2 30.5
Parents will pay −5.0 −4.9 −4.8 10.1
Loan 4.6 5.4 4.9 −10.9
Bursary −4.2 3.8 3.7 −2.4
Combination of answers 2.0
Some schooling (not Grade 12) 3.1 2.0 −2.5
Completed Grade 12 −3.0 2.5 2.5 2.2 −2.2
Fewer than three years of study
after Grade 12
A three-year qualification 4.0 −3.2
More than tree years of study after
−5.6 −6.4 −2.1 9.6
First in family 2.0 −3.0
Both parents to university −2.5 −3.3 3.5
Parents not but a brother or sister 5.9 4.1 −8.2
Many members of family attended −6.3 −6.4 −2.6 10.8
The results above contain a variety of interesting trends. Some confirm findings in other
studies and others (especially amongst the Low SES [LSES] and the Very High SES
[VHSES] groups) seem anomalous and require further investigation. Using the standardised
residuals to identify the details of the location of statistically significant relationships, three
main groups emerged from the results above: The Low SES group displayed a number of
distinguishing attributes (Group 1); the Medium Low SES (MLSES), Medium High SES
(MHSES), and High SES (HSES) groups have a lot in common (Group 2); and, the VHSES
group emerges as distinct in some ways (Group 3).
As was found by Manik (2014, p. 159), and confirmed in Groups 1 and 2 (as discussed
below), the various types of “deprivations” suffered by poor students were not mutually
exclusive and, as a matter of fact, tended to overlap. In the case of Group 3, the various types
André van Zyl: The contours of inequ ality 9
of advantage were also found to overlap. The findings also support the position of Visser and
Van Zyl (2013) with regard to the linking of population groups to relative advantage and/or
disadvantage in the South African context. Moreover, the findings support Kuh et al. (2007),
who found a statistically significant link between the finance methods students use and
The three groups that emerged from the analyses were, then, as follows:
Group 1: Low-SES students
The LSES group consisted mostly of African students who tended to congregate on specific
campuses of the University of Johannesburg. This group of students used a combination of
funding sources and in many cases they had to try any means they could to access the
required funds. As a result, this group tended to be worried that a lack of funding would
stop them from completing their studies. Such students were also less likely to be able to
access relatively costly institutional accommodation; as a result, they often had to travel for
more than one hour to get to campus. On a social level, these students tended not to have a
lot of parental support, confirming Modipane’s (2011) notion that relative socio-economic
status was linked with the likelihood that parents would support their children towards
academic success. Group 1 students also expected it to be difficult to make friends in the
new environment. Many of these students had to change their intended course of study at
a late stage – indicating that they are likely not to be enrolled for their first-choice course.
These students were also the most poorly prepared group academically, being more likely
to have an academic literacies (AL) and quantitative literacy (QL) National Benchmark Test
(NBT) score in the basic band (and less likely to be in the proficient band).
It is clear that students in this group have many risk factors and seemingly
insurmountable obstacles in their way, but they still manage to gain entrance to university.
The seemingly anomalous findings of this paper might give an indication of some of the
enabling factors that allow students to make this heroic leap. These factors include that
such students come from homes with many books, which is likely to indicate a reading
culture and a value placed on education. These students were also likely to have read a
number of books during the previous year and their parents seem to have tried to access
further education. Another enabling factor seemed to be that Group 1 students had been
taught English by someone who, in their perception, is an English first-language speaker. In
summary, a literacy culture and value of education at home and a good English foundation
seem to be enablers to get these very poor students into higher education.
Group 2: Medium-low, Medium-high and High-SES students
The second group consists of students from the MLSES, MHSES and HSES groups. This
“middle group” has a lot in common and tended to show very similar patterns. This group
consisted mostly of African students who tended to be distributed more evenly (less so
for MLSES) amongst the four campuses of the university. Socially, they expected some
difficulty in making friends, but they did not have a propensity to change their course at a
10 Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1-16 | 23 07- 62 67 | DOI : 10.14 42 6/jsa a.v4 i1.141
late stage. This group tended to come from homes with a moderate number of books and
they were likely to have read at least a few books in the previous year. Students were likely
to report that English was not their first language and that the main person who taught
them English was not an English first-language speaker. These students also tend to report
that they worked relatively hard at school, but they tended to be more likely to be in the
basic band (and less likely to be in the proficient band) for both the NBT AL and QL
tests. These students tended either to stay in institutional accommodation or in communes
relatively close to campus; they were less likely to stay at home. This meant that students
in this group tended either to be able to walk to campus or had to travel for more than
one hour to get to campus. Students tended to be less worried about money and they
either used a bursary or a loan to fund their studies. It is likely that many of these students
qualified for, and were able to access, National Student Financial Aid Scheme (NSFAS)
loans. The parental education of students in Group 2 tended to be up to Grade 12 level,
with few students having parents with more than a three-year qualification after school – as
a result, these students also tended to be first-generation university entrants.
Group 3: Very-high-SES students
The last group were from the VHSES group and tended to represent the privileged
minority. They were less likely to be African and were unevenly distributed amongst the
institutional campuses. Students in Group 3 tended to report that they wanted to study, but
contrary to the findings of Modipane (2011), their parents were not very involved in their
studies. Socially, students either expected it to be very easy or very difficult to make friends,
and they were not likely to have considered changing course. These students were much
more likely to be English first-language speakers, and they tended to be more likely to
score in the proficient NBT bands. Group 3 students tended to stay at home or on campus
and have a moderate (less than one hour) commute to get to class. On the financial front,
students tended not to be worried about money, and their main funding source was their
parents (they were less likely to use a loan or a bursary). These students also tended to come
from homes where higher education was something normal and where many members of
their family had gone before them.
The analyses of the VHSES group also contained some seemingly anomalous findings,
which put their seemingly strong position to succeed in higher education (explained
above) at risk. More so than expected, VHSES students reported having no or one book in
the house where they grew up and fewer than expected reported that they had more than
10 books (all categories). More students than expected in this group reported not having
read any books for fun in the previous year, and fewer than expected reported having read
10 or fewer books. More students than expected in this group reported having studied
for fewer than five hours a week at school and fewer than expected reported studying for
15 or more hours a week. These results seem to suggest that the advantaged background
of VHSES students allows them easier access to higher education, but at the same time
their poor literacy and study habits put them at risk of finding the transition into higher
education particularly difficult.
André van Zyl: The contours of inequ ality 11
Although it is difficult for institutions to address the financial problems that students
experience directly, detailed early advice may be one possible strategy to lessen the impact
of a lack of financial resources on student success. Students who anticipate the financial
struggles they could encounter before they arrive are a lot more likely to persist when
compared with those who are surprised by this challenge (Hawley & Harris, 2005, p. 133).
Although it is a well-known fact that SES in South Africa is unequally divided and still
strongly delineated along racial lines, these conclusions in themselves can often obscure the
truth about the challenges that students from different SES groups face. It is clear that both
poverty (and its effects) and wealth (and its effects) create very high levels of inequality in
an entering cohort, as well as in their experiences of higher education.
From these findings, it becomes clear that each of the broad SES groups brings its
own strengths and weaknesses to the higher education endeavour. Students from the very
low SES band come with many obstacles, but they also have unexpected strengths (such
as literacy-friendly home and school environments). These students are most likely to drop
out because of financial reasons and they seem to have less access to NSFAS than groups
that are slightly higher on the SES ladder. Students in the middle group tend to be less
worried about money and able to access external funding, but they seem to be less well
prepared from social background, schooling and academic perspectives. Students in the
VHSES group, on the other hand, seem to bring potential strengths, but their academic and
literacy habits as well as parental support and commitment seem to be lacking.
These different descriptions – drawn from the University of Johannesburg – clearly
show why a one-size-fits-all approach to student support and development will not work.
This also holds important implications for many other South African institutions, as their
student populations are increasingly representative of the country’s population. As is the
case with regards to many other attributes, students from different SES levels clearly have
different needs, and institutions of higher learning should customise their interventions to
the identified needs of these groups.
Agresti, A., & Fin lay, B. (2009). Statistical Methods in the Social Sciences. (4th ed.). New Jersey, USA:
Pearson International Edition.
Astin, A. (2005). Making sense out of degree completion rates. Journal of College Student Retention,
7(1–2), pp. 5–17.
Berkovitz, R.A. & O’Quin, K. (2006). Predictors of graduation of readmitted “at-risk” college
students. Journal of College Student Retention, 8(2), 199–214.
Braunstein, A.W., Lesser, M. & Pescatrice, D.R. (2008). The impact of a program for the disadvantaged
on student retention. College Student Journal, 42(1), 36–40.
Breier, M. (2010). From “financial considerations” to “poverty”: Towards a reconceptualization of the
role of finances in higher education student drop-out. Higher Education, 60, 657–670.
Caison, A.L. (2005). Determ inants of systemic retention: Implications for improving retention
practice in higher education. Journal of College Student Retention, 6(4), 425–441.
12 Journal of S tudent Aair s in Africa | Volume 4(1) 2016, 1-16 | 23 07- 62 67 | DOI : 10.14 42 6/jsa a.v4 i1.141
Counci l on Higher Education (CH E) (2013). A proposal for undergraduate curriculum reform in South Africa:
The case for a f lexible curriculum structure. CHE publication: Pretoria.
De Beer, K.J. (2006). Open access, retention and throughput at Centra l University of Technology.
South Afr ican Journal of Higher Education, 20(1), 33–47.
Hawley, T.H. & Harr is, T.A. (2005). Student characteristics related to persistence for first year
community college students. Journal of College Student Retention, 7(1–2), 117–142 .
Hausmann, L.R.M., Schof ield, J.W. & Woods, R.L. (2007). Sense of belonging as a pred ictor of
intentions to persist among A fr ican Amer ican and white first year college st udents. Research in
Higher Education, 48(7), 803–839.
Hinkle, D.E., Wiersema, W. & Jurs, S.G. (1988). Applied Statistics for the Behavioural Sciences. Boston:
Houghton Miff lin.
Isaak, M.I., Graves, K.M. & Mayers, B.O. (2006). Academ ic, motivational and emotional problems
identified by college students in academic jeopardy. Journal of College Student Retention, 8(2),
Jones, B., Coetzee, G., Bailey, T. & Wickham, S. (2008). Factors that facilitate success for disadvantaged
higher education students: An investigation into approaches used by REAP, NSFAS and selected higher
education institutions (Executive summary). Athlone, SA: Rural Education Access Programme.
Johnston, V. & MacLeod, L. (2004). Using a research-led approach to informing retention strategies for students
from low income and lower socio-economic groups. Internal Napier Un iversity publ ication. Retrieved
16 June 2007 from www.napier.ac.uk/qes/studentretentionproject/SRPhome.asp.
Kreysa, P.G. (2006). The impact of remediation on persistence of under-prepared college students.
Journal of College Student Retention, 8(2), 251–270.
Kuh, G.D., Kinzie, J., Buckley, J.A., Bridges, B.K. & Hayek, J.C. (2007). Piecing together the student
success puzzle: Research propositions, and recommendations. Association for the Study of Higher
Education (ASHE) Report, 32(5), 200 pp.
Letseka, M., Breier, M. & Visser, M. (20 09). Poverty, race and student ach ievement in seven higher
education institutions. In: M. Letseka, M. Cosser, M. Breier & M. Visser (Ed s), Student Retention
and Graduate Destination: Higher Education and L abour Market Access and Success, pp. 25–40. Cape
Tow n : HSRC Press.
Manik, S. (2014). Shifting the d iscourse: Student departure in the context of relative deprivations.
South Afr ican Journal of Higher Education, 2 8(1), 14 8–163.
Martins, J.H. (2006). Household cash expenditure by living standards group. Journal of Family Ecology
and Consumer Sciences, 34, 1–9.
McLoughlin, II, P.J. (2012). The transition experiences of high-achieving, low-income underg raduates
in an elite college environment. Journal of the First-Year Experience & Students in Transition, 24(2),
Modipane, M.C. (2011). Initial experiences of first entering students at the University of Limpopo:
Implications for coping with academ ic work/studies. South African Journal of Higher Education,
Pallant, J. (2005). SPSS Survival Manual: A Step by Step Guide to Data Analysis Using SPSS Version 12.
Maidenhead, UK: Open Universit y Press.
Reason, R.D. (2009). An examination of persistence research through the lens of a comprehensive
conceptual framework. Journal of College Student Development, 50(6), pp. 659– 682.
Schreiber, B., Leuscher-Mamashela, T. & Moja, V. (2014). Tinto in South Africa: Student integ ration,
persistence and success, and the role of student affairs. Journal of Student Affairs in Africa, 2(2), i–v.
André van Zyl: The contours of inequ ality 13
Scott, I., Yeld, N. & Hendry, J. (2007). A case for improving teaching and learning in South African
higher educat ion. Higher Education Monitor, 6, 1–86.
Statistics South A fr ica. (2012). Statistica l relea se (Revised): Census 2011. Retrieved 15 September 2015
fro m ww w.st a t s sa.g ov.z a /publ icat i ons/ P 030 14/ P 0301420 11.pd f .
Statistics South A fr ica. (2014). Poverty trend s in South Africa: An examination of absolute poverty
between 2006 and 2011. Retrieved 15 September 2015 from http://beta2.statssa.gov.za/
publications/Report- 03-10-06/Report- 03-10-06March2014.pdf.
Tinto, V. (2014). Tinto’s South African lectures. Journal of Student Affairs in Africa, 2(2), 5–26.
Van Rooyen, E. (2001). Die voorspelling van die akademiese prestasie van studente in ’n
universiteitsoorbruggingsprogram. South African Journal of Higher Education, 15(1), 180–189.
Van Zyl, A. (2010). The predictive value of pre-entry attributes for student academic per formance
in the South African context. Unpubl ished doctoral disser tation, Universit y of Johannesburg,
Van Zyl, A., Gravett, S. & De Bruin, G.P. (2012). To what extent do pre-entr y attributes predict f irst
year student academic performance in the South African context? South African Journal of Higher
Education, 26(5 ), 10 95–1111.
Veenstra, C.P. (2009). A strategy for improving freshman college retention. T he Journal of Quality and
Participation, January, 19–23.
Visser, H. & Van Zyl, D. (2013). Assessment of academic readiness to achieve student success and
retention. South African Journal of Higher Education, 27(2), 330–352.
Walton, E., Bowman, B. & Osman, R. (2015). Promoting access to higher education in an unequal
so cie t y. South African Journal of Higher Education, 29(1), 262–269.
Wessel. R.D., Bell, C.L., McPherson, J.D., Costello, M.T. & Jones, J.A. (2006). Academic
disqualif ication and persistence to g raduation by financial aid categor y and academic ability.
Journal of College Student Retention, 8(2), 185–198.
Willia ms, M.L., Leppel, K. & Waldauer, C. (2005). Socioeconomic status and college major: A
re-examination of the empirical evidence. Journal of The First Year Experience, 17(2), 49–72.
Yorke, M. & Longden, B. (Eds). (2004). Retention and Student Success in Higher Education. Maidenhead,
UK: The Societ y for Research into Higher Education, Open Universit y Press.