Cognitive ability is the result of interaction between genetic endowment and the environment. Current research highlights that education can be a central factor in the environment’s influence on cognitive ability (Baker et al., 2012, 2015; Barber, 2005; Downey, von Hippel & Broh, 2004; Ostrosky, Ardilia, Rosselli, Lopez-Arango & Uriel-Mendoza, 1998). Regardless of the specific content area being taught, what happens at school has an impact on neurocognitive development. Learning basic literacy, numeracy and other academic subjects, even only for a few years under basic conditions, leads to a number of cognitive enhancements resulting in schooled children thinking and reasoning in a significantly different manner in comparison to unschooled children. Immersed in an environment that prioritises cognitive abilities, whilst learning a specific set of skills at school, children’s scope and depth of cognition and aspects of their executive functioning are also fundamentally enhanced (Baker et al., 2012; Downey et al., 2004; Howard-Jones et al., 2012; Ostrosky et al., 1998; Schneeweis et al., 2014; Stevens & Bavelier, 2012).

Within this dynamic, studies have identified the impact of educational quality on cognitive ability and have specified that this aspect needs additional investigation (Donnelly, 2001; Luo et al., 2003). Furthermore, it is evident that there is significant investigation and debate surrounding educational attainment level and cognitive ability, but scant exploration on the relationship between type of secondary education (incorporating subject choice) accomplished and level of GCA (Kuncel et al., 2010). There is specifically a lack of research into the reasoning skills of adolescents and the role of subject-based content at the secondary educational level on the youth’s reasoning achievements (Ariёs et al., 2015). A need has been identified for research into what it is about prior learning at school that enhances subsequent cognitive abilities. Specifically, there is a call to investigate whether syllabi or different subjects employed at schools have a distinguishing impact on neurocognitive development, specifically the executive intelligence functions (Baker et al., 2012). Researchers have called for their findings into the relationship between mathematical achievement and GCA in young adults should be replicated in a large more diverse sample to further explore the extent of this association (Rohde & Thompson, 2007).

Taking into consideration the background and motivation for this study, there were four primary research questions underpinning the project:

Is there a significant relationship between the type of secondary education (incorporating subject choice) obtained and verbal reasoning ability?

One of the most critical management decisions is the appropriate selection of a candidate for a vacant position (Azar, Sebt, Ahmadi & Rajaeian, 2013; Grobler et al., 2011). Recruitment and selection are, however, costly human resource management exercises and errors in judgement can be extensive in terms of time, energy and money (Grobler et al., 2011; Lough & Ryan, 2010; Moore, 2006; Paterson & Uys, 2005). There is no set, typical and/or generally accepted human resource selection process and no two organisations conduct selection in the same manner (Louw, 2013; Van der Merwe, 2002). A widely used technique, the multiple-hurdle (or successive-hurdle) approach, results in the candidate pool becoming increasingly smaller after each stage in the selection process (Grobler et al., 2011; Nel et al., 2011; Noe et al., 2011). To assist in this decision-making process, various tools are used to assist recruitment and selection practitioners, including the initial screening of candidates’ curricula vitae, reviewing application forms, conducting interviews, carrying out assessment and testing, as well as doing medical and reference checks (Chan & Kuok, 2011; Grobler et al., 2011; Louw, 2013).

General cognitive ability is regarded as a valid predictor of both educational and vocational performance, providing valuable appraisals of creativity and career potential (Kuncel, Hezlett & Ones, 2004; Ng & Feldman, 2010). The reason for the first broad test of cognitive ability, developed by Binet in 1905, was to be able to predict individual differences in educational outcomes and research continues to underpin a moderate to strong association between cognitive ability and educational achievement (Deary et al., 2007). The use of psychometric tests as educational and occupational selection tools is hence justified on the basis of firm evidence that their scores accurately predict real-world success and have considerable value at both practical and theoretical levels (Deary et al., 2007; Foxcroft & Roodt, 2013). Given their usefulness in making employment decisions, GCA tests have been used in human resource selection for over 80 years (Outtz, 2002). Aptitude testing can denote the future performance of job applicants accurately. Psychological tests are considered as valuable instruments for two reasons. They can predict what the person is able to do currently as they reflect existing skills and knowledge, plus they are able to provide insight into how well an individual is likely to benefit from training and develop in the future (Kuncel & Hezlett, 2010). In a meta-analysis into the validity of GMA and specific cognitive abilities (i.e. verbal, numerical, perceptual and spatial) for predicting job performance and training success, both were found to be valid predictors of the two criteria across a wide variety of occupational groups (Bertua et al., 2005). Through the use of a multi-intelligence aptitude test battery, Tao, Chen and Chen (2009) concluded that an organisation can evaluate whether a job respondent has the required aptitude rapidly and effectively and use this as an important selection technique.

Psychological assessment in South Africa has come under close scrutiny over the past few decades, resulting in national labour legislation and regulations being passed by various stakeholder bodies after the 1994 elections (Nel et al., 2011; Nzama, De Beer & Visser, 2008; Paterson & Uys, 2005). The Employment Equity Amendment Act (EEAA) (2013) particularly indicates that employers may not reject candidates on the basis of a lack of skills or experience. Furthermore, this Act requires potential, the ability to acquire the required skills and experience within a reasonable time period, to be acknowledged in selection practices. Educationally disadvantaged learners are not a homogeneous group and psychological testing therefore needs to assist in accurately identifying between those who have potential but have been disadvantaged and those, although disadvantaged, who do not have potential (Murphy & Maree, 2006; Theron, 2007; Van der Merwe & de Beer, 2006).

Applicants undergoing psychological assessments are also protected by specific regulations governing the administration thereof. All assessment practitioners using psychological assessment procedures in the workplace must be registered with the Professional Board of Psychology and are required to adhere to a code of practice (Health Professions Council of South Africa, 2010; Mauer, 2000; SIOPSA, 2006). Adhering to these rules and regulations is of major concern to South African organisations using psychometric testing (Van der Merwe, 2002). The cost of using psychological assessments in the selection process is also a significant deciding factor (Bartram, 2004; Louw, 2013; Zibarras & Woods, 2010).

Another convenient measure used to assist managers in selection decisions is secondary education, including grade obtained, overall marks achieved and per subject, as well as the type of institution the secondary education was obtained at and the subjects included therein (Grobler et al., 2011; Louw, 2013; Sackett & Lievens, 2008). Certain professions or vocations show preference for certain types of secondary education. For example, artisans are traditionally selected from applicants who have a technical-type secondary education, which is obtained through a technical high school or a technical vocational education and training (TVET) institution (Schafmeister, 2013). A study researching the effectiveness and fairness of using matriculation and aptitude test results as predictors of further academic performance found that both have predictive validity (Van der Flier, Thijs & Zaaiman, 2003).

This study will assist human resource managers, recruitment and selection practitioners, or any individual involved in personnel selection to make informed decisions regarding potential employees with regard to their cognitive abilities and secondary education. The next section offers a literature review of the theoretical frameworks underpinning the study, incorporating a review of the empirical literature in the research field.

There have been numerous theories and definitions of GMA and GCA over the decades. The underpinning theory supported in this study is Carroll’s (1993) two-factor theory of intelligence, namely, fluid intelligence (Gf) and crystallised intelligence (Gc). Carroll revised and extended Cattell and Horn’s fluid-crystallised model, which is currently regarded as the most expedient model in both educational and business contexts (Hunt, 2014; Nisbett, 2013; Thorsen, Gustafsson & Cliffordson, 2014). Crystallised intelligence is defined as ‘consolidated knowledge … built out of the knowledge of the culture and out of the education of a subject’ (Ter Laak, Gokhale & Desai, 2013, p. 341). Fluid intelligence, in contrast, relies significantly less on stored knowledge, depends more on working memory and is involved in solving new problems (Nisbett, 2013; Ter Laak et al., 2013). According to Cattell’s investment theory, there is a causal relationship between Gf and Gc, with Gf being a broad knowledge acquiring ability and hence involved in all tasks related to knowledge attainment. Working memory, abstract reasoning and perceptual or processing speed are regarded as Gf measures (Colom, Escorial, Shih & Privado, 2007), whilst general knowledge, vocabulary and comprehension are instruments of crystallised intelligence (Furnham, 2010). This study explores three forms of Gf, namely verbal reasoning, non-verbal reasoning and visual perceptual speed.

There is voluminous literature supporting the extensive importance of GCA testing (Kuncel et al., 2004). The measurements of GMA and GCA are regarded as extremely useful instruments as they are reasonably effective predictors of school grades, occupational performance and several other aspects of achievement in life (Kuncel & Hezlett, 2010; Nisbett, 2013). It is purported that GCA is more than likely the single measurable human characteristic that impacts on almost every practical aspect of an individual’s life, from education to occupation, economic and social consequences (Furnham, 2008). CATs have revealed strong predictability of success in military and civilian occupational training, overall job performance, effectiveness in leadership positions, various indices of academic achievement as well as creativity assessments (Kuncel & Hezlett, 2010).

Whilst GCA has an impact on an array of outcomes, it is also influenced by many factors. Studies have found it challenging to clearly separate the relative importance of each of these factors as they are deemed interrelated and possibly even have a reciprocal impact on one another. Amongst other factors, GCA as measured through CAT performance is affected by age, gender, genetics, parental intelligence, personality, prenatal and immediate postnatal care, race and socio-economic status factors (Colom et al., 2007; Furham & Monsen, 2009; Leeson, Ciarrochi & Heaven, 2008; Nisbett, 2013). In the transformation of potential (Gf) into fulfilled potential (Gc), several elements are influential, including personal factors (i.e. interests, motivation and persistence) as well as the availability and quality of both formal (e.g. school) and informal (e.g. family and community) learning experiences (Soares et al., 2015; Thorsen et al., 2014).

Rather than being a static, predetermined construct endowed on individuals by virtue of their heredity and genetics, GCA is understood to be a malleable individual trait (Malmberg et al., 2008; Nisbett, 2013; Thorsen et al., 2014). Research has confidently exposed two contributing factors to the malleability of GCA: increased schooling and improvements to the curriculum (Hunt, 2014; Nisbett, 2013). Substantial research into the impact of schooling on cognitive functioning has confirmed that this environmental factor has shown an ontogenetic effect on individual contagion ‘making schooling into a neurocognitive developmental institution’ (Baker et al., 2012, p. 8). Education and academic training appear to modify the brain’s size and organisation. These changes may be either structural or the result of the redeployment of neural networks, which implies neural plasticity (Howard-Jones et al., 2012; Hunt, 2014). There is evidence of training causing changes in brain anatomy, such as the enlarged splenium of the corpus callosum and increased tissue volumes in multiple cortical regions during academic tasks such as reading, writing and language comprehension (Baker et al., 2015). A study revealed that, in the auditory and language processing areas, the white matter of the corpus callosum is thinner in illiterate people than in literate people (Ardila et al., 2010). Researchers established that targeted academic training for schooled children and adolescents can meaningfully alter the brain’s patterns of activity, and hence expertise in specific cognitive spheres. Following arithmetic training and the consequent heightened activation of the students’ brains, the attention and executive processing-rich zones of the frontoparietal areas became more automatic and efficient knowledge retrieval zones (Ischebeck, Laura, Karl, Michael & Margarete, 2007).

Whilst the neurobiological plasticity underpinning human cognitive development is certainly robust enough, there is growing evidence that its development is consistently tethered to the surrounding context. Environmental exposures and demands, such as education and academic training, are intervening and restraining variables that can influence neurocognitive growth (Baker et al., 2012). This is compounded by the appreciation that the number of years of schooling as well as the quality of education fulfils a substantial role in determining an individual’s intellectual capacity (Van Tonder, 2007). Within South Africa, this is a particular challenge as a weight of evidence supports the contention that there is an ongoing crisis within the country’s educational system. The South African educational system is noted as being exceptionally inefficient, severely underperforming and grossly unfair (Ramdass, 2009; Spaull, 2013). However, the focal undertaking of any educational system is to prepare the country’s young people effectively for the world of work (Van de Werfhorst, 2014).

It is clear from the literature review that there is evidence to support a link between cognitive abilities and education. The exact nature of that link and the implications thereof are less clear. In order to shed some light on this area, this research study aimed to investigate whether there is a relationship between the type of secondary education (incorporating subject choice) obtained and the applicants’ specific cognitive aptitudes and hence their selection for potential employment. The research sought to explore whether certain types of secondary education (incorporating subject choice) better prepared individuals for the world of work as seen in the results obtained through a selection process. It was believed that this could assist human resource practitioners to overcome some of the challenges inherent in implementing psychological assessment, specifically the EEAA (2013) stipulations.

Within this study, secondary education refers collectively to technical high schools, academic high schools and TVET institutions. The terminology ‘secondary education’, ‘Grade 12 qualification’ and ‘matriculation’ are therefore used interchangeably. Specifically, this study denotes five types of secondary education (incorporating subject choice) named and defined as follows:

A secondary education with mathematics and science as subjects is a Grade 12 qualification obtained through an academic high school.

The literature review leads to the study’s four hypotheses:

Hypothesis 1: There is a significant relationship between levels of verbal reasoning and type of secondary education (incorporating subject choice).

Hypothesis 2: There is a significant relationship between levels of non-verbal reasoning and type of secondary education (incorporating subject choice).

Hypothesis 3: There is a significant relationship between levels of visual perceptual speed and type of secondary education (incorporating subject choice).

Hypothesis 4: There is a significant relationship between types of secondary education (incorporating subject choice) and success in the screening process.

The quantitative research design and methodology employed in this study will subsequently be explained, followed by a presentation of the study’s findings.

The underpinning research philosophy for this research project was a positivist perspective. A descriptive quantitative approach was adopted as hard data, in the form of numbers, were collected and statistically analysed in this study (Blanche, Durrheim & Painter, 2006; Neuman, 2012). A cross-sectional convenience survey design, which involved standardised questionnaires, was implemented this data.

Descriptive statistics for the three aptitude subtests are displayed in Table 1. The medians for all three subtests were in the moderate (stanines 4–6) range with the verbal reasoning subtest’s median being slightly higher than that obtained for both the non-verbal reasoning and visual perceptual speed subtests. Table 1 also depicts that modes of seven, in the high range, were achieved for each of the three aptitude subtests.

Chi-square tests of independence were performed to examine the relationship between the type of secondary education (incorporating subject choice) and the stanine scores attained in the three aptitude tests, grouped according to low, moderate and high scores. Tables 2–4 summarise the applicants’ (N = 2463) results in the three aptitude subtests according to type of secondary education (incorporating subject choice) attained. Table 5 indicates the applicants’ overall success in the screening process.

The findings from the study’s four hypotheses led to the generation of two conclusions. These conclusions outline the different selection outcomes achieved by matriculants with different types of secondary education.

The first conclusion pertains to the recommended types of secondary education (incorporating subject choice) following the first phase of selection outcomes. Based on the results obtained, two types of secondary education (incorporating subject choice) are recommended, namely, the secondary education (gained through an academic high school) types inclusive of mathematics and science or just mathematics. Respondents with a Grade 12 qualification (gained through an academic high school) which included mathematics performed consistently better in comparison to the other four types of secondary education. Specifically, this type of secondary education inclusive of the mathematics subject outperformed the other matriculation types in non-verbal reasoning. The applicants with a secondary education (gained through an academic high school) inclusive of mathematics and science consistently performed well across the three aptitudes assessed. The common dominator in these two types of matriculation is the mathematics subject.

This conclusion is in line with the findings of several researchers who have highlighted the import of mathematical ability in academic achievement (Lubinski, Webb, Morelock & Benbow, 2001; Rohde & Thompson, 2006; Shea, Lubinski & Benbow, 2001). Mathematics is regarded as being essential for approximately 80% of the qualifications offered at a higher education level (Ramdass, 2009). Recent neurological research has established that enhanced abilities in mathematics may well influence cognitive abilities (Baker et al., 2015; Rohde & Thompson, 2007; Schneeweis et al., 2014; Stevens & Bavelier, 2012). This study has confirmed that the common denominator for a successful selection outcome as an automotive operator is an academic-type secondary education inclusive of the mathematic subject. In order to adequately cope with the various transformations and competitive challenges being experienced, the automobile industry therefore needs to focus specifically on academic Grade 12 qualifications inclusive of mathematics as a selection criterion for automotive operators.

The second conclusion pertains to the qualification types (incorporating subject choice) that are not recommended based on the selection outcomes of the first phase of this multiple-hurdle selection process. Whilst the secondary education (gained through an academic high school) inclusive of the science subject outperformed the other matriculation types in verbal reasoning, this matriculation type did not perform well in the non-verbal reasoning subtest. Applicants with the technical and general types of secondary education consistently performed poorly in the three aptitude subtests. It is therefore concluded that these three types of matriculation are not as successful in their selection outcomes as the two matriculation types (gained through an academic high school) inclusive of the mathematics subject.

The overall results of the study revealed statistically distinct relationships between the applicants’ type of secondary education and the three aptitudes assessed in this study. It is therefore concluded that the applicants’ type of secondary education was as effective as a predictor of potential as the cognitive aptitude subtests’ scores attained by these candidates. This conclusion highlights that secondary education can be used as a more time- and cost-efficient preliminary screening measure, specifically for entry-level technical positions. This finding is in contradiction to that of Berry, Gruys and Sackett (2006) who upheld that selection using CATs was a more efficient screening instrument than educational attainment.

The contribution of this research study is fourfold. Firstly, this study has confirmed the well-founded belief that standardised psychometric measures are efficient predictors of performance (Deary et al., 2007; Foxcroft & Roodt, 2013; Furnham, 2008; Kuncel & Hezlett, 2007; Rohde & Thompson, 2007; Watkins et al., 2007). Secondly, given the scant research on the relationship between the type of secondary education and cognitive abilities, this study adds to the human resource management field as the first study to successfully provide such data within a selection process for automotive operators. Thirdly, this study’s findings have accentuated the need to employ a multiple rather than a singular selection approach. This is of particular import in a context where the potential labour pool has been educated through a mass education system and where supply exceeds demand. These factors could plausibly lead to the inflation of credentials requiring employers to acquire additional information sources to discriminate between applicants (Kuncel et al., 2010; Van de Werfhorst, 2014). Fourthly, this study has highlighted that the assessment of aptitude versus using type of schooling achievements (incorporating subject choice) is a useful debate in the light of employment equity challenges. Secondary education could be instructive for human resource managers and practitioners in the screening processes they follow. The study’s findings highlight that through using a multiple-hurdle selection approach, the use of type of secondary education can be successfully employed as the initial hurdle. This could add significant value to human resource managers and practitioners as they seek to optimise their allocated resources. Using type of secondary education, rather than more expensive and time-consuming psychological tests, could save time and money. This approach thereby also alleviates the practitioner from the concerns associated with psychological testing and compliance with EEAA (2013) regulations.

This study was conducted as part of a large recruitment process completed by a South African automotive assembly plant. The sampling frame for this study was a database of prescreened potential operators. In comparison to the 2001 census, the sample is not in line with either provincial or national data pertaining to age, race or gender (Lehohla, 2004). This means that the main shortcoming of this research is that the results cannot be generalised to either the South African or global automobile industry.

A second limitation is that the study did not explore the non-g factors that may have had a substantial influence on both educational achievement and the cognitive abilities demonstrated in the findings. These factors include, amongst others, school attendance and engagement, interest, motivation and effort, the endowment of appropriate learning experiences, teaching quality and structure. These potential relationships should be explored in future research.

A third limitation could be the way in which the five types of secondary education (incorporating subject choice) were defined for this study. Whilst these types of secondary education are not referred to in literature, they were distinct categories of selection for the automotive industry concerned. It is recommended that future research considers whether there is a relationship between the subject marks obtained by applicants and the aptitudes assessed in this study.