The impact of self-efficacy beliefs on first-year accounting students’ performance: a South African perspective

ABSTRACT

This study measures the levels of self-efficacy beliefs to determine how this correlates with academic success in introductory tertiary accounting within a South African context. Also, self-efficacy beliefs are compared to determine if significant differences exist based on gender, academic language, type of study funding and different professional programmes studied. The study applies social cognitive theory and gathers quantitative data using questionnaires with statistical analysis to determine self-efficacy belief levels in a first-year accounting module. The means of several variables were compared via t-tests or ANOVAs. Hereafter, regression analysis was applied to determine if self-efficacy beliefs significantly impact academic performance. Findings indicate that most first-year accounting students felt confident in their self-efficacy beliefs. Significant differences were observed in students’ self-efficacy beliefs based on gender, academic language and type of programme studied, but not based on type of study funding. Several self-efficacy beliefs correlated significantly with academic performance, though differences were observed for different types of programmes. As limited consideration of the impact of self-efficacy beliefs within the discipline of accounting prevails in the literature, and even less so in South Africa, the study serves to enhance accounting educators’ understanding on how self-efficacy beliefs impact academic performance at the first-year level.

KEYWORDS:

• Academic performance
• accounting education
• self-efficacy beliefs
• South Africa

Introduction

Self-efficacy is the beliefs, thoughts and feelings individuals have regarding their personal capabilities, which affect how they function and, ultimately, how they perform (Bandura, 1977; 1997). The current study investigates the impact of self-efficacy beliefs in the accounting education context and aims to answer the following main research question: ‘How do self-efficacy beliefs correlate with academic success in an introductory tertiary accounting module?’.

Although the importance of the relationship between self-efficacy beliefs and their impact on academic performance have been widely advocated, the scope of research conducted within the discipline of accounting is relatively limited (Burnett et al., 2010; Byrne et al., 2014). The current study is therefore important in the quest for accounting educators to better understand the potential impact of students’ self-efficacy beliefs and, ultimately, how academic success in accounting studies can be improved.

The study is of interest to accounting educators around the globe as the findings could serve to deepen their understanding of the self-efficacy beliefs of accounting students, as well as the impact of such beliefs on first-year academic performance. A deeper understanding of these self-efficacy beliefs could guide the implementation of new interventions and instructional methods, such as a variety of workshops that could enhance self-efficacy beliefs, to improve the success of first-year accounting students.

Some students, and more specifically first-years, struggle to adjust in their transition from secondary to tertiary education, as the higher education environment requires self-efficacy through effort, persistence and perseverance (Basith et al., 2020; Brook & Roberts, 2021; Hussey & Smith, 2010; Joynt & De Villiers, 2019; Kahn & Nauta, 2001; Tinto, 1975; 1988; 1997; Umar & Bello, 2019). When self-regulatory learning processes (comprising factors such as goal setting, self-monitoring, self-evaluation, and strategy) are applied, self-efficacy beliefs can indeed provide students with a sense of agency, which motivates their learning ultimately (Abu Saa et al., 2019; Ahinful et al., 2019; Basith et al., 2020; Naidoo & Govender, 2021; Zimmerman, 2000).

Some studies have confirmed the positive correlation between motivation and academic performance (Afzal et al., 2010; Bardach et al., 2019; Basith et al., 2020; Beatson et al., 2020; Pintrich & De Groot, 1990; Struthers et al., 2000; Umar & Bello, 2019). Furthermore, evidence prevails that student motivation and learning are closely related to students’ self-efficacy beliefs (Almer et al., 1998; Brown & McCartney, 1998; Byrne et al., 2014; Christensen et al., 2002; Duger, 2021; Schunk, 1991; Zimmerman, 2000). Thus, the more students believe that their ability to learn can be attributed to forces they can control, the higher their motivation (and academic performance) would be (Mooi, 2006).

Self-efficacy beliefs include several conceptions individuals have about themselves, especially regarding the degree of confidence to perform a target behaviour, which is different from the type of role instantiated by performing the behaviour, or the degree of self-esteem ascribed to that role (Solberg et al., 1993). Lindsay (2014) describes self-efficacy beliefs as having confidence in oneself and in what one is able to do. High levels of self-efficacy indicates not only that a person possesses a specific skill, but also the belief that this skill can be demonstrated effectively under specific circumstances (Pajares, 1996; Schunk, 1991; Solberg et al., 1993; Zimmerman, 1995). For functioning at the higher education level, both Solberg et al. (1993) and Gore et al. (2006) view self-efficacy as the belief in one’s ability to engage successfully in higher educational-related tasks and behaviours. They further contend that the construct of self-efficacy could serve to deepen educators’ understanding of students’ persistence and performance at this level.

To answer the research question, the current study measures the levels of self-efficacy beliefs and determines how this correlates with academic success at three assessment levels (i.e. participation mark, examination mark and final mark) in an introductory accounting module at a university in South Africa. In addition, the self-efficacy beliefs are compared to determine if significant differences exist based on gender, academic language, type of study funding and whether a professional programme is studied.

Accounting education in South Africa faces fairly distinctive challenges (Baard et al., 2010; Du Plessis et al., 2005; Miller, 2020; Samkin & Stainbank, 2016; Van der Merwe, 2014), which provides for interesting context for the study conducted. Aspects such as academic language and study funding are delicate issues in the South African accounting education context. Students’ academic language (i.e. language of academic instruction provided by universities) are not necessarily the same as their first (or even second or third) language (Van der Merwe, 2014). The availability and source of study funding, which is linked to the history of South Africa and the resultant socio-economic inequality (Sesant et al., 2015), is another complication which might impact students’ self-efficacy belief levels and, ultimately, academic success (Papageorgiou & Callaghan, 2020). Students’ funding sources are often connected to their socio-economic background.

Another distinctive characteristic of the South African accounting education environment is its focus on directly preparing students for a number of professional designations. Professional accounting bodies include the South African Institute of Chartered Accountants (SAICA), the South African Institute of Professional Accountants (SAIPA) and the Institute of Commercial Forensic Practitioners (ICFP). International professional accounting bodies include the Chartered Institute of Management Accountants (CIMA) and the Association of Chartered Certified Accountants (ACCA). These professional bodies have differing educational and skills requirements (e.g. CIMA, 2019; SAICA, 2019). Consequently, each of these bodies attract a different profile of accounting student in terms of aptitude and interests. On the other hand, students who are not studying towards a professional accounting qualification, but nevertheless take accounting as a subject as part of a non-professional degree programme, may exhibit further differences in terms of their self-efficacy.

Contribution

The study contributes to the limited body of knowledge on the self-efficacy beliefs of accounting students. Understanding the relationship between self-efficacy beliefs and academic performance are of practical and managerial value to accounting educators and professional accountancy bodies to plan and implement enhanced interventions with a positive impact on first-year accounting students’ success within accountancy degree programmes. Contrary to other studies, this study is unique in that it measures the academic self-efficacy beliefs of a wide spectrum of first-year accounting students studying towards obtaining a professional designation at a variety of accountancy bodies or towards a non-professional degree programme in which accounting is simply one of the subjects.

The remainder of the paper is structured as follows: First, general literature on the relationship between self-efficacy beliefs and performance is reviewed. The theory which guided the research design is also stated and motivated. This is followed by a review of international and South African studies in which self-efficacy beliefs are considered to influence performance and success in accounting education and training. Hereafter, a description is given of the research methodology that was used. Finally, the empirical findings of the study are reported and discussed, followed by a summary of the key findings and conclusions.

Literature review

The relationship between self-efficacy beliefs and performance

The construct of self-efficacy needs to be embedded within a theory that explains a network of relationships among various factors (Bandura, 2006); thus, the current study is positioned in social cognitive theory (SCT). This theory, which was advanced by Bandura (1986) as an extension of his social learning theory, holds that an individual's knowledge acquisition can be directly related to observing others within the context of social interactions and experiences. Bandura (1997; 1999) positioned self-efficacy as a key component within SCT by highlighting that self-efficacy beliefs play a critical role in guiding human cognition, motivation, and behaviour. An integral concept of Bandura’s SCT is the principle of ‘guided mastery’, which can be described as building students’ inner confidence through taking ‘baby steps’ with the guidance of their educators.

Schoenfeld et al. (2017) used SCT to explain how accounting students’ career interests, goals and intentions are related to their self-efficacy beliefs and outcome expectations. As this theory posits that an individual’s self-regulated behaviour is driven by the variables shaping self-efficacy beliefs, self-efficacy is regarded as the conduit to understanding why individuals initiate certain behaviour, the effort they expend in engaging in this behaviour, and their persistence in facing obstacles (Gore et al., 2006).

Early studies in the fields of engineering and science (Brown et al., 1989; Lent et al., 1986; 1987; Siegel et al., 1985) identified that self-efficacy beliefs play a more significant role in predicting performance and persistence than traditional measures such as aptitude. Multon et al. (1991) found a moderate relationship between self-efficacy and performance and indicated that between 11% and 14% of the variance in academic performance and persistence could be accounted for by an individual’s academic self-efficacy beliefs. Results of a study by Beatson et al. (2018) showed a positive relationship between midterm examination results and student sense of self-efficacy at the end of an introductory accounting course. The relationship between students’ self-efficacy and their English language achievement was investigated by Mahyuddin et al. (2006) who concluded that students with high self-efficacy often display greater performance compared to those with low self-efficacy levels. In addition, various studies support the fact that the choice to engage in a task, the effort expended in performing it, and the persistence shown in accomplishing it, as well as the standard of the performance, are all affected by self-efficacy beliefs (Bandura & Schunk, 1981; Bouffard-Bouchard, 1990; Busch, 1995; Byrne et al., 2014; Mone, 1994; Mooi, 2006; Robertson & Sadri, 1993; Zimmerman, 1995). Moreover, the conceptual meaning and predictive validity of self-efficacy beliefs in relation to performance could vary depending on measurement timing (Mone et al., 1995).

Zimmerman et al. (1992) indicated that students’ efficacy beliefs regarding self-regulatory behaviours are related to their efficacy beliefs regarding academic goals and ultimate achievement. Thus, from an operational perspective, self-efficacy at the higher education level refers to students’ degree of confidence in their ability to successfully perform higher education-related tasks such as taking notes, asking questions and participating in discussions during lectures (Solberg et al., 1993). When investigated as a mediating variable in training studies, self-efficacy beliefs were found to be affected by improvements in students’ methods of learning (especially those involving greater self-regulation) and to be predictive of achievement outcomes (Zimmerman, 2000).

Chemers et al. (2001) confirmed that academic self-efficacy beliefs are significantly and directly related to academic expectations and performance and that highly efficacious students perceive themselves to be less stressed and coping better in adjusting to higher education. This is consistent with Bandura’s (2006) finding that the stronger the sense of personal efficacy, the greater the perseverance and the higher the likelihood that the activity will be performed successfully. Elias and Loomis (2002) argue that the strategy of measuring academic self-efficacy belief can be adapted to evaluate students’ confidence in their ability to successfully complete general undergraduate courses. Basith et al. (2020) confirm that students with high levels of confidence (i.e. self-efficacy) contribute to students’ success, one of many factors that affect students’ overall achievement. In support of the latter, Popa-Velea et al. (2021) found a significant association between low optimism, low resilience, high perceived stress and poor academic performance. Also, in a study conducted by Duger (2021) it was confirmed that Generation Z’s individualism, collectivism cultural tendencies and self-confidence levels have a significant and positive effect on their social-normative leadership motivation. Clearly, deduced from the aforementioned literature discussed, it is evident that a relationship exists between the level of self-efficacy and the successful performance of a specific activity.

Self-efficacy beliefs and performance within accountancy

Busch (1995) investigated gender differences in the self-efficacy beliefs and academic performance of first-year business administration students, but found no significant differences in the subject field of accounting. Christensen et al. (2002) studied the relationship between students’ beliefs regarding their abilities and their examination performance in a cost accounting course and found that students who initially had low outcome expectations subsequently performed better, while students with high outcome expectations performed more poorly. Thus, interestingly, the more conservative a student’s self-efficacy beliefs, the better the final performance was based on their study.

By applying students’ predictions of their grade as a measure of self-efficacy, Burnett et al. (2010) found that grade perception was significantly related to the students’ actual examination results. This confirms a positive relationship between self-efficacy beliefs and actual performance. More recently, Byrne et al. (2014) measured the self-efficacy beliefs of first-year accounting students in Ireland. Their analysis revealed that students lacked the confidence to fully participate in academic-related tasks such as taking notes, engaging in independent reading and studying effectively. They also found that 40% of the students were unable to judge the standard required to perform well in formal assessments. Based on these findings, they concluded that factors such as confidence in one’s ability to understand the content of the course, attempt questions in advance of tutorials and meet deadlines were positively associated with significantly better performance in accounting modules. Findings of Tepper and Yourstone (2018) show that introductory accounting students with similar entrance grades may outperform others owing to the effect of certain non-cognitive variables pertaining to self-efficacy. These included the students’ perceived skill level, tendency to become discouraged and expected performance relative to others. In Nigeria, Umar and Bello (2019) deciphered the relationship between accounting students’ the self-efficacy beliefs and their intentions and outcome expectations to become Chartered Accountants. The conclusion was that accounting student’s level of commitment and confidence is directly related to their intention of pursuing chartered accountancy as a career. In a study conducted amongst accounting students in Malaysia, the findings of Mansor et al. (2020) revealed that burnout has a significant negative relationship towards students’ academic performance, compared to perceived stress which has a significant positive relationship. An evaluation based on gender showed a negative significant relationship between gender and academic performance among students studying in the fields of accountancy and business (Al-Mutairi, 2010; Alhajraf & Alasfour, 2014).

The South African context

According to Bojuwoye (2002) and Fraser and Killen (2003), factors that affect first-year performance and success need to be investigated and understood properly, because they are linked to academic success. A number of South African studies have focused on the factors that affect success and performance of first-year accounting students at university level. However, most of these studies mainly applied biographical data to investigate student performance in the field of accounting. For example, Van Rensburg et al. (1998) focused on factors such as marks obtained in Grade 12 (final year of secondary school in South Africa) and previous exposure to accounting as a subject to determine their impact on the academic performance of first-year accounting students. Fraser and Killen (2003), as well as Baard et al. (2010), investigated the factors that affect the success of first-year accounting students at South African universities by investigating the relationship between biographical data, academic performance and class attendance. Similarly, Du Plessis et al. (2005) reported on the impact of factors such as South African students’ language proficiency in English, gender, motivation, prior experience in mathematics and accounting, as well as personality type, on the first-year accounting pass rate of distance learning students. The latter study was replicated for further validation by Müller et al. (2007) who confirmed that the variable of prior exposure to accounting had no significant impact on the performance of first-year accounting students. Bosua and Van der Nest (2015) investigated the impact of mathematical literacy at the Grade 12 level on the final marks of students taking a first year accounting module at a South African university and found no positive influence. More recently, Rossouw and Brink (2021) determined the impact of prior knowledge and concluded that Accounting as a subject at secondary school level is not necessarily essential in order to be successful in obtaining a professional Accounting degree in South Africa.

With regard to measuring the impact of self-efficacy beliefs on student performance, Steenkamp et al. (2009) used student perceptions to identify the variables that have the greatest impact on student success or failure in first-year accounting. Their results showed that poor class attendance, inadequate preparation, insufficient time and a lack of English tuition were perceived by students to be the main factors contributing towards poor performance. Joynt and De Villiers (2019) assessed the efficacy of a pre-university course in accounting and found it to have a positive impact on students’ academic performance. An investigation by Naidoo and Govender (2021) amongst cost and management accounting students at South African Universities of Technology revealed that the integration of General Education (i.e. the skills of written communication, oral communication, critical thinking, quantitative analysis, science, understanding and computer literacy) into accounting curricula positively correlated with the self-efficacy of learners and their confidence to be academically successful.

Based on the above literature overview, it is evident that a gap prevails with regard to self-efficacy beliefs as a measure and predictor of academic performance, particularly in the South African specific context when compared between different professional designation aimed accounting degree programmes. Specifically, research on the self-efficacy beliefs that impact students’ academic success, will contribute to the literature in this area.

Methodology

Research design

The current study was conducted within a positivistic paradigm and aimed to measure the self-efficacy beliefs and their impact on academic performance in accounting. The design was a quantitative study employing a questionnaire (which measured students’ perceptions of their levels of self-efficacy) to collect data, with statistical analysis to interpret results. Various quantitative studies have identified self-efficacy as a key variable associated with academic decisions and performance in the domain of social cognitive theory (Byrne et al., 2014; Lent et al., 1994; Ramos-Sanchez & Nichols, 2007; Schoenfeld et al., 2017; Zajacova et al., 2005) and as such this method was deemed suitable.

Measurement of self-efficacy

The relationship between self-perception and academic performance is in harmony with social cognitive theory, as elucidated earlier. Owen (1986) suggests that self-efficacy can be measured and used to assess a composite of affect, cognition and performance. The level of self-efficacy belief refers to the strength of perceived efficacy and depends on the difficulty of a particular task or required skill. Levels are measured by the amount of one’s confidence about performing a given task. These properties of self-efficacy judgments are measured using task-specific questionnaire items that vary in difficulty and capture degrees of confidence (Zimmerman, 2000).

Self-efficacy measures focus on performance capabilities rather than personal qualities. In addition, perceptions of efficacy depend on a mastery criterion of performance instead of normative or other criteria. When investigating performance at the higher education level, students’ perceptions are regarded as good sources of information, because they know their own situation well, know how other students think and feel, and have first-hand experience of the learning environment (Steenkamp et al., 2009).

Questionnaire design and validity

The questionnaire administered in this study was developed mainly from (and with the permission of) a questionnaire designed by Byrne et al. (2014) to measure the levels of self-efficacy beliefs of first-year accounting students in Ireland. This instrument was selected because of its proven reliability and because the objective of the current study and that of Byrne et al. (2014) was similar. However, in this study, Byrne et al.’s (2014) questionnaire was adjusted slightly for the unique South African accounting education context. A pilot study was conducted in the first semester of the course to ensure the reliability and validity of the adjusted questionnaire before formal testing commenced in the second semester.

The questionnaire contained 33 items, each referring to a different task-related academic activity. Respondents were required to indicate how confident they perceived themselves to be in their ability to perform each activity. These 33 measures were identified and selected after consultation with the lecturers responsible for first-year accounting in the various accounting degree programmes at a South African university. Bandura (2006) suggested that the activities used in a self-efficacy survey be formulated as statements beginning with ‘I can … ’ rather than ‘I will … ’. ‘I can … ’ is viewed as a judgment of capability, while ‘I will … ’ is a statement of intention. Consequently, the activities were phrased to follow the statement: ‘I feel confident in my ability to … ’, and were rated on a seven-point scale ranging from 1 (‘not at all confident’) to 7 (‘completely confident’) to measure the level of perceived self-efficacy.

The target population comprised 1057 students registered for first-year financial accounting semester modules in the second semester in different accounting degree programmes at a South African university (the same subject, but different modules). Ethical clearance was obtained from the university prior to data collection. Based on exposure to different assessment types at university (and thus different results could be expected), students were divided into three groups: (i) Group 1, professional accounting degree programmes leading to designations with SAIPA, ACCA and CIMA (ii) Group 2, professional accounting designations pertaining to SAICA and ICFP; and (iii) Group 3, non-professional (general) degree programmes of which accounting is included as a first-year subject (see Table 1).

Table 1. Respondents’ composition.

Respondent composition Professional degree programme	Group 1 N = 48	Group 2 N = 123	Group 3 N = 200	Total % (N = 371)
SAIPA/ACCA	32	–	–	9
CIMA/CGMA	16	–	–	4
SAICA	–	96	–	26
ICFP	–	27	–	7
Other*	–	–	200	54
*Other degree programmes of which first-year accounting is either an elective or compulsory module.
Target population composition Number registered per module	Group 1 % N = 233	Group 2 % N = 286	Group 3 % N = 538	Total % (N = 1 057)
Number of respondents as a percentage of the target population	21	43	37	35

All questionnaires were distributed via a secure web-based platform which could be accessed by any student registered for any one of the first-year accounting modules in the relevant programmes. Because self-efficacy judgments specifically pertain to future functioning, they should be made before students perform the relevant activities (Zimmerman, 2000). Thus, the questionnaires were distributed to students at the start of the second semester. In order to reduce the effect of respondent bias, the completion of the questionnaire was completely voluntary. Although respondents were required to provide their university student number (to enable comparison to actual academic performance), the risk of social evaluative concerns was addressed by indicating that all responses would remain confidential and that all data would be reported only on an aggregated basis, ensuring complete anonymity. To encourage respondents to provide honest answers, the survey clearly explained the purpose of the research, as well as the important contribution the research will make towards enhancing academic performance in first-year accounting. All data captured were analysed using SPSS and all interpretations were verified by a qualified, experienced an independent statistician. The specific statistical methods, tests and techniques that were used are elucidated in the ‘empirical research results’ section later.

Table 1 presents the respondent composition of 371 valid responses obtained, representing a response rate of 35%. Although the response rate seems low, the respondent pool still represented a wide range of students studying accounting degree programmes accredited with a variety of professional accountancy bodies (Group 1 and Group 2), as well as other degree programmes not accredited by professional accountancy bodies (Group 3). The levels of representation of the three groups in this study were consistent with enrolment levels at the particular university.

The demographic profile of the study participants is shown in Table 2. A fairly even spread of male (48%) and female (52%) students participated. From Table 2 it can also be observed that the majority of students that participated were white, Afrikaans-speaking students with private funding, which is somewhat reflective of the general student profile of the particular institution. The Broad-Based Black Economic Empowerment Act (South Africa, 2003) distinguishes between African, Coloured, Indian/Asian and White races. Some studies (e.g. Ali & Narayan, 2019; Solberg et al., 1993) highlight the nuance of ethnicity when considering self-efficacy beliefs of students. The context per Table 2 is therefore useful to keep in mind when considering results. Given the country’s history, discourse on socio-economic transformation, which is closely linked to race or ethnicity, is embraced in South Africa (Sadler & Wessels, 2019).

Table 2. Respondents’ profile.

Variable	Group 1 N = 48	Group 2 N = 123	Group 3 N = 200	Total % (N = 371)
Gender:
Male	25	55	98	48
Female	23	68	102	52
Race/ethnic group:
African/Black	4	20	32	15
Coloured	1	2	6	2
Indian/Asian	–	2	6	2
White	42	99	155	80
Other	1	–	1	1
Academic language:
Afrikaans	40	90	128	70
English	8	33	72	30
Type of study funding source:
Bursaries	6	18	27	14
Personal study loan	6	14	29	13
Private financing	33	75	128	64
Any combination of the above	3	16	16	9

Results and discussion

In agreement with other studies on self-efficacy beliefs (Byrne et al., 2014; Lent et al., 1986; Silver et al., 1995), the responses obtained were summarised in aggregate and interpreted as students’ self-perception relative to their ability.

Levels of self-efficacy beliefs

Table 3 indicates the levels of students’ confidence (self-efficacy scores) in aggregate for Groups 1 to 3 in respect of each of the 33 academic tasks/activities in the questionnaire. The table also presents the average percentages for each task based on a seven-point scale which are summarised in three categories, namely not confident (scores 1 to 3), unsure (score of 4), and confident (scores 5 to 7). A factor analysis was not performed as the authors chose to report the results of the 33 questionnaire items individually, so as to convey the full richness of gathered data through a descriptive statistical analysis. Surprisingly, the majority of students felt confident in their ability to perform all 33 tasks. When comparing all the tasks under the ‘confident’ category (i.e. scores ranging between 5 and 7 out of 7), the task to ask questions in lectures had the lowest average of 58%, while the task to meet all the requirements to successfully obtain the degree that I am registered for scored the highest average percentage of 91%. Interestingly, the task which scored the lowest average confidence percentage (i.e. asking questions in lectures) was the same task reported by Byrne et al. (2014). This is concerning as learning opportunities in the classroom environment may not be utilised optimally. In support of this finding, the results show that students preferred asking a classmate for assistance when they do not understand: The level of confidence to ask for help from my classmates if I don’t understand (task 3.32) scored an average of 90% under the ‘confident’ category, while the tasks associated with asking questions to persons other than peers, such as to ask questions in lectures (task 3.1), to ask questions during facilitation sessions (task 3.4), to ask for help from academic trainees (clerks) if I don’t understand (task 3.7), to ask for help from student facilitators if I don’t understand (task 3.10), and to ask for help from my lecturers if I don’t understand (task 3.11) all scored lower percentage levels under the ‘confident’ category that ranged between 58% and 72% where students indicated their confidence level ranging between a score of 5 to 7 out of 7, yielding lower confidence level percentages when compared to the confidence level percentage of 90% of task 3.32 to ask questions to peers. Large classes might be contributing to these low levels of confidence, as students might feel embarrassed asking questions in front of their peers. Also, syllabus overload combined with limited lecturing time (Miller, 2020) might contribute towards limited opportunities for students to ask questions or indicate to lecturing staff that they do not understand the material.

Table 3. Rating of respondents’ self-efficacy scores (N = 371).

	I feel confident in my ability to:	Mean^ (1-7)	Std. Dev.	NC** (1-3) %	U** (4) %	C** (5-7) %
3.1	Ask questions in lectures	4.52	1.69	24.3	17.5	58.2
3.2	Prepare sufficiently for lectures	4.67	1.44	20.5	18.9	60.6
3.3	Respond to questions asked in lectures	4.78	1.48	18.6	15.1	66.3
3.4	Ask questions during facilitation sessions	4.80	1.79	21.8	12.4	65.8
3.5	Approach academic trainee(s) (clerks) to receive feedback	4.84	1.70	19.9	14.3	65.8
3.6	Respond to questions asked during facilitation	4.88	1.61	18.1	12.7	69.3
3.7	Ask for help from academic trainee(s) (clerks) if I don’t understand	4.91	1.72	19.4	11.3	69.3
3.8	Make a good attempt to answer the facilitation questions in advance	4.95	1.57	17.3	14.0	68.7
3.9	Approach student-facilitator(s) to receive feedback	4.98	1.69	18.1	14.0	67.9
3.10	Ask for help from my student facilitators if I don’t understand	5.02	1.67	16.2	11.6	72.2
3.11	Ask for help from my lecturers if I don’t understand	5.14	1.62	17.3	10.2	72.5
3.12	Approach my lecturers to receive feedback	5.20	1.63	14.3	12.7	73.0
3.13	Draw up a study plan	5.41	1.52	10.8	8.9	80.3
3.14	Follow and make sense of the material being covered in lectures	5.44	1.27	9.2	5.9	84.9
3.15	Answer a discussion-type question	5.49	1.28	6.5	11.3	82.2
3.16	Apply my knowledge to solve previously unseen questions	5.54	1.26	5.9	10.0	84.1
3.17	Understand theoretical accounting principles in questions	5.57	1.18	6.2	7.5	86.3
3.18	Apply theoretical accounting principles in questions	5.59	1.17	6.2	8.1	85.7
3.19	Judge the standard required to get good marks in my exams	5.63	1.20	4.9	8.1	87.1
3.20	Make sufficient notes during lectures	5.67	1.37	8.9	6.5	84.6
3.21	Make sense of the theoretical/conceptual aspects of my accounting module	5.68	1.21	6.5	4.3	89.2
3.22	Engage in academic discussions with my classmates	5.70	1.37	9.2	4.6	86.3
3.23	Produce my best work in exams	5.72	1.23	5.9	5.4	88.7
3.24	Source and write up additional notes to support the material covered in lectures	5.75	1.27	7.3	4.0	88.7
3.25	Explain material covered to a classmate	5.76	1.30	6.5	6.5	87.1
3.26	Meet the deadlines for my homework	5.78	1.36	8.6	7.0	84.4
3.27	Plan my time to revise effectively for exams	5.78	1.16	5.9	4.9	89.2
3.28	Make sense of the material I read in my study material	5.79	1.12	3.8	6.5	89.8
3.29	Pass the first-year accounting module at the first attempt	5.80	1.46	6.7	8.9	84.4
3.30	Make sense of the feedback I receive regarding my tests/exams	5.86	1.15	4.3	6.5	89.2
3.31	Study effectively on my own	5.86	1.23	5.9	3.5	90.6
3.32	Ask for help from my classmates if I don’t understand	5.98	1.25	5.7	4.6	89.8
3.33	Meet all the requirements to successfully obtain the degree that I am registered for	6.14	1.10	2.4	6.2	91.4

^Seven-point scale: 1 = not at all confident to 7 = completely confident.

**NC = not confident (scores 1, 2 or 3), U = unsure (score of 4), C = confident (scores 5, 6 or 7).

Differences based on gender

Out of the 33 tasks, only six significant differences in the mean scores of male versus female students were identified, with p-values (Sig.) less than .05 (Pallant, 2013) (see Table 4). Male students were found to be more confident than female students in four of the six tasks, namely making sense of the theoretical/conceptual aspects of their accounting module, asking questions in lectures, judging the standard required to get good marks in their exams, and meeting all the requirements to successfully obtain the degree that that they are registered for. The last task was also the only task in the study by Byrne et al. (2014) for which male students were reported to have significantly higher levels of self-efficacy than females. Furthermore, in this study female students seemed to feel significantly more confident to make sufficient notes during lectures and source and write up additional notes to support the material covered in lectures. To a degree, these results correspond with other studies which also report general higher levels of self-efficacy beliefs in males than in female students (Subramaniam & Freudenberg, 2007).

Table 4. Levels of self-efficacy analysed by gender (N = 371).

	I feel confident in my ability to:	Male N = 178		Female N = 193
		Mean (1-7)^	Std. Dev.	Mean (1-7)^	Std. Dev.	Mean diff.	Sig. p-value*
4.1	Make sense of the theoretical/conceptual aspects of my accounting module	5.87	1.04	5.51	1.34	0.36	.00
4.2	Make sufficient notes during lectures	5.48	1.37	5.85	1.35	−0.38	.01
4.3	Ask questions in lectures	4.74	1.63	4.32	1.72	0.41	.02
4.4	Source and write up additional notes to support the material covered in lectures	5.60	1.29	5.90	1.23	−0.30	.03
4.5	Judge the standard required to get good marks in my exams	5.76	1.04	5.51	1.33	0.26	.04
4.6	Meet all the requirements to successfully obtain the degree that I am registered for	6.25	0.98	6.03	1.19	0.22	.05

^Seven-point scale: 1 = not at all confident to 7 = completely confident.

*p-value: < .05, indicates a significant result, assuming a random sample.

Differences based on academic language

To determine any significant differences in the level of self-efficacy based on academic language, a t-test was performed. In this test, students’ levels of self-efficacy belief were compared with regard to their academic language of choice, either Afrikaans or English. Academic language refers to the language in which students attend lectures, study and answer questions, which is not necessarily their native language. The university in this study has a multi-lingual language policy in which the same lecturers teach both Afrikaans and English classes for all first-year accounting modules (Groups 1 to 3). The results of this t-test in Table 5 only include the three tasks that were found to differ significantly based on their mean scores with p-values less than .05 (Pallant, 2013). For all three these tasks it was evident that students using English as their academic language seemed more confident to prepare sufficiently for lectures, to study effectively on their own and to ask questions during facilitation sessions. The latter is aligned with findings of Mahyuddin et al. (2006). A possible reason for the lower level of confidence among Afrikaans students could be ascribed to the fact that the majority of first-year textbooks are available in English only. Hence, students from an Afrikaans schooling background and using Afrikaans as their academic language may grapple with English textbooks. Besides preparation and study time for accounting, these students may be spending additional time on translation, interpretation and understanding of English material. This could be seen as an obstacle to sufficient preparation for lectures and effective, autonomous studying, although time management, self-regulation and planning skills might also be contributing factors.

Table 5. Levels of self-efficacy analysed for academic language (N = 371).

	I feel confident in my ability to:	Afrikaans N = 258		English N = 113
		Mean (1-7)^	Std. Dev.	Mean (1-7)^	Std. Dev.	Mean diff.	Sig. p-value*
5.1	Prepare sufficiently for lectures	4.56	1.45	4.93	1.41	−0.37	.02
5.2	Study effectively on my own	5.77	1.33	6.07	0.93	−0.30	.03
5.3	Ask questions during facilitation sessions	4.67	1.82	5.11	1.68	−0.44	.03

^Seven-point scale: 1 = not at all confident to 7 = completely confident.

*p-value: < .05, indicates a significant result, assuming a random sample.

Differences based on type of study funding

An analysis of variance (ANOVA) test was performed to determine whether the students’ level of self-efficacy was affected by types of study funding. The four funding options were (i) bursaries; (ii) personal study loans; (iii) private financing; or (iv) any combination of options (i) to (iii). No significant differences for any of the academic tasks were identified regarding any of the four funding options, with none of the mean score differences exhibiting a p-value of less than .05.

Differences based on type of degree programme

An ANOVA was performed to identify any significant differences in the levels of self-efficacy beliefs in students studying accounting degree programmes accredited by professional accountancy bodies (Groups 1 and 2) as opposed to students enrolled for accounting as a subject, but studying towards a non-professional degree programme (Group 3). Table 6 only lists the 13 academic tasks for which significant differences were observed among the three groups, all with a p-value of less than .05 (Pallant, 2013). Overall, based on the mean scores, students in Group 1 were found to be the most confident, whereas students in Group 3 were found to be the least confident in performing the majority of these 13 tasks. Eleven of these tasks scored the highest mean values for Group 1 and the lowest mean values for Group 3. The higher self-efficacy levels of Groups 1 and 2 could be ascribed to the fact that students in these groups are studying towards a professional accounting degree with accounting being their main subject area, compared to students in Group 3 who take accounting as a support module for a different programme. The rational expectation is that students in Groups 1 and 2 might have a greater interest in and are more comfortable with accounting-related educational activities, resulting in a higher confidence level to execute the tasks compared to students in Group 3 who were not pursuing a professional accounting qualification.

Table 6. ANOVA on self-efficacy scores between Groups 1 and 3.

I feel confident in my ability to:		Group 1 (N = 48)		Group 2 (N = 123)		Group 3 (N = 200)
		Mean (1–7)^	Std. Dev.	Mean (1–7)^	Std. Dev.	Mean (1-7)^	Std. Dev.	F	p-value*
6.1	Follow and make sense of the material being covered in lectures	5.85	0.77	5.58	1.00	5.26	1.48	5.48	.00
6.2	Pass the first-year accounting module at the first attempt	6.46	0.65	5.70	1.47	5.71	1.56	5.74	.00
6.3	Ask for help from my lecturers if I don’t understand	5.52	1.13	5.42	1.50	4.87	1.74	6.14	.00
6.4	Meet the deadlines for my homework	5.75	1.34	6.09	1.12	5.60	1.46	5.18	.01
6.5	Understand theoretical accounting principles in questions	5.96	0.71	5.68	1.10	5.42	1.28	5.00	.01
6.6	Apply theoretical accounting principles in questions	5.88	0.89	5.76	0.92	5.43	1.33	4.75	.01
6.7	Make a good attempt to answer the facilitation questions in advance	5.52	1.01	4.85	1.52	4.88	1.69	3.73	.02
6.8	Make sufficient notes during lectures	6.02	1.02	5.83	1.11	5.50	1.56	4.10	.02
6.9	Respond to questions asked during facilitation	5.44	1.27	4.93	1.55	4.72	1.69	4.00	.02
6.10	Ask questions during facilitation sessions	5.19	1.32	5.01	1.78	4.58	1.86	3.53	.03
6.11	Ask for help from my student facilitators if I don’t understand	5.60	1.22	4.97	1.68	4.92	1.73	3.46	.03
6.12	Engage in academic discussions with my classmates	5.69	1.34	5.94	1.16	5.55	1.47	3.27	.04
6.13	Source and write up additional notes to support the material covered in lectures	6.00	0.95	5.91	1.04	5.60	1.44	3.35	.04

^Seven-point scale: 1 = not at all confident to 7 = completely confident.

*p-value: < .05, indicates a significant result, assuming a random sample.

Furthermore, the comparison between Group 1 and Group 2, both representing professional accounting degree programmes accredited by professional accounting bodies, indicated that students in Group 1 had higher levels of self-efficacy beliefs for the majority of the tasks (11 out of 13) listed in Table 6. A possible reason could be the fact that the admission requirements for the degree programmes in Group 2 (i.e. SAICA and ICFP) are higher than those for the degree programmes in Group 1 (i.e. SAIPA, ACCA and CIMA) at the university where the research was conducted. Despite the fact that students in Group 2 are most likely academically stronger students compared to Groups 1 and 3, the Group 2-degree programme requires much more effort and persistence from Group 2 students in terms of their studies than that what is expected form students studying in Groups 1 and 3 degree programmes. The Group 2 degree programme is associated with higher levels of complexity, a more comprehensive syllabus, higher difficulty levels of assessment etc. The latter differences between the three degree programmes are well-known facts amongst students in Groups 1 to 3. Hence, in general students Group 2 students need to operate under more pressure in order to successfully transition and adapt to the greater challenges and higher demands required by the Group 2-degree programme. Therefore, Group 2 students could generally be more stressed (Popa-Velea et al., 2021), which results in elements such as fatigue and burnout that are negatively associated with self-efficacy levels (Mansor et al., 2020). Consequently, students in Group 2 might be less confident in performing the tasks, because they perceive these tasks to be more difficult and challenging while operating in a more pressurised and challenging degree programme environment. According to Zimmerman et al. (1992), the more capable students judge themselves to be, the more challenging the goals they embrace.

Correlation with academic performance

In the module studied, academic performance is measured based on marks. The participation mark encapsulates the weighted average marks of all formative assessments completed by students during the course of the semester up to the point in time when formal lectures ended and exam preparation commenced. The examination mark represents the mark for the final summative assessment written by students at the end of the semester. The participation mark and the examination mark each contributes a weighting of 50% to the final module mark. The module is passed if a final mark of 50% or more is achieved.

To gauge the extent to which the levels of self-efficacy beliefs explain the variation in students’ academic performance, Pearson correlation coefficients were determined and thereafter a stepwise multiple regression analysis was applied. Academic performance data were acquired from the university’s student academic record repository. The students’ academic performance was assessed using the participation mark accumulated during the course of the second semester and the examination mark at the end of the second semester, each contributing 50% towards the final mark for the relevant module. Firstly, the Pearson correlation coefficients for all 33 academic tasks listed in the questionnaire were determined for each of the academic performance variables, namely the participation mark, the examination mark and the final mark. Of the 33 tasks, 28 measures correlated with the participation mark, 25 measures correlated with the examination mark and 30 measures correlated with the final mark significantly (listed in Table 7). Only the significantly correlated variables were entered into a stepwise regression model for each of the academic performance measures.

Table 7. Self-efficacy scores significantly correlated with academic performance (N = 371).

		Participation mark (PM)	Examination mark (EM)	Final mark (FM)
7.1	Follow and make sense of the material being covered in lectures	0.228**	0.214**	0.243**
7.2	Study effectively on my own	0.256**	0.240**	0.273**
7.3	Respond to questions asked in lectures	0.259**	0.251**	0.281**
7.4	Meet the deadlines for my homework	0.202**	0.112*	0.171**
7.5	Produce my best work in exams	0.192**	0.151**	0.185**
7.6	Approach my lecturers to receive feedback	0.124*		0.121*
7.7	Approach academic trainee(s) (clerks) to receive feedback	0.107*		0.105*
7.8	Engage in academic discussions with my classmates	0.161**	0.146**	0.169**
7.9	Ask questions during facilitation sessions	0.109*		0.116*
7.10	Make sense of the theoretical/conceptual aspects of my accounting module	0.170**	0.157**	0.181**
7.11	Judge the standard required to get good marks in my exams	0.172**	0.149**	0.176**
7.12	Make a good attempt to answer the facilitation questions in advance	0.160**	0.131*	0.159**
7.13	Prepare sufficiently for lectures	0.118*		0.104*
7.14	Ask for help from my lecturers if I don’t understand	0.191**	0.151**	0.188**
7.15	Ask for help from academic trainee(s) (clerks) if I don’t understand		0.112*	0.114*
7.16	Ask for help from my student facilitators if I don’t understand		0.102*	0.107*
7.17	Explain material covered to a classmate	0.182**	0.191**	0.205**
7.18	Answer a discussion-type question	0.193**	0.159**	0.194**
7.19	Apply my knowledge to solve previously unseen questions	0.162**	0.203**	0.202**
7.20	Ask questions in lectures	0.166**	0.168**	0.185**
7.21	Plan my time to revise effectively for exams	0.224**	0.184**	0.223**
7.22	Make sense of the feedback I receive regarding my tests/exams	0.143**	0.108*	0.138**
7.23	Source and write up additional notes to support the material covered in lectures	0.135**	0.154**	0.159**
7.24	Make sufficient notes during lectures	0.132*	0.134**	0.146**
7.25	Respond to questions asked during facilitation	0.111*		0.111*
7.26	Understand theoretical accounting principles in questions	0.225**	0.150**	0.205**
7.27	Apply theoretical accounting principles in questions	0.229**	0.160**	0.213**
7.28	Make sense of the material I read in my study material	0.230**	0.190**	0.230**
7.29	Pass the first-year accounting module at the first attempt	0.199**	0.200**	0.219**
7.30	Meet all the requirements to successfully obtain the degree that I am registered for	0.162**	0.165**	0.178**

**significant at the 1% level.

*significant at the 5% level.

Finally, a stepwise multiple regression analysis was performed for Groups 1 to 3 for each of the three academic performance measures. The purpose was to determine which measures of self-efficacy significantly contributed to explaining variation in the students’ academic performance.

In the first regression model (see Model 1 in Table 8) the participation mark was the dependent variable for each group. No measures of self-efficacy could be entered in the model to predict participation mark significantly for Group 1. Two measures of self-efficacy were entered in the model to predict participation mark significantly for Group 2: understand theoretical accounting principles in questions; and respond to questions asked in lectures. These two measures combined explained 13.8% (R²) of the variation in the students’ participation mark performance and were highly significant. For Group 3, three measures of self-efficacy: make sense of the material I read in my study material; meet the deadlines for my homework; and respond to questions asked in lectures were entered in the model to predict participation marks significantly. These measures in aggregate explained 14.4% (R²) of the variation in the participation mark performance of students in Group 3. All three these measures were found to be highly significant at the 1% level for Group 3.

Table 8. Model 1 – Participation mark (dependant variable).

		Coefficient	Beta	t-statistic	p-value	VIF
Group 1 (SAIPA/ACCA, CIMA/CGMA): (N = 48)
8.1.1	None	–	–	–	–	–
Group 2 (SAICA and ICFP): (N = 123)
8.2.1	Understand theoretical accounting principles in questions	2.553	0.251	2.806	.006	1.118
8.2.2	Respond to questions asked in lectures	1.582	0.204	2.273	.025	1.118
R² = 0.138; Adj R² = 0.124; F-value = 9.599 (p-value = .000)
Group 3 (Other): (N = 200)
8.3.1	Make sense of the material I read in my study material	1.735	0.174	2.354	.020	1.256
8.3.2	Meet the deadlines for my homework	1.486	0.163	2.292	.023	1.161
8.3.3	Respond to questions asked in lectures	1.465	0.167	2.246	.026	1.266

R² = 0.144; Adj R² = 0.131; F-value = 10.971 (p-value = .000).

In the second regression model where the examination mark was used as the dependent variable (see Table 9) only one measure of self-efficacy was entered in the model to predict examination marks significantly for Group 1: study effectively on my own. This measure was found to be highly significant (p-value of .012) and explained 13% (R²) of the examination mark performance of students in Group 1. For Group 2, two measures of self-efficacy were entered to predict examination marks significantly for Model 2: responding to questions asked in lectures; and making sense of the feedback received regarding tests/exams. The first measure was found to be highly significant at the 1% level, while the second measure was significant at the 5% level. Together, these two measures accounted for a disparity of 17.4% (R²) in the examination marks of Group 2. Only one measure of self-efficacy, apply my knowledge to solve previously unseen questions, was entered to predict examination marks significantly for Model 2 in respect of Group 3, and was found to be highly significant at the 1% level, explaining 9.6% (R²) of the variation in examination marks of Group 3.

Table 9. Model 2 – Examination mark (dependant variable).

		Coefficient	Beta	t-statistic	p-value	VIF
Group 1 (SAIPA/ACCA, CIMA/CGMA): (N = 48)
9.1.1	Study effectively on my own	5.010	0.360	2.617	.012	1.000
R² = 0.130; Adj R² = 0.111; F-value = 6.849 (p-value = 0.012)
Group 2 (SAICA and ICFP): (N = 123)
9.2.1	Respond to questions asked in lectures	2.802	0.293	3.425	.001	1.060
9.2.2	Make sense of the feedback I receive regarding my tests/exams	3.479	0.236	2.762	.007	1.060
R² = 0.174; Adj R² = 0.160; F-value = 12.649 (p-value = 0.000)
Group 3 (Other): (N = 200)
9.3.1	Apply my knowledge to solve previously unseen questions	2.877	0.310	4.588	.000	1.000

R² = 0.096; Adj R² = 0.092; F-value = 21.051 (p-value = .000).

In the final regression model (see Table 10), the final mark (comprising the participation mark [50% weighting] and the examination mark [50% weighting]) was applied as the dependent variable for all the groups. Similar to Model 2, only one, and also the same, measure of self-efficacy was entered to predict final marks significantly for Model 3 for Group 1: study effectively on my own. Again, this measure was found to be highly significant, explaining 10.7% (R²) of the variation in the final mark. The same two measures of self-efficacy as for Model 2, respond to questions asked in lectures; and making sense of feedback received regarding tests/exams, were entered to predict final marks significantly for Model 3 for Group 2; again the first measure was found to be highly significant, and the second measure significant at the 5% level. These two measures combined accounted for 16.6% (R²) of the variation in the final mark performance of Group 2. For Group 3, two measures of self-efficacy were entered to predict final marks significantly for Model 3: respond to questions asked in lectures; and making sense of the material I read in my study material. Only the first measure was found to be highly significant with a p-value of .003, while the second measure was significant at the 5% level. These two measures explained 12.6% (R²) of the disparity in final marks for Group 3.

Table 10. Model 3 – Final mark (dependant variable).

		Coefficient	Beta	t-statistic	p-value	VIF
Group 1 (SAIPA/ACCA, CIMA/CGMA): (N = 48)
10.1.1	Study effectively on my own	3.704	0.327	2.347	.023	1.000
R² = 0.107; Adj R² = 0.088; F-value = 5.510 (p-value = .023)
Group 2 (SAICA and ICFP): (N = 123)
10.2.1	Respond to questions asked in lectures	2.371	0.289	3.365	.001	1.060
10.2.2	Make sense of the feedback I receive regarding my tests/exams	2.875	0.227	2.650	.009	1.060
R² = 0.166; Adj R² = 0.153; F-value = 11.980 (p-value = .000)
Group 3 (Other): (N = 200)
10.3.1	Respond to questions asked in lectures	1.754	0.222	3.038	.003	1.203
10.3.2	Make sense of the material I read in my study material	1.799	0.201	2.748	.007	1.203

R² = 0.126; Adj R² = 0.117; F-value = 14.226 (p-value = .000).

Conclusion: findings, limitations and recommendations

The study had set out to answer the following main research question: ‘How do self-efficacy beliefs correlate with academic success in an introductory tertiary accounting module?’. The study was thus designed to investigate the relationship between self-efficacy beliefs and academic performance of first-year university accounting students. More specifically, the study addressed the relationship between self-efficacy and gender, academic language, types of study funding, type of degree programme and academic performance.

In respect of gender differences, it was noted that male students’ levels of self-efficacy beliefs were higher for four measures of self-efficacy than those of female students. This may be a reflection of the current demographics of the South African accounting profession which is male dominated (SAICA, 2020). Regarding academic language, a relatively unique problem facing South African higher education is the fact that the country has 11 official languages, of which very few have been developed as academic or scientific languages. Moreover, textbooks are primarily available in English. Therefore, a mismatch between the academic language chosen by some students and the language in which course content (textbooks) is available does seem to have an impact on students’ self-confidence in respect of studying effectively on my own. On the other hand, the source of study funding does not seem to have a significant impact on students’ self-efficacy belief levels. This finding is reassuring, given the calls for free higher education in South Africa (Sesant et al., 2015), because it indicates that a lack of study funding alternatives might not necessarily have a negative impact on students’ levels of self-efficacy beliefs.

Overall, the regression models showed that the task, to study effectively on my own, had a significant impact on academic performance with regard to the examination mark and the final mark for Group 1. The fact that no measures of self-efficacy entered Model 1 for the participation mark might indicate that students in Group 1 had not worked effectively during the course of the semester, but only started studying intensively before the final examinations. For Group 2, it was noted that students who felt more confident to respond to questions in lectures seemed to perform better overall as this variable loaded significantly in all three models. This finding could indicate that students who are confident to respond to questions in lectures outperformed those who were less confident. A further task that loaded significantly for Group 2 was the students’ ability to make sense of feedback they received on tests and examinations. This finding could signal that their ability to positively respond to feedback might have boosted the students’ self-confidence, instead of discouraging them, which could, in turn, have contributed towards improved academic performance. For Group 3, similar measures of self-efficacy entered Models 1 and 3 in respect of the tasks to make sense of the study material they read and to respond to questions asked in lectures. This could indicate that, if students were able to make sense of their study material, they tended to be more confident to ask questions in lectures. In contrast, the only measure that entered Model 2 for Group 3 relates to the students’ ability to apply their knowledge to solve previously unforeseen questions. Thus, the more opportunities these students have to apply their knowledge to previously unseen questions, the larger their confidence seem to be to enhance their performance in final examinations.

The above findings addressed the research question through the identification and quantification of self-efficacy beliefs of first-year accounting students that are most significantly correlated with academic performance. It also highlights significant differences in self-efficacy beliefs based on gender, academic language, type of study funding and whether a professional programme is studied.

A common limitation associated with the measurement of self-efficacy is that the generalisation of findings may result in ambiguity about exactly what was measured, because the level of tasks and activities to be executed could differ between educational institutions with different module content, learning outcomes, circumstances and situational demands (Bandura, 2006). Therefore, the results reported in this study should be interpreted with due care, as the study was limited to students taking a first-year accounting module at a single South African university.

As mentioned in the literature review, the conceptual meaning and predictive validity of self-efficacy beliefs in relation to performance could vary depending on measurement timing (Mone et al., 1995). It is, therefore, submitted that a replication of this study at a different time during the academic year could result in different findings. However, the timing applied in this study was based on guidelines provided by prior studies that measured self-efficacy. Thus, the approach followed and the findings obtained are held to be valid, reliable and a valuable contributor towards the current body of knowledge on measuring the impact of self-efficacy beliefs on the academic performance of accounting students in South Africa.

Mone et al. (1995) suggest that academic performance could be significantly affected by changes in self-efficacy, rather than by changes in self-esteem. It is, therefore, recommended that the performance of first-year accounting students be improved by implementing self-efficacy interventions, instead of self-esteem interventions. These interventions need to foster a culture in which students are motivated to ask questions and feel comfortable to seek additional assistance from lecturing staff. Furthermore, it is recommended that students be better equipped in respect of their verbal communication, time management and self-regulation abilities, as this should develop them into more confident, self-regulated learners. Coherent with Bandura’s principle of ‘guided mastery’, this could be achieved through courses, interactive workshops and motivational speaking sessions in which students are encouraged to respond positively to feedback on their academic performance. Goal structures (i.e. classroom practices that emphasise performance goals to enhance motivation) are also important, as they may be positively related to self-efficacy beliefs of accounting students (Bardach et al., 2019). Further research could be conducted on the extent to which verbal persuasion can change the level of self-efficacy in students.

In light of the findings of this study, it is recommended that accounting educators pay particular attention to the structuring of study material, as well as the way feedback is given on tests and examinations, to ensure that all students are able to make sense thereof. In addition, students should be given more opportunities to apply their knowledge to previously unseen questions. Another aspect that seems to underpin some of the challenges experienced by students, is the fact that many students do not study in their first language. More research is required in this area to identify possible solutions to the problem (which may or may not include further development of African languages in the accounting sciences).

Based on the findings, it is clear that levels of self-efficacy beliefs might have a significant impact on academic performance, as explained by social cognitive theory. Therefore, because self-efficacy belief can affect students’ engagement in and approach towards learning activities, it needs to be managed and improved through appropriate interventions. This study contributes to the extremely limited evidence of the link between self-efficacy and academic performance in the uniquely South African accounting education environment and can also assist in better understanding some of the nuances affecting transformation of the profession through education. Finally, the findings of this study contribute as it assists in identifying areas where self-efficacy is potentially problematic with regards to specific academic-related tasks required to be performed by first-year accounting students and its impact on such students’ academic performance. Understanding the relationship between self-efficacy beliefs and academic performance are of practical and managerial value to accounting educators and professional accountancy bodies to plan and implement enhanced interventions with a positive impact on first-year accounting students’ success within accountancy degree programmes.

Disclosure statement

No potential conflict of interest was reported by the author(s).