Introduction

In the current competitive environment surrounding higher education, academic success of students is a growing priority for all universities. The experiences gained by first year university students are vital to establishing baseline knowledge, positive attitudes, self-confidence and commitment to study (Kuh 2001; Pargetter et al. 1998). For some students, commencing a degree is not easy and this can have a negative impact on academic success. Furthermore, poor academic success can lower self-confidence and self-esteem (McInnis et al. 2000). From an institutional perspective, poor academic outcomes can influence the reputation of a university as academic success is associated with institution quality (Price et al. 1991; Vivekananda et al. 2003).

For the purposes of this study, first year academic success was considered in terms of a student’s academic performance and retention. Numerous models and conceptual frameworks have been developed to explain academic success (Bean 1980, 1982; Bean and Metzner 1985; Evans 2000; Tinto 1975). Overall, academic success can be described as a complex process that involves the interplay of factors relating to the institution (e.g. support programs, type of degree) and factors relating to the individual student (e.g. demographic factors).

Academic performance

Academic performance is influenced by a student’s intake of knowledge and their ability to demonstrate and apply learned information. In the United States grade point averages (GPAs) are often used to measure performance, whereas in Australia weighted average marks (WAMs), percentage pass marks and GPAs are calculated (Dalziel and Peat 1998; Evans 2000; Everett and Robins 1991; Murray-Harvey 1993; Win and Miller 2004).

Student support programs, such as orientation programs and mentor schemes can yield substantial benefits to the first year experience of students and hence their academic performance (Vivekananda et al. 2003). These programs positively influence ‘student fit’ with the institution by teaching students about the campus, increasing the possibility of student involvement in campus activities, providing assistance to at-risk students and helping students cultivate behaviours necessary to succeed both academically and socially (Cabrera 2001; Gardner 2005; Higbee et al. 2002; Higgins 2004; Kuh 2001; McInnis et al. 2000; Murtaugh et al. 1999).

Degree preference is also an important factor to consider when looking at academic performance. McInnis (2002) noted that the number of combined degrees offered by Australian institutions has increased over time, although specific information about these degrees, in terms of their influence on academic success, is lacking. In Western Australia, university applicants nominate four preferences of the tertiary institution and degree into which they would like to be admitted. If a student is granted entry into a degree that is not his/her first preference, it is conceivable that the student’s interest in that degree may not be as high as a student who has received a first preference.

The relationship between matriculation scores and academic performance at university is well established (Abbott-Chapman et al. 1992; Dalziel and Peat 1998; DeBerard et al. 2004; Downes 1976; Everett and Robins 1991; Murphy et al. 2001; Pike and Saupe 2002). In Western Australia, Tertiary Entrance Rank (TER) is the matriculation score used by universities to measure a student’s academic ability. TERs range from zero to 99.95 and relate to how well a student has performed in his/her Tertiary Entrance Examination (TEE) relative to all other applicants that year. Once at university, it has been found that students with high TERs out perform students with low TERs (Everett and Robins 1991). The type of secondary school attended by a student has also been researched in relation to academic performance at university. Once at university, students from government secondary schools have been found to out perform students from non-government/private schools (Abbott-Chapman et al. 1992; Evans and Farley 1998; West 1985; Win and Miller 2004). West (1985) suggests that these differences may be due to disparities in the school system that require government school students to be more self-directed and responsible for their own learning as compared with non-government school students who are more likely to receive extra tutoring and coaching.

The influence of age and sex on academic performance is varied with some studies finding these variables to be important (Dalziel and Peat 1998; Graunke and Woosley 2005; McClelland and Kruger 1993; Murray-Harvey 1993; Ofori and Charlton 2002; Tay 1994; Win and Miller 2004) and others finding no significant associations (De Clercq et al. 2001; Hoschl and Kozeny 1997; Stacey and Whittaker 2005; Tutton and Wigg 1990; Walmsley 1990). Student characteristics such as socio-economic status, language spoken in the home and ethnicity have also been used to predict academic performance (Graunke and Woosley 2005; McClelland and Kruger 1993; Tay 1994). In Australia it has been found that Indigenous students are often less successful than other students in a tertiary setting, despite being highly focused on developing their talents and finding lectures more intellectually stimulating than other students (Hillman 2005; Krause et al. 2005; McClelland and Kruger 1993).

Retention

Retention can be defined in terms of the number of students who continue to be enrolled in a degree after a certain time period (e.g. one year). In Australia, the annual national higher education retention rates are generally between 70 and 80 per cent, with some variations by institution and faculty (Abbott-Chapman et al. 1992; Department of Education Science and Training (DEST) 2004).

According to Tinto (1975), academic performance while at university is the single most important factor in predicting student retention. Numerous studies have found a positive relationship between first year university academic performance and retention (Bean 1982; Huon and Sankey 2000; Krause et al. 2005; Murtaugh et al. 1999; Potts et al. 2003). A positive relationship between matriculation scores and student retention at university has also been reported (Abbott-Chapman et al. 1992; Arulampalam et al. 2004; Bean and Metzner 1985; Johnes and McNabb 2004; Murphy et al. 2001; Murtaugh et al. 1999; Potts et al. 2003).

As with academic performance, student support programs can positively influence the retention of first year students through the provision of knowledge, skills and socialisation opportunities (Zepke and Leach 2005). In addition, students who participate in mentor or orientation programs are more likely to be retained in their degree than non-participants (Campbell and Campbell 1997; Murtaugh et al. 1999).

The relationship between retention, sex and age is inconclusive. Some studies have found no relationship (Kirby and Sharpe 2001), while others have found females more likely to be retained (Arulampalam et al. 2004; Bradsen and Farrington 1986; Johnes and McNabb 2004), females more likely to discontinue (Abbott-Chapman et al. 1992), older students more likely to discontinue (Murtaugh et al. 1999; Scott 2004) and older students more likely to be retained (Johnes and McNabb 2004).

In the United States, the link between retention and ethnicity has been widely researched, with no clear pattern emerging (Bean and Metzner 1985; Gardner 2005; Murtaugh et al. 1999). Possible reasons for this include differences accounted for by socio-economic status, past academic achievement or institutional variations (Bean and Metzner 1985). In Australia and in New Zealand, Indigenous students have been found to be more likely to discontinue their degree than other students in first year (Hillman 2005; Scott 2004). According to both Tinto (1975) and Bean and Metzner (1985), students who come from families that are more educated, more affluent and more able to pay for their university education, are more likely to persist with their degree.

The Bachelor of Health Science degree (BHS) at The University of Western Australia (UWA) commenced in 2000. This course can be completed either as a single degree (four years full time) or as a combined degree with a Bachelor’s degree in Commerce, Economics, Law or Music (up to six and a half years full time). The BHS is a generic degree that incorporates both a public health and science major with business units. The degree is structured to prepare graduates for a health related career. As the academic success of students is a priority for all universities, the aim of this study was to investigate the first year academic performance and retention of BHS students at The University of Western Australia using the 2000 to 2005 cohort. The first year of university was specifically chosen as this is the time when the highest amount of academic failure and discontinuation occurs (Hillman 2005; McInnis et al. 2000).

Methods

Participants

The study consisted of 381 students and was a census of all full-time students who commenced the BHS between 2000 and 2005. Part time students were excluded from the study as it was acknowledged that factors influencing their performance and retention differ from full time students (Bean and Metzner 1985; Hillman 2005; Krause et al. 2005).

Design and procedure

This retrospective cohort study considered two outcomes, academic performance and retention. Academic performance was defined in terms of a student’s first year weighted average mark (WAM). A WAM is an indicator of overall academic performance and ranges from 0 to 100. Student retention was defined as the proportion of first year students that were still enrolled in the BHS the following year. If a student discontinued their course, transferred to a non-BHS degree or took a leave of absence, they were classified as ‘not retained’.

The factors considered for inclusion in the models of first year academic performance and retention are shown in Table 1. The independent variables included in the analysis were those for which data were available either from a university department or the student records system. As student psychological data are not collected by the University, the relationship between psychological variables, academic performance and retention was not considered in this study.

Table 1 Independent variables—name and description
Full size table

The analysis strategy comprised an initial descriptive investigation of the data. To increase the efficiency of the starting multiple regression model, single associations between each independent variable (Table 1) and outcome variable (performance or retention) were calculated. Single association variables with a conservative p-value of 0.20 were included in the starting regression model. A parsimonious final model was identified using a backward elimination strategy. Interaction terms considered plausible were investigated as part of the model. Finally, the model was assessed for multicollinearity. As academic performance is a continuous outcome variable a multiple linear regression analysis was used. For the analysis of retention, a logistic regression analysis was conducted as the outcome was dichotomous. All data were analysed using SAS for Windows (Version 9.1). Permission to undertake this study was granted by the UWA Human Research Ethics Committee.

Results

Table 2 describes the variables of interest for all 381 subjects. Overall, most students were enrolled in a single degree (61%), were female (73%) and were 18 years or less (68%). One half of all students did not select BHS as their first preference and most did not participate in the UWA mentor scheme (66%).

Table 2 Description of Study Participants—First year, Full time, Bachelor of Health Science Students (2000–2005)
Full size table

Academic performance

Single associations between academic performance and each independent variable were assessed for inclusion in the starting multiple linear regression model. Single association variables that attained a conservative significance level of p ≤ 0.20 were sex, language spoken at home, Indigenous status, type of secondary school, payment of university fees, matriculation score, type of secondary school mathematics course, type of secondary school English course, degree type and participation in the UWA mentor scheme.

When modelled together, matriculation score, Indigenous status, secondary school type, sex, payment of university fees, and secondary school English course were found to be significantly associated with first year academic performance. The final, parsimonious model is shown below in Table 3. A moderate amount of the variance (32%) in student marks could be explained by this model.

Table 3 Multiple Linear Regression Model—Predictors of academic performance (WAMs) in First year, Full time, Bachelor of Health Science Students (2000–2005)
Full size table

Overall, three interactions were considered plausible in terms of their possible modifying effect. These interactions were (1) sex and matriculation score, (2) sex and secondary school English course, and (3) degree type and matriculation score. None of the assessed interaction terms was statistically significant and therefore these terms were not considered further. Multicollinearity was also assessed and not found to be statistically significant.

According to the final model, a non-Indigenous female student, who attended a government secondary school, who previously completed secondary school English Literature, attained a high matriculation score and paid their university fees upfront could be predicted to gain a higher first year mark in the BHS at UWA than other types of students.

Retention

Variables that met the single association criteria (p ≤ 0.20) for inclusion in the starting multiple logistic regression model were sex, Indigenous status, degree preference, participation in the UWA mentor scheme and first year WAM group. Participation in the UWA mentor scheme and first year WAM group were found to be significantly related to retention in the first year of the BHS. As Indigenous status was so close to significance (p = 0.07) and due to the size of the odds ratio, this variable was included in the final model. The final, parsimonious model is shown below in Table 4.

Table 4 Multiple Logistic Regression Model – Predictors of Retention in First Year, Full Time, Bachelor of Health Science Students (2000 to 2005)
Full size table

Overall, four interactions were considered plausible in terms of their possible modifying effect. These interactions were (1) sex and participation in the UWA mentor scheme, (2) sex and first year WAMs, (3) degree type and participation in the UWA mentor scheme, and (4) degree type and first year WAMs. None of the assessed interaction terms was found to be significant and therefore these terms were not considered further. Multicollinearity was also assessed and not found to influence the analysis.

According to this model, a non-Indigenous student who participated in the university mentor scheme and who scored a first year WAM in the pass/credit range (i.e. 50–69.9) was more likely to be retained the following year in the BHS at UWA than other types of students.

Discussion and conclusions

Consistent with the literature, the factor of most influence on the academic performance of first year students was matriculation score and the factor of most influence on retention was first year marks.

Matriculation scores were found to be the strongest predictor of academic performance in first year students. This is an important finding as it highlights the importance of previous academic success. For those university staff in charge of student entry and quotas this finding is noteworthy as it highlights the importance of minimum entry requirement for a course. That is, if entry requirements are set too low, the institution may be setting students up for poor performance in first year. High matriculation scores and academic performance may also be a sign of maturity or organisation therefore more research is needed to better understand this factor.

Academic performance in the first year of university was found to be the strongest predictor of first year retention. The literature suggests that students with high academic performance are more likely to be retained at university than students with low academic performance (Bean 1982; Huon and Sankey 2000; Krause et al. 2005; Murtaugh et al. 1999; Potts et al. 2003). In this study, students with failing first year grades were more likely to discontinue than students with pass/credit grades, the most common reason for discontinuing being unsatisfactory academic performance. However, students with distinction/high distinction grades were found to be less likely to be retained in the BHS than students with pass/credit grades, the most common reason for withdrawal being to transfer to another degree. It is interesting that 25% of distinction/high distinction students left the BHS in their first year. This finding warrants further investigation to understand the reasons behind this decision, in particular the relationship between degree preference and the degrees into which the students transferred.

The UWA mentor scheme integrates first year mentoring with orientation activities. Students who participated in this program were twice as likely to be retained as students who did not participate. This is an important finding as it highlights the importance of this program with regard to first year persistence. This may also imply that a university has a direct means by which the retention of students can be positively influenced. When all factors were considered, academic performance was not found to be associated with participation in the UWA mentor scheme. However, it should be noted that student retention rather than academic performance is the primary focus of this program. As the UWA mentor scheme is voluntary it is possible that successful students seek out and enrol in such programs and that this factor is actually a measure of social inclusion or extroversion rather than program success. As a result further exploration of this factor is recommended.

The type of secondary school attended by students was found to be associated with academic performance. When all other factors were taken into account, students who attended government secondary schools were found to have higher marks in first year than students who attended non-government/private secondary schools. This finding is consistent with the literature (Abbott-Chapman et al. 1992; Evans and Farley 1998; West 1985; Win and Miller 2004). The type of English course completed by the student in secondary school was also found to be associated with academic performance. Most university courses require effective use of the English language in terms of literacy, understanding and communication (Jalili-Grenier and Chase 1997). Students who previously completed the secondary school English Literature course performed better than those who completed the secondary school English course. English as a Second Language (ESL) students were found to perform as well as English Literature students, however there were not enough ESL students in this study for results to be conclusive. The mathematics course previously completed by students was not found to be associated with academic performance. The type of mathematics and English course completed in secondary school was not found to be associated with retention. These findings will be of interest to people guiding students towards a career in health science as it demonstrates the benefit of certain secondary school courses. This finding also highlights a further avenue for research to assess which other secondary school courses are associated with success in this degree.

Few studies have measured the importance of degree preference or degree type on academic success. If a student is granted entry into a degree that is not his/her first preference, it is conceivable that his/her interest in that degree may not be as high as a student who has received a first preference. If a student is completing a combined degree and therefore a higher semester load than a single degree student they may not be able to commit as much time or effort to each unit. However, when all factors were taken into account, degree preference and degree type were not found to be associated with academic performance or retention.

In the literature, the relationship between academic success, sex and age was inconclusive. In this study, first year academic performance was found to be associated with sex, as overall, female students had higher marks than male students. DeBerard et al. (2004) suggest two reasons for gender related differences in academic performance at university, these being degree structures that are more suited to a specific gender or course imbalances in which student populations are predominantly male or predominately female. The latter is an interesting proposition as the BHS at UWA is predominantly composed of female students. In this study, first year retention was not found to be associated with sex or age.

Indigenous status and language spoken at home were used as measures of cultural and linguistic background. Language spoken at home was not found to be associated with either retention or academic performance. In Australia, Indigenous students are often less successful than other students in a tertiary setting (Hillman 2005; Krause et al. 2005; McClelland and Kruger, 1993). In this study, academic performance was found to be associated with Indigenous status, in which the first year weighted average marks of Indigenous students, were found to be lower than non-Indigenous students. In Australia and in New Zealand, Indigenous students have also been found to be less likely to continue with their degree than other students (Hillman 2005; Scott 2004). This was also found in this study, as Indigenous students were three times more likely to discontinue their degree than non-Indigenous students. These findings are cause of concern however it should be kept in mind that there were only small numbers of Indigenous students in this study, therefore further research should be conducted. Finally, the method used by first year students to pay for their university fees was not found to be associated with retention but was found to be predictive of academic performance. Results showed that those who paid their fees up front performed better than those who deferred their fee payment.

From an Australian perspective, results from this study could be generalised to students completing similar degrees, providing the Australian university they are attending is similar to UWA in terms of institution size and student population. The ability to generalise findings to overseas institutions is lower due to differences in admission policies and the campus experience of students. For example, overseas institutions, particularly American colleges, have open door admission policies compared with Australian admission policies which are mostly competitive and based on previous academic performance (McInnis et al. 2000). Differences also occur in terms of the residential nature of some overseas institutions in which a majority of students live on campus rather than commuting from home to university each day, which is the method used by most full time Australian students.

The strength of this study was that a census of all first year, full time, Bachelor of Health Science students from 2000 to 2005 were included in the analysis. Therefore, most of the usual sources of bias that occur in a cohort study (e.g. selection bias, loss to follow-up) were not experienced. Bias in terms of misclassification, may have occurred, but its effects are unlikely to be large. For example, non-differential misclassification may have occurred in the form of random data entry mistakes when information is input into the student records system. However, the possibility of this occurring was considered to be low, as information (e.g. student marks) are checked before they are finalised. Finally, many factors relating to the institution and to the individual student impact on academic success. As existing university data were used in this study, a limitation was that only those variables for which data were available were analysed. As a result, information about the psychological characteristics of students or participation in student clubs was not available and therefore not considered. Furthermore, no information relating to student’s overall life situation (i.e. work, living conditions, friends) were available for inclusion. Consequently, the final models of first year retention and first year academic performance established in this study may be incomplete. Further studies should therefore consider broadening the information collected and perhaps consider using a prospective study design.

Valuable information about the academic success of students was provided in this study. This information can be used to inform policy, for planning purposes and to assist in the development or review of first year student support programs. To extend the findings of the current study, further research could be conducted with other student cohorts such as second or third year BHS students. Research about part time students would also be informative. Finally, the information provided in this study is relevant to the operational priorities of any university. Therefore, these findings should be considered and used to improve academic outcomes such as performance and retention and therefore enhance the quality of the student learning experience.