Who leaves and who stays? Retention and attrition in engineering education

Abstract

At a time of high demand for engineering graduates, the mean graduation completion rate of engineering undergraduates in Australia has been identified as approximately 54% (with considerable variation across institutions and sectors). Such a proportion of non-completions has been viewed as an excessive loss to the qualified workforce of Australia. Broad brush, government-collected statistics do not, however, provide the level of detail required to understand who leaves, when and why they leave and where they go. This paper reports on a pilot study undertaken to precede and inform final decisions on research design and methodology for a multi institutional project seeking to understand and reduce student attrition from engineering degrees across Australia. The aim of the project is to produce guidelines on curriculum formulation and delivery strategies to reduce attrition from engineering programmes while meeting course outcomes.

The pilot study was conducted at an institution which has a relatively diverse range of students (a high proportion of whom study part time) and engineering degree structures incorporating traditional and internship-based degrees. Results from a cohort analysis which tracked pathways to completion or non-completion of the degree for the cohorts from two specific entry years are presented. From this analysis, groups of students who “persisted over long periods”, “switched to another degree” or “withdrew from the university” were identified and interviewed. Their experiences and stories formed an essential pathway to a better understanding of the dynamics of retention/attrition and factors which required further investigation before the multi institutional study began.

Introduction

The consultative review of the Australian engineering education system, Addressing the quality and supply of engineering graduates (King, 2008), undertaken in 2007–2008 by the Australian Council of Engineering Deans, reported that the demand for engineering graduates in Australia and globally is increasing, with an estimated shortfall of 20,000 engineers in Australia alone. It also noted that the mean graduation completion rate of engineering undergraduates during the period 2001 to 2006 was approximately 54%. This corresponded to a mean annual course success rate in the range of 0.77 (domestic part-time male students) to 0.908 (international full time females) and mean course retention rates of 0.653 (international part time males) to 0.923 (international full time females). This loss of approximately 46% of the commencing cohort was not only viewed as an excessive loss to the qualified workforce but also, in the case of domestic students, a loss of return on public investment. The aggregated attrition data for engineering presented in the review confirmed data from several Australian higher education reports (Krause et al., 2005; Marks, 2007; Olsen et al., 2008) but provoked disquiet and considerable discussion which focused on the need for a better understanding of the impact on retention of institutional differences, new degree structures, part time study, the introduction of “learning spaces” and targeted retention strategies.

An overarching project aimed at underpinning and strengthening Australia’s education system for engineering education through informed and systematic curriculum renewal and delivery support has been funded by the Australian Learning and Teaching Council. One of the four themes in this project specifically addresses understanding and reducing attrition. The purpose of this project theme is to gain a better understanding of who leaves, when and why they leave and where they go to, in order to identify strategies that can be implemented within university administration, curriculum specification and delivery that will increase retention to completion of the qualification. The overall research question for the attrition project theme could therefore be framed as how could Australian engineering degree granting institutions and their staff adapt their current processes and practices to improve retention of students to completion of the engineering degree? One of the proposed outcomes is the production of guidelines on curriculum formulation and delivery strategies to reduce attrition from engineering programmes while meeting outcome standards.

The first phase of the project was a best evidence synthesis of literature on retention/attrition internationally in engineering education, and in the higher education sector within Australia. Initial analyses of attrition and completion data using Department of Education, Employment and Workplace Relations (DEEWR) statistics were also part of this initial scoping exercise.

This exercise revealed that for engineering education in Australia:

• Attrition is higher in engineering than in other professional disciplines such as medicine and veterinary science, but lower than more open entry degrees such as sciences or arts

• Attrition is higher for part time students than full time students

• There is less migration into than out of engineering from other degrees

• Attrition is higher for domestic students than international

• Attrition for the 19–23 year age group is the lowest of the age groups

• Attrition varies markedly by institutional characteristics such as status, admission score, urban/regional

• On average, attrition and failure rates are higher for male students than for female students (although female participation in engineering averages nationally is less than 15%)

• Average retention is 85% per year which would result in 52% still enrolled or eligible to graduate after four years.

The second phase of the project is an in-depth analysis at each of two case study institutions with the aim of piloting models and processes and highlighting issues related to attrition to be explored in the third phase of the project. One institution is a traditional, highly-ranked, research-intensive university. The other is one of the Australian Technology Network (ATN) universities which has incorporated more varied degree structures. This paper reports on the initial results from the ATN university, hereafter known as ATN1.

The third phase of the project will be to extend the cohort analysis model and instruments developed in the pilot studies to a selection of ten engineering schools with common and contrasting characteristics. Similarities and anomalies in retention/attrition patterns revealed by these cohort analyses will be further explored by the individual institutions.

What do we know about retention?

Retention, persistence and completion in post-school education have been the focus of increased attention internationally in recent years. Reports from the USA (Hauptman, 2008), the UK (National Audit Office, 2007; van Stolk et al., 2007), New Zealand (Scott, 2005) and Australia (Olsen et al., 2008) demonstrate that governments throughout the Western world increasingly expect improved learning outcomes for money spent on post-school education.

Surveying the wealth of literature on student retention within engineering, Heywood (2002) provides an excellent synthesis and review for the preceding twenty years. He particularly highlights the theoretical foundations and motivation for this focus on student retention in both the USA and the UK, also noting the cyclic and unchanging nature of many of the issues.

An annotated bibliography on attrition in engineering education, recently completed for this project, has revealed that the majority of the research literature has focused on investigating causes of attrition and predicting students at risk of dropping out using statistical analyses, surveys and qualitative studies, rather than evaluating intervention strategies. Such studies have sought to identify candidates with the capacity and motivation to complete a degree course and practice engineering (Bernold and Anson, 2007; Besterfield-Sacre et al., 1997; French et al., 2005, Matusovich et al., 2008). The theoretical model most commonly referred to in this student retention/drop-out literature is that of Tinto (1988). His model gains support because it immediately appeals to common sense, with its central notion of “integration” claiming that whether a student persists or drops out is quite strongly predicted by their degree of academic and social integration. This evolves over time as integration and commitment interact, with “dropping out” depending on the level of commitment at the time of the decision. Moller-Wong and Eide (1997) extended this theory to group the factors attributed with influencing attrition into five categories: academic background, academic and social integration, attitude and motivation and institutional fit. Reports from several large scale longitudinal studies (Brainard and Carlin, 1998; Eris et al., 2007; Haag et al., 2007; Marra et al., 2008; Ohland et al., 2008; University of Hull, 2002; Seymour and Hewitt, 1997) support that categorisation and provide compelling evidence of the importance of factors such as self-efficacy, appropriate academic background and motivation, and engagement when it comes to persistence in engineering degrees.

Critiquing the model of Tinto, Georg (2009) suggests that it emphasises individual attributes and does not give enough consideration to the institutional characteristics of a subject area such as the extent of regulation of a programme of study and the quality of teaching and advising. His own research findings, however, suggest weak commitment to the course of study has a dominant influence on departure choices, with institutional factors having a more modest influence limited to maintaining or improving teaching quality.

Of the research recommending and reporting on strategies to improve retention, a high proportion focuses on the first year experience, adding hands-on, problem solving courses (Knight et al., 2007) to engage students through relevant, real life examples of engineering practice. Evaluation of the effectiveness of the implementation of curriculum redesign or support strategies is less readily available and evidence is often based on “one-off” implementations (Aziz, 2008; Hammoudeh and Barrett, 2002; Light and Davis, 2004; Ohland et al., 2004). Only a few studies such as Ohland et al. (2001) have attempted to evaluate the effect on retention of the implementation of curriculum redesign by tracking retention data over a number of years, pre and post implementation. Looking back over 30 years of retention/attrition studies, Tinto (2005) suggested that it is not only important to identify effective action, it is also necessary to implement it fully and in ways that will endure and enhance student retention over time.

Internationally, engineering education provides many areas of similarity to the Australian context, making much of the research directly applicable. Australian engineering programmes, particularly at first year level, have led the way internationally in the area of integrated curricula and there is a clear need to evaluate the impact of restructured curricula. The current study will provide answers to the suggestion that “drop-out” may have shifted from first year to later years, as found in Ohland et al. (2001). Whilst significant differences in retention rates continue between institutions and sectors, demonstrating that retention and attrition are not mono-causal but the result of complex interactions, evidence-based research will assist in identifying and evaluating strategies and practices which appear to have been successful in improving progression and completion rates. From this research it is hoped that “best practice” guidelines for institutional policies and practices will emerge.

Research question and goals

The research question for the pilot study, reported here, is:

“What are the characteristics of those who leave before completion of an engineering degree at this university, at what stage do they go, where do they go to, and what are the driving forces behind their choice not to persist with their engineering studies?”

Methodology

The pilot study institution ATN1

ATN1 is located in the central business district of the largest city in Australia, with approximately 3000 on-campus undergraduate engineering students enrolled in 12 accredited engineering specialisations. The majority of students commute into the city, with limited university-provided accommodation. There are over 40 languages other than English spoken in the homes of the engineering student body, reflecting a diversity of ethnic backgrounds. The ‘flagship’ undergraduate engineering degree is the Bachelor of Engineering/Diploma of Engineering Practice (BEDipEngPrac) degree, which incorporates two semesters of work experience or internship. This is a five year degree undertaken by approximately 70% of the entering cohort. A traditional coursework- only four year degree is offered, predominantly to international students. Approximately 16% of the entering cohort study a double degree combining engineering with business, arts, science or law and a smaller group (∼3.5%) study for a three year non-accredited engineering technologist degree.

Sources of data for the pilot study

One of the first difficulties encountered in research on attrition is consistency and clarity in defining “drop-out”. Government statistics tend to use the definition of “not re-enrolling in the current degree” but, in seeking to understand the impact on the future engineering workforce of students dropping out of an engineering programme, it is necessary to be more definitive. A student may terminate his/her engineering studies without completing a degree programme, but this must be differentiated from cases where students change institution but continue in engineering study, change to a non-engineering programme at the same or other institution or interrupt their studies for other reasons. This clarification, hidden by aggregated statistics, needs uncovering before strategising at the institutional level to reduce attrition can be effective.

The mixed methods approach used in this study was appropriate where answers are sought to questions not only of “who, how many and “when?”, but also “why?” The first source of information, investigating the dynamics of attrition at an individual level, was a cohort analysis providing the opportunity to compare the characteristics of those students who stay to completion with the destination, timing, academic performance and characteristics of those who leave. An exit questionnaire was sent to all those had left prior to completion in the most recent year (2008). Broad brush data on destination and the factors influencing the decision not to continue was sought. It was recognised that decisions to persist or not persist with study of a degree are rarely as straightforward as a questionnaire might imply. Interviews were therefore conducted to hear the ‘stories’ of students who had persisted in their studies over a longer period than usual, switched to another degree or not persisted and left the university. A summary of each of these methods of data collection follows.

Cohort analysis

Two factors needed consideration before selecting an entry year for a cross institutional comparative study. The first was to pick an entry year for which a significant proportion of entrants would have graduated by 2009, recognising the potential diversity in pathways and degree structures across institutions. The second pragmatic consideration was the ability to extract consistent and reliable student data from student database management systems, noted in other studies (Cao and Gabb, 2006) to be susceptible to categorical or data entry errors.

Looking at the records of students graduating in 2008, it was found that approximately 47% of the BEDipEngPrac graduates had taken between six and 11 years to complete. Clearly, perceptions that only a small proportion of students completed in the minimum time were accurate. In consultation with the university’s Statistics and Planning Unit, the decision was made to choose 2003 as the year for inter-institutional comparison, with data from the 2006 entering cohort also to be analysed. Human resource and project funding constraints precluded the analysis of each year since 2003.

The following attributes were identified for each of the selected cohort: citizenship (classified as domestic or international fee paying), gender, and number of points of credits or exemptions available on entry from previous academic study. Where available, the following attributes were collected for further in-depth analysis: admission index, entry pathway, birthdate, school type, languages spoken in the home, and engineering specialisation (discipline).

A series of enrolment information, comprising course enrolled, number of subjects attempted, and number of subjects completed, was downloaded as a spreadsheet for each student for each semester. This was then coded as:

• Currently enrolled

• Enrolled at the same institution in a non engineering degree

• Not enrolled because of engineering degree completion (graduation)

• Not enrolled at the institution

Two codes specific to the pilot study institution were included:

• E Not enrolled because student on work experience/internship

• G Not enrolled but returned to enrolment in a later semester

For each semester, full time/part time status and the level of academic success achieved was also coded. A table (similar to Table 1) containing the data for each student in the entering cohort was constructed, with each student ID cross-linkable to individual attributes as identified above.

This base data enabled enrolment patterns and levels of attrition to be identified for a variety of attributes.

Exit questionnaire

Whilst acknowledging that obtaining responses to a postal survey from students who had left an engineering degree had been found by other studies to give low response rates (Baillie, 2000), a questionnaire was deemed to be an effective method of acquiring data from a wide range of non-persisting students. As recommended by Cresswell (2008), it was decided that a previously validated questionnaire would be adopted: one which had been used across all disciplines (including engineering) at the other pilot study institution. It had been extensively tested and validated and was potentially suitable for use in the third phase of the project with collaborating institutions.

The questionnaire was posted to domestic students who had withdrawn from engineering in 2008 or early 2009 (n= 170) with an accompanying letter of explanation, invitation to interview and a stamped addressed envelope.

Those questionnaires which were returned as “address unknown” were followed up, where a number was available, by a phone call. Where contact could be made the questionnaire was re-sent or a brief discussion was held by phone. It was noted that approximately 75% of the mobile phone numbers provided were still in use, suggesting that a future strategy to increase response rate might be to use an appropriately worded phone call or text message.

Ultimately only 16 copies of the exit questionnaire were returned (9.4% response rate), although telephone contact was made with five additional ex-students. Because the survey was anonymous it was not possible to know who had returned the questionnaire, except where they had sent back the accompanying ‘invitation to interview’ letter.

Interviews

Three categories of students were invited to interview: persisters (those students who were identified from preliminary cohort analysis data as taking longer than minimum time to complete the degree), non-persisters (those who had not re-enrolled in any degree at ATN1 during 2008 or 2009) and switchers (those who had left engineering in 2008 or 2009 but enrolled in a non engineering degree at ATN1). Switchers and persisters were sent an invitation to interview by email and the invitation was included in the exit questionnaire mailout for the non persisters. The aim of the interviews was to add depth to the factual information available from transcripts and survey data by hearing the students’ “stories” and perceptions of the factors which influenced their decision to persist with or leave engineering.

In the month before lectures finished in 2009, interviews were conducted with six persisters, six switchers and six non-persisters by the project research fellow. The interviews followed a semi structured format, extending the exit questionnaire around the themes of: motivation for choosing engineering, self perception of preparedness, academic background, academic progress and engagement, and factors influencing the decision to persist with or leave engineering. The interviews were professionally transcribed and are being coded using nVivo software. Detailed analysis of the interviews is underway, but incomplete at the time of writing. Each of the switchers and non-persisters interviewed had completed the exit questionnaire and, although it is already apparent that the interviews contain much rich detail, their essence is contained in the responses to the exit questionnaire outlined below.

Table 1 Simplified representation of the base data used in the cohort analysis

Student ID	2003a	2003s	2004a	2004s	2005a	2005s	2006a	2006s	2007a	2007s	2008a	2008s	2009a
1007	A	A	A	A	A	A	G	A	A	A	D	D	D
1008	A	A	A	A	A	A	A	A	A	A	A	A	D
1009	A	A	A	E	A	A	A	A	A	A	A	A	C
1010	A	A	A	A	D	D	D	D	D	D	D	D	D
1011	A	A	A	E	A	A	A	A	A	A	A	C

Cohort analysis — information on retention pathways

It is not possible in the space available to adequately convey the wealth of information provided by the cohort analysis, and further analysis is ongoing. Initial findings which both affirm and contradict previous perceptions by the host institution are provided. Although data was collected as a semester based time series (in preparation for comparisons with collaborating institutions), the findings are presented by year, with counts of students in each category at the beginning of the year. In Australia the academic year matches the calendar year.

Although the primary focus of the study is on domestic students (in keeping with the overarching aim of the project to increase the potential engineering workforce in Australia), the patterns shown in Figure 1 show the difference in completion rates between domestic and international students. Attrition for international students appears higher after one year, although attrition proportions become more similar in later years. Although attrition for domestic students after six years is higher than that for international students, the difference is not statistically significant. By contrast, the difference in graduation rate is visually and statistically significant. It is to be noted that the majority of international students pursue the four year degree option, whereas domestic students tend to pursue the five year degree option.

Figure 1 Retention pathways for the 2003 entering cohort by citizenship

Although these initial graphs do not attempt to distinguish between the different engineering degrees studied, the very low proportion of domestic students who completed their degree in five years of study (13.6%) was of concern. Staff at ATN1 were aware that the proportion of students completing in the minimum time was low, but it was lower than expected.

When the information for domestic students leaving engineering is split to differentiate between those who switch to a non-engineering degree at ATN1 and those who leave the institution completely it is seen from Figure 2 that only 5% of the original cohort appear to switch to a non engineering degree at the same institution. This proportion had been perceived as higher by some of the collaborating institutions.

The highest annual attrition for those leaving the institution and those switching to a non- engineering degree occurs after one or two years, and it had been the perception of ATN1 staff that these students may have been influenced by their experience during their internship, usually undertaken in the second year. Fine grained analysis of the base data refuted this perception, as no student in the cohort switched or left the institution after their first semester on work experience. In fact, interviews confirmed that for some students the unavailability of a suitable internship was a factor in their decision to discontinue.

Comparisons of attrition patterns for those entering with credit (advanced standing)

Potential differences in attrition patterns between those who entered at first year level and those who came in with sufficient advanced standing to be deemed as entering above first year level (and hence the likelihood of a shorter degree completion time) were of interest. Compared to more traditional universities, ATN1 had a high proportion of students entering with advanced standing (23.6% in 2003). These students predominantly entered after completion of an engineering technician qualification or transferred from one of the regional universities. Using the classification that advanced standing of at least ¼ of a year’s point loading would qualify as entering above first year, it was found that 65 (14%) of the entering cohort fitted into this category. With their proven commitment to engineering study it was expected that this group would have a lower attrition rate - this is demonstrated in Figure 3.

Of note was the almost complete absence from the second group of students switching to a non-engineering degree. Clearly, students entering after prior study were confident and committed to the engineering programme. The seemingly higher graduation rate of 44.6% after six years of study compared to 31% for the first year entrants is negated if the advanced standing of one year credit is accounted for. A more equitable comparison would be to look at the graduation rate after five years for the non- first year entrants: this is 30.8%, remarkably similar to the graduation rate after six years for the first year entrants. There is an apparently higher attrition rate for non-first year entrants after one year of study (15.4% compared to 10.5%) and it was expected that interviews would provide further insight into possible causes.

Figure 2 Retention pathways for the 2003 domestic entering cohort, including information on switching to a non-engineering degree at the same institution

Full time and part time study

The low proportion of completions in the minimum time alluded to earlier appeared to reinforce the tendency of ATN1 students to shift to part-time study as they progressed through their degree, particularly after completing an internship. Figure 4 illustrates this tendency. The high proportion of part-time students entering with advanced standing points is attributed to the likelihood of them being in employment whilst studying, right from entry. This data is given by semester rather than year because of the confusion the internship semesters give to annual full-time/part-time calculations.

By gender

Australian statistical data had suggested that, on average, male attrition is slightly higher than female attrition within engineering, in agreement with the findings of Cosentino de Cohen and Deterding (2009). The data from ATN1 were in agreement with these findings, but an interesting gender differentiated feature was observed when non-persisters were split between “left ATN1” and “left engineering but continued at ATN1 in a non-engineering degree”. Figure 5 demonstrates that 4.5% of the male cohort entering at first year level (remembering that no switchers were identified for entry above first year level) and 15.2% of the equivalent female cohort transferred from an engineering degree to a non-engineering degree at ATN1. This result was statistically significant at the 95% confidence level. It would appear that, particularly after the second year of study, female students tend to switch out of engineering, whereas a higher proportion of male students persist with their study, although neither the graduation nor currently enrolled differences is statistically significant. As noted earlier, this is not linked to work experience and questions arise whether switching is related to the engineering education culture at the institution, lack of understanding about the engineering degree content and career, or achievement levels.

Figure 3 Retention pathway for 2003 cohort split by entry at first year level or above

Figure 4 Full-time part-time enrolment pattern for the 2003 entering cohort, split for entry level

Transferring to a non-engineering degree

The transcripts of all students in the 2003 cohort who switched from an engineering to a non-engineering degree but stayed at ATN1 were examined to see if any trend could be identified. Two patterns were evident. One group entered with reasonable entry grades, were successful in their study and appeared to be using their engineering grades to gain entry to a preferred degree for which their original admission score did not qualify them. The other group struggled in engineering, particularly in mathematics and subjects known to be “killer” subjects. Over the years these “killer” subjects have presented a major barrier to progression in electrical, computer and software engineering and are the subject of a separate project within the institution.

Comparison of overall attrition rates from 2003, 2005 and 2006

Attrition data from 2006 was deemed to be of special interest to ATN1 because previous statistics aggregated for census purposes showed a dramatic reduction in attrition (from 20% to 9%) between 2005 and 2006. In actual fact, the analysis for the cohorts entering in 2003, 2005 and 2006 demonstrates that the 2006 attrition after year one was 12.9% — not a significant reduction. More accurately, the increase in attrition after the first year for the 2005 entry cohort was statistically significant when compared to both the 2003 and 2006 entering cohorts.

Figure 5 Retention pathway for entry cohort at first year degree level, differentiated by gender

When attrition is viewed as entry year, entry year +1,+2 and +3 (as displayed in Table 2) it becomes apparent that after three years attrition for all three entry cohorts has evened out.

This type of evidence leads to questioning the focus on curriculum redesign and support strategies at first year level. A variety of interventions were in place for the 2006 entering cohort, yet the effect does appear to have been merely a shift in the timing of attrition rather than a reduction, confirming the findings of Ohland et al. (2001).

Findings from the exit questionnaire

The exit questionnaire contained two open ended questions:

1. What factors influenced your decision to leave the university?

2. Was there anything that the university could have done differently to support you with your studies?

Specific information was then sought about destination on leaving the university, intention to return to studying engineering at ATN1 or another institution, and preparedness for their engineering studies. The major portion of the questionnaire asked the respondents to state how influential a list of statements were in affecting their decision to withdraw from engineering degree studies at ATN1 (choosing from: not at all true, slightly true, very true and extremely true). The responses to this part of the questionnaire are displayed in Figure 6.

Table 2 Comparison of attrition after three years from entry

Entering year	Left engineering +1 year	Left engineering +2 years	Left engineering + 3 years
2003	13.9%	23.5%	28.4%
2005	20.0%	26.3%	31.2%
2006	12.9%	24.8%	31.7%

Whilst recognising the lack of reliability in using a sample size of 16, the main causes given for withdrawals, as displayed in Figure 6 were:

1. I was not performing as well academically as desired

2. Time taken to travel to university was limiting

3. I had a lack of enjoyment or interest in course content

4. I felt there was a lack of information available to help guide me in my choice of programme

5. I had too many external commitments to allow time for university study

6. I had difficulty in understanding academic content and concepts

7. I had doubts regarding my ability to perform well enough.

As has been evidenced in the research literature (French et al., 2005; Matusovich, 2008), lack of academic progress can be strongly linked to deteriorating self confidence, motivation, commitment, engagement and lack of enjoyment and interest — all strong predictors of attrition. Only one of the questionnaire respondents indicated that they came in well prepared for engineering, whereas 13 out of the 16 respondents suggested they were either poorly or mildly prepared for their engineering studies. In particular, a lack of confidence and familiarity with mathematics at an appropriate level for entry were identified by a third of the respondents. Interview data indicates that a high proportion of non-persisters and switchers were either under-prepared academically or in their understanding of engineering and, where it was possible to check transcripts, these perceptions were confirmed. Finer grained scrutiny of the cohort analysis data will use logistic regression models to explore whether a causal relationship exists between lack of academic progress and attrition.

Unique to ATN1, and of concern as major causes of attrition, were ‘time travelling was limiting’ and ‘lack of guidance in choice of programme.’ Public transport and parking present challenges in a large, central-city university, particularly one which has a high proportion of part time students. In an age of ever-expanding communication technologies and possibilities, ATN1 will clearly need to address these issues if the attrition rate of motivated students is to be lowered. Lack of suitable guidance in choice of programme resulting in students entering engineering without a clear understanding of the course content and career options is also a major concern. A mismatch in expectations and reality of the engineering course and content was evident, with several of the non-persisters interviewed discussing a lack of connection with the course material. It has been noted in the research literature (Hammoudeh and Barrett, 2002; Ohland et al., 2008) that the goal of retention strategies is to retain those students who are qualified and interested in engineering. So what of those who enter without strong motivation and commitment? Improving advising processes and early exposure to the engineering profession appears to be a necessity if students are to be prevented from entering engineering without a commitment to the course.

The sample size of responses to the exit questionnaire is too small to give a clear picture of destination on leaving ATN1, but the responses did indicate a reasonably even split between continuing engineering at another institution (five), studying a non-engineering degree (five) and becoming employed (six, of which four were in an engineering related area).

Conclusions and recommendations

Several issues have emerged from these initial findings. The first is the need to investigate attrition by engineering specialisation — “killer” subjects discussed by interviewees appeared to cluster around one engineering specialisation. The second is a need to investigate the link between the internship experience and drop-out. Although no student appeared to have dropped out immediately after an internship, a connection did seem to be evident between timing of the internship, a shift to part-time study and delays in degree completion. Links between part-time study and drop-out, particularly for those who had entered with advanced standing, highlight the urgent necessity (expanded on by students interviewed) to investigate ways of optimising timetabling constraints and student needs.

Figure 6 Factors influencing the decision to withdraw from engineering study at ATN1

Evidence that advising systems could be improved needs closer examination by the institution. Questionnaires and interviews with non-persisters and switchers may provide a biased viewpoint of the effectiveness of the current advising system. Evidence being gathered in this study appears to support the need for better risk management, not only improving advising about programme choice but also identifying students at risk of failing as early as possible and implementing the necessary support systems.

As the pilot study is expanded to work with collaborating institutions, the cohort analysis model has demonstrated its potential to increase our understanding of the “who”, “how many” and “when” questions of attrition. The limitations to the use of survey methods to expand our understanding of the “why” questions lie in obtaining feedback from those who leave the university. Not only are contact details subject to change, but students who drop out are likely to have more personal difficulty discussing their situation than someone who has successfully graduated.

The initial findings mentioned here for the cohort analysis and the exit questionnaire are only a starting point but they have already highlighted some confirming and contradictory trends. Further in-depth analysis, including interview data and documentary evidence for triangulation, are expected to extend and expand the understanding of attrition causes and dynamics at the pilot study institution.

Acknowledgements

The project leaders would like to acknowledge the assistance of Peter Antony, for extraction of data for the cohort analysis model, and the Australian Learning and Teaching Council, who have provided funding for the project Curriculum specification and support systems for engineering education that address revised qualification standards from which this research is drawn.