HEFCE Report 00/15

Study of comparative costs of first degree and sub-degree provision

Report to the HEFCE
by J M Consulting Ltd

This version of the document contains the Contents and Summary Report only. The full Report is available in both Word (282Kb) and RTF (455Kb) formats.

Contents

1 Summary report

1. Purpose of the study

2. The aim of this project was to compare the costs of different years of full-time undergraduate degree and sub-degree programmes in a selection of universities and further education colleges. The questions the project was intended to address were:

   a. Is there any systematic difference in costs between years of a sub-degree or degree programme?

   b. Is there any systematic difference in costs between sub-degree programmes and degrees in those institutions which offer both types of provision?

3. If such systematic differences existed, they could have implications for future funding policy and methods.

   Method

4. We have conducted this study by a method which compares the relative costs of the different inputs (teaching staff time, learning resources, infrastructure, etc) that are used to service these different years and types of provision. We have called these the input factors. This method does not provide the actual costs of the provision in the way that a ‘top-down’ full-cost approach would, but does review the relative effect of the different cost drivers at a programme and year level which are required for this study. This approach has the advantage of being very much less costly to apply, and of being much less burdensome for the institutions which contributed to the study.
5. The study did not look at differences in costs between institutions.
6. We have applied this approach in a sample of 11 higher education institutions (HEIs) and 10 further education colleges (FECs) as listed in the Annex of Evidence. These institutions were chosen to be reasonably representative of the range of types of institutions where degree and sub-degree programmes co-exist in the same subject areas, and also included two HEIs not offering sub-degree programmes.
7. The study’s main focus was on the relative costs of a full-time student on a three-year undergraduate degree programme, compared with a two-year Higher National Diploma (HND) programme. Other programmes were also reviewed, but the samples were much smaller: the Diploma in HE (DipHE) gained through credit accumulation, Foundation (HEFCE funded) and taught masters.
8. We compared the degree programmes with sub-degree programmes offered in the same departments. In most of the institutions visited we covered two subject areas (one classroom and one laboratory-based) and more than one degree and sub-degree programme (or pathway) in each of these.
9. The study did not look at Higher National Certificates (HNCs), part-time provision, professional diploma programmes, or franchised provision.
10. The purpose of the study was to look for any differences that are large enough and systematic enough to be potentially of relevance to funding policy. It is inherent in the method used that it will not reveal small differences in costs, and that it may not reliably detect particular factors that are only relevant in a small number of institutions or programmes, or in a particular type of institution or programme outside our sample.
11. The study looked at variations at programme level, not between institutions or between departments. Although the number of institutions and departments visited is small when compared with the whole sector, our research did cover many programmes and pathways within each institution. The conclusions based on the sample are considered to be robust.

    Context – nature of provision

12. The first output of the study is a clearer picture of the nature of this type of provision. This does not directly answer the questions at paragraph 1.1, but it is important contextual information for policy makers.
13. The essential feature we found is the extreme complexity and diversity of arrangements, and the changes currently (and constantly) being made in both resource use and programme provision by institutions. These factors indicate a need for caution about drawing any definitive conclusions unless these are supported by consistently based evidence across a significant number of programmes.
14. The questions that we were asked to address, while valid in policy terms, imply a degree of simplicity and consistency in this area which we have not observed in the institutions we have visited. Three key factors define this complexity and can be regarded as conclusions of the study.

    Conclusion 1. The concept of different years of a course is of limited validity, and the notion of ‘level’ does not provide a robust alternative.

15. Most student programmes are now unitised, and units or modules are not necessarily taken in any specific year.
16. Many institutions have a policy of allowing flexible entry and access to programmes at different points, and options and electives (particularly in modular schemes) are often extensive. Moreover, resource pressures have led to specific policies of maximising student numbers on modules; and students can and do take modules at different points in their programme. This means that students taking a module may be in a ‘different year’ to others on another programme. Industrial/overseas placement years add further complexity when categorising students into a particular ‘year’ group.
17. Degree programmes are normally categorised into ‘levels’ (1 to 3). However, this term is not consistently used, and some institutions do not classify their HND modules into year or level. Some use ‘Part I’ and ‘Part II’ instead: the latter covers modules in both years 2 and 3. Students may take a module in a year of their programme that is different from that of many of the other students taking that module.

    Conclusion 2. Sub-degree provision may not be a distinct element within the institution.

18. DipHEs are often not distinguishable from degree programmes. Most HEIs offer DipHEs as an ‘unplanned exit route’ after two years (240 credits) from their degree programmes, with the students actually registered for the degree award. However, other DipHEs (not covered in our sample), particularly those in professional disciplines, are awards in their own right.
19. HNDs are historically vocational, with an emphasis on skills, while degree programmes have been considered to be more academically based. However, a number of factors mean that this distinction is now not so clear:
    • professional degree programmes
    • the still increasing popularity of ‘vocational’ disciplines such as business studies and computing
    • the demands of students (now fee-paying) for programmes with ‘use-value’
    • the need for remedial skills to be taught to A-level students.
20. These have all helped to align more closely the different academic/knowledge, practical/skills and personal/social base of the degree and sub-degree programmes.
21. Many course/department managers felt that their degrees were either strongly vocational now or should change to incorporate more of the vocational features and skills support evident in their HNDs. Some top-up degrees (an HND graduate studying a third year and gaining a degree i.e. a 2+1 model) are recognised as less academic than a three year comparative degree in the same department.
22. Many course/department managers felt that their HNDs were (or should be more so) as academically based as a degree, predominantly because of student demand for access to a three year degree (i.e. 2+1 model, not 2+2 model). Employer expectations for a degree qualification, rather than an HND, are a driving factor here. (This is not true of all HNDs, but applies to most in our sample even where the HND also provided a valued exit route in its own right.)
23. In any case, modular schemes, with integrated teaching and learning on many of these programmes, have led to significantly similar curricula (even if assessment or tutorial support are different). Teaching and learning can be shared not just between HND and degree students studying the same subject, but with other groups of both full and part-time students on other degrees or HND/HNCs.
24. It is a now common practice for HND graduates to progress to the second or third year of a degree. This has led to the specific focus on learning outcomes at each year of the HND and degree which facilitate ‘bridges and ladders’ across the programmes. The clearest example of this is where an HND is fully embedded in a degree programme, and some students are joint registered. However, in practice there is a very wide range of approaches to common or differing teaching and/or assessment methods, depending largely upon the academic view of the student intake and their different needs.
25. In many programmes, the only difference between the HND and degree programme is that the students on the HND have only one A-level, not two and have ‘less confidence’, but often equal ability, given appropriate learning opportunities. This allows a significant, or total, commonality of approach, as long as adequate support is provided in this. This can be compared to other programmes where HND students have had a very different educational or personal background, and as a result, academics plan a significantly different teaching and learning experience for them from that for school-leavers.

    Conclusion 3. The current pattern of resource allocation is a poor guide to future policy.

26. Finally, in this section, it is important to note that policy and practice are changing. Most institutions we visited are responding to the Government’s agenda of access and increased participation. In practical terms, this means taking students who are often less well equipped to cope with the traditional HE qualifications (whether HND or degree) and retaining these students (minimising attrition). Institutions are adapting their ‘input factors’ to respond to these needs with a strong focus on ‘year 1’ support. There are cost implications to this and many institutions either have shifted or would like to shift their resources accordingly.
27. Institutions made the point that it would run counter to Government policy to devise a new funding method that consolidated or fossilised historic patterns of resource input at a time when they are trying to break away from these patterns in order to support students with non-traditional entry qualifications and flexible progression routes. They also point out that such support should be planned on a student lifetime basis not on a ‘one year at a time’ basis (employing concepts such as ‘investment’ and ‘progression across a continuum of learning’).

    Costs of provision

28. We have examined the relative costs of different years of study in both HND and degree programmes in a range of subject areas across the group of 21 institutions listed in Section 2. Our main conclusions on costs are as follows.

    Conclusion 4. There is a multiplicity of interlinking ‘input factors’ in these programmes but staff input is the area of cost which is most closely affected by the year and type of programme.

29. We reviewed a large number of factors which influence the relative costs of the different elements of provision covered by the study. The main costs in HE institutions are staff (typically 55-60%), learning resources and staff/student services (around 10%), and physical infrastructure (10-15%). An allocation of each of these (generally via the department, programme, and finally module) totals the costs of a programme, with costs per student influenced by student numbers on a particular programme or module.
30. Academic staff costs in academic departments constitute about 35% of institutional costs in the sector and these prove to be the most relevant costs to consider when looking at the variations between types of provision (see below).
31. We found that, in practice, the differential use of learning resources and staff/student services between years or types of provision was occasionally important (e.g. higher use of library or of counselling in a particular year); but this was not common to all programmes. The use of learning resources and sometimes more specialised equipment/software was often considered more expensive in the final year, but there is no quantitative information to inform this.
32. Estates and physical infrastructure usage comprises a significant area of resource. The relative use by programme or year can only be considered on a module by module basis, and depends largely upon subject, and student numbers per module. There were many examples of estates constraints (e.g. on room size) that had led to repeats of modules or lectures, with related resource implications.
33. We found examples where a year or type of programme had a higher or lower use of equipment (particularly specialised equipment), or of laboratories (requiring technician support) or studios, etc. However, these distinctions were often fairly subtle, and not easy to quantify: we identified no factors that varied consistently between years or type of programme.
34. We examined other cost areas, such as administration, management, and central services. The year and type of programme can influence these costs, but they are driven much more significantly by other drivers which exist at institution and department level.
35. The most important cost driver for this review (and the main focus of our work) was therefore the intensity of staff input - in particular, the nature of the teaching and learning sessions and the time spent on assessment and preparation.
36. The staff input cost per student was influenced by a complex web of interelated factors. There are three main factors:
    • subject (e.g. class sizes for laboratory/practical class work)
    • academic teaching and learning strategy (mix of lecture, seminar, tutorial, practical, project; contact time vs student reading time; type of assessment), and
    • the number of students on a module.

    These are themselves influenced by:

    • market considerations (i.e. demand and recruitment success)
    • pedagogical considerations (e.g. nature of student experience; nature of the modular scheme; specialisation/options/electives)
    • learning objectives; professional body requirements (professional qualification); validating body requirements (award); institutional policy on resits and repeats, and on support to students, etc
    • nature of the students (whether different cohorts should have different learning experiences; popularity of the modules; needs of any one student)
    • size of student cohort on a programme, on a pathway
    • accommodation (i.e. size and fitness of teaching rooms)
    • health and safety
    • resources available to the department/faculty/institution.
37. In summary, we found that the most significant cost factors from those above are, in most cases:
    • the number of students taking a module
    • the number and size of the groups where staff teach/interact with students
    • inclusion of project or dissertation work
    • assessment and preparation time.
38. We discuss these in more depth in the Annex of Evidence, but it is worth noting here that:

    a. Staff input time in the form of contact hours could be quantified. However, in many departments a key driver of this was the lecturers, who have the discretion to determine the teaching and learning methods they use. This has led to very wide variations in the type and number of taught sessions used in modules common to a particular type of programme or year.

    b. Similarly, particularly on large modular schemes, there are very large variations in student numbers attending an individual teaching session, and this is irrespective of the number of entrants on any one programme.

    c. Project/dissertation support is less well quantified, although some ‘norms’ seem to be commonly used in workload planning.

    d. The time taken in assessment and preparation was considered significant (in HE, generally at least double the contact time) but very rarely quantified. Notional allowances were usually used (if at all) in workload planning models, and views ranged widely on the actual amount of time spent and the factors that are driving the time (which types of assessment).

    Conclusion 5: The observed differentials are extremely sensitive to assumptions, particularly those on the pathways/options selected, student numbers per module, size of seminar groups, dissertation support taken up, and relative time in assessment or preparation. Our findings can only be considered representative to within at least plus or minus 20% (of the total value). As a result, differentials (and the underlying cost drivers) would need to be both consistent and significantly larger than this to be a firm basis for a change in funding method.

39. This sensitivity and variability means that for almost any significant cost differential that can be identified in one programme or department it is possible to find another programme (or pathway within that programme) with either no significant difference or even the opposite differential. To illustrate this we found examples where:
    • year 2 of a degree is more expensive than year 1
    • year 1 of a degree is more expensive than year 2
    • there is no significant difference between these years.
40. The differentials calculated are very sensitive to assumptions in a number of important areas:
    • on the cohort size per module (which could change year on year, irrespective of the number of students on any one programme);
    • the time spent in supporting projects/dissertations, particularly in non-laboratory subjects, which depend on the wide range of factors summarised above, as much as year or type of programme;
    • the time spent in assessment and preparation (about which there is very little hard data);
    • the choice of pathway and/or options in a department or particular programme where a different selection can produce very different results;
    • the number and size of seminar groups, which depend on assumptions about academic staff preference, resource availability, availability of rooms of particular size, how cohorts of students are allocated to groups, the maximum size of a group, etc.
41. These sensitivities and variability mean that the differentials that we calculated could only be regarded as representative within plus or minus 20%. For example, if year 1 costs were calculated to be 120% of year 2 costs (a differential of plus 20%), with some changes to assumptions this value could easily be recalculated to a figure somewhere in the range of 100% to 140% (or even wider).

    Conclusion 6. While there are variations in cost, there is no systematic difference at a level which is material for policy purposes either between years of provision or between sub-degree and degree provision in this sample of institutions.

42. There are some variations in the costs of direct teaching time which, while not systematic, material or universal to the extent required for policy purposes, can be observed in a significant number of institutions and programmes. The most significant of these are:

    a. The costs per student of year 3 of degree programmes is often marginally more expensive than year 2 (nearly half the sample was up to a third as much) due to project work or dissertations in the third year and specialisation or options leading to smaller modular sizes.

    b. HND programmes are often marginally more expensive than year 1 or year 2 of degree programmes even when teaching is shared, due to the extra tutorial and skills support needed (nearly half the sample showed costs per student per year up to a fifth higher).

43. However, in both of these, costs are significantly affected by many factors that are not consistent programme to programme (even if the final differentials happen to be the same) nor are they necessarily programme- or year-dependent. Not all factors that lead to these cost differentials form an appropriate basis for informing the funding methodology - for example, attrition often plays a part (sometimes significantly) in increasing the costs per student in one year over another - and yet recognising this would ‘reward inefficiency’.
44. In any case, the above summary only represented half the sample and the two most common findings. A wide range around these results was found, as is shown in Figure 1. This shows the main factors leading to differentials, and the levels of the differentials found, for year 1 (Y1) of a degree programme compared to year 2 (Y2), and year 3 (Y3) compared to Y2.
45. We found that:

    Y3 costs per student from direct teaching time were:

    • marginally more expensive (up to 30%) than Y2 in nearly half the sample
    • equal to Y2 in a quarter of the sample
    • significantly more expensive (over 30%, in a range up to 100%) in a quarter of the sample.

    (The comparison here is to Y2, taken as a ‘base year’. Where Y1 is higher or lower than Y2 – as applies to half the sample – then the differential between Y3 over the average Y1/Y2 cost would be lower or higher respectively.)

    Figure 1: Differentials in the cost of direct teaching time, between years

    Not available in this version

46. HND costs per student in Y1 and Y2 were:
    • marginally more expensive (up to 20%) than a comparable degree programme in nearly half the sample
    • equal to the degree costs in more than a third of the sample
    • significantly more expensive (50% plus) in a small number of those sampled.
47. Y1 costs were similar to those of Y2 (almost all within plus or minus 20% of Y2 costs per student). Lower costs per lecture in Y1, from fewer options and fewer pathways (therefore higher numbers per module) were diluted by the extent of the teaching in small groups (in HE, perhaps half the contact time experienced by one student; in FE even higher). Lower costs from higher student numbers in Y1 were also diluted by a higher contact time and additional tutorials/skills support in that year when compared with Y2.
48. FEC cost structures were obviously very different from those in HE, but as this study looked only at differentials within an institution, this was not relevant here. Teaching patterns are different, although not completely so (HE also has a considerable amount of small group teaching, and some FE cohorts are higher than some in HE). We found as wide a range in differentials in FECs as were found in HEIs. There were no discernible differences when compared by subject or institution as a whole.
49. Our sample number in each subject type was small, and the specific subject within a subject type varied (e.g. civil, coastal or electrical engineering; in business studies a language pathway, or a particular departmental focus on computing/IT). As might be expected from the wide range of cost drivers reported above, we found almost as wide a range within subjects as between subjects.
50. The analysis shown here relates only to direct academic staff time (lectures, seminars, tutorials, dissertation support). This probably represents less than half of the attributable academic staff costs as it excludes preparation, assessment and administration time. Variations in these and other cost areas (assessment and preparation time, technician support, estates, other central costs, etc) are much less likely to be influenced by student year or type of programme. The differentials reported here for direct academic staff time will therefore be significantly diluted when the full costs of provision are taken into account.

    Conclusions for funding policy

51. Overall, we conclude that this area of provision is complex and is likely to continue to change significantly in most institutions, but in differing ways and to different extents depending upon local institutional factors.
52. There are arguments of principle against basing funding models on a year-by-year analysis of costs as this may discourage institutions from treating the student programme as an integrated whole. (Institutions without a ‘standard’ portfolio may be disadvantaged; some may encourage direct entry to their Y3 programmes; resource distribution inside some institutions may be altered and might lead to changed teaching strategies in some departments.)
53. There are some commonly-observed variations in academic staff cost between years of provision, but these are not seen in all programmes or institutions, and the size of the differentials is usually quite modest. The most common of these variations is to observe a higher cost of year 3 of a programme relative to year 2; and of the HND relative to the first two years of a degree programme. The differences where observed are often of the order of 10%-20%, but, as frequently, more or less than this. (These relate to only part of the total costs of the student, with even less variation likely in the balance of the costs).
54. There are occasional examples of higher differentials in academic staff costs in year 3 costs when compared with year 2. There is some commonality in the drivers behind these (specialisation and projects/dissertation work), but there are many examples where other institutions use modes of delivery that reduce the impact of these same cost drivers (sometimes with this specifically in mind), achieving lower differentials as a result.
55. With a constant funding for students of all years and programmes based on average sector costs, institutions will be disadvantaged only if their profile is out of balance with the normal in the sector. This will happen if they have an unusually high number of direct entrants into a higher-cost Y3, or increased provision through a higher-cost sub-degree programme. (We found these significantly higher costs in a quarter of our sample.)
56. In summary, the differentials we calculated on direct academic staff costs were neither consistent or significant. Other cost areas, which make up the majority of teaching costs, are even less likely to show an influence from year or type of programme.
57. On this basis, we conclude that there is no difference in cost either between years or between levels of provision that would be material and systematic enough to justify any impact on funding policy.