Do students profit from feedback ?

Undergraduate students in psychology were given the opportunity to exchange the traditional exam with portfolio assessment. The students received written feedback, by way of a standard feedback form, on two of the three essays of the portfolio. To investigate whether students attend to and act on the feedback, a comparison was made between unofficial marks on the first draft of the first essay and the official marks on the full portfolio at the end of the semester. With approximately 20% of the first drafts being unacceptable in the end only 1.6% of the portfolios failed to reach the level of acceptance. The result is taken to indicate that the students did indeed attend to and profit from the written feedback.


Introduction
"Nonsense.""Rubbish.""Disastrous.""This leads to nowhere.""It is only your minimal material knowledge that prevents me from giving you an F." The above statements are, contrary to what one perhaps might expect, exact citations from the kind of feedback students of law at the University of Bergen were given on a written assignment.Not surprisingly, this created quite a stir among the students.When asked why he presented feedback in this manner, the lecturer replied: "I admit that I was a bit too brutal in some cases.However, after the first assignment, the students complained that my comments were too scanty, so I decided that this time I would give them some well-founded information as to what they were doing wrong" (Studvest, 2004).In this paper I will take a closer look at some aspects of the relationship between learning and feedback.The proposed question -do students profit from feedback -will be discussed on the basis of data from a study among undergraduate students of psychology where the students were given the opportunity to exchange the traditional 4-hour written exam with portfolio assessment.The students were offered written feedback on 2 of the 3 essays included in the portfolio, and were given the opportunity of rewriting the essays on the basis of this feedback.

The Norwegian context
Following Parliament Proposition no 27-2000-2001, the 'Quality Reform of Higher Education' was implemented in Norway in 2003.As a result, Norwegian institutions of higher education have seen some massive changes in educational programs, teaching and assessment procedures, as well as leadership structures.The reform may be seen as part of a wider European integration process following the Bologna declaration of 1999, in which 29 ministers of education around Europe agreed upon a common degree structure.As part of the Quality Reform, all institutions of higher education are expected to develop systems in which students are secured a closer followup and a proper feedback on their academic achievements.The same institutions are urged to introduce alternatives to the traditional exam, and to make use of more 'student active' teaching.The latter may be seen as a response to the rather severe criticism launched by the OECD, in which Norwegian universities were described as "…research institutions conducting exams…" (OECD, 1997).According to the OECD report, too much emphasis has been on control of students and on grading, at the expense of teaching and co-operative learning.
The Quality Reform, and the manner in which it was implemented, has created much debate.With the odd exception (e.g.Bleikli, 2005), the local debates in our universities have focussed on how much the reform will cost in terms of money and extra teaching resources, nearly all of them concluding with fear that research will suffer and that universities will be reduced to suppliers of teaching and degrees, making them into schools rather than universities.As the reform has been followed by a new system of financing, in which the institution is refunded based on the number of students who pass an exam or finish a degree, there is also an expressed concern that this may affect grading and that it may result in a reluctance to fail students.Such a concern was publicised in an editorial in the membership magazine of the Norwegian Research Association in the autumn of 2004.With reference to experiences from the British educational system, the editorial more than suggested that the increased annual production of credits, and the reduction in failure rates seen across different institutions between 2003 and 2004, were the results of the new financial system (Myking, 2004).

The importance of proper feedback
Taking a closer look at the way in which our educational system traditionally has solved the task of giving feedback to students, we find that this primarily has been done in terms of marks or grades.In cases where additional feedback has been given, focus has typically been on what is wrong or not so good.This is true even in subjects like psychology and pedagogy, despite the long acknowledged fact that positive feedback carries more information and that it may serve to strengthen motivation and self-efficacy (Bandura, 1998;Wormnes & Manger, 2005).Bjørgen (1989), in a discussion of what he refers to as 'ten myths about learning', puts it this way: "We have available to us today an overwhelming amount of research concerning the learning process, with both animals and human beings as experimental subjects.One of the conclusions which stands out most clearly from this mass of material is that the most effective learning is inspired by the carrot rather than the stick: by rewarding what is right, the behaviour that is correct.Drawing attention to what is wrongincorrect answers and inappropriate reactions -seems in most cases to be unhelpful" (Bjørgen, 1989, p. 23).
The importance of positive and constructive feedback is also underlined by Ramsden (2002).He is very critical of a practise in which students only receive a mark or a grade: "It is impossible to overstate the role of effective feedback on students' progress in any discussion of effective teaching and assessment.Students are understandably angry when they receive feedback on an assignment that consists only of a mark or grade.I believe that reporting results in this way, whatever the form of assessment, is cheating students.It is unprofessional teaching behaviour and ought not to be tolerated" (Ramsden, 2002, p. 193).Hattie (1987), in a review of 87 meta-analyses of studies related to what is known to affect students' learning, shows that feedback is the one factor that bears the greatest importance.Rogers (2001) draws much the same conclusion.Here it is argued that the most effective learning happens when we really want and need to learn, when we know how we will apply our knowledge, when we are rewarded one way or another for having it, and when we are supported, stretched and challenged.Topping (1998), in reviewing studies on peer assessment, concludes that feedback from peers may have as good, or even better, effect than the effects of teacher assessment.He also demonstrates that 18 out of 25 studies comparing teacher and peer marks or grades show 'acceptably high reliability ' (p. 257).Students find the process demanding but anxiety reducing, and learning gains in terms of test or skill performance are frequently reported.In an overall summary, Topping concludes that peer assessment is adequately reliable and valid, students find it demanding but anxiety reducing, there is a marked learning gain, and it may improve confidence and result in better presentation and appraisal skills (p.268).Similar conclusions are drawn by Althauser & Darnall (2001), whose study also shows that there is a relationship between the type of feedback and outcome.The better the written peer review, the higher the quality of revised essays.Willis (1993) makes the important point that there is a close association between curriculum, assessment and reporting, and that if one element is to be changed, it is necessary to change others to ensure consistency.Assessment methods have been shown to shape students' learning approach, and before any discussion of assessment reform is begun, consideration should be given to what kind of learning is desired, it is claimed (Willis, 1993).A similar point is made by Jackson (1995), who also makes the point that a deep approach to learning can be encouraged by peer teaching, in that it promotes self-consciousness among students about how they learn.Falchikov (2001) argues along much the same line, showing that positive feedback 'being pleasurable to presenters, seemed to boost confidence and prepare them for the criticism' (p.274).
As is evident from the interview referred to in the introduction, the whole idea of giving feedback may be interpreted in various ways.One very common way of practicing feedback is correcting wrong responses, or plainly marking parts of a written text with words like 'needs rewriting' or the like, without supplying any information that may actually help in the process.There is some evidence that correcting wrong responses, or not giving feedback at all, do have a positive effect on learning (Kulhany, 1977;Meyer, 1986).This is, however, typical of situations in which the learning material is very simple, or limited to situations which allow the learner to see the feedback before responding (Kulhany, 1977).As pointed out by Weaver (2006), students consider feedback as unhelpful to improving learning when comments are too general or vague, when they lack guidance, focus on the negative, or when they are unrelated to assessment criteria.
In a discussion of the benefits of feedback on written work, Thompson (1994) cites students who say that knowing that they will have a 'second chance' makes them more relaxed and willing to be more exploratory.Giving students feedback and the opportunity to rewrite their work is instrumental in providing them with experience of discovering their own standards.This may in turn result in a basic change in how students experience their writing, from regurgitating known information to writing as making meaning.Thompson (1994) shows that students typically receive marks that are 5 to 10 points higher on the second draft than on the first draft.Time, as well as quality of the feedback seem to be crucial elements.In this case students were supplied with typed comments in the range of 700 to 1,000 words shortly after handing in the first draft, and given three weeks to hand in a second draft.In discussing this approach, Thompson makes the point that the process is time-consuming for students and instructors alike, and that it may not be feasible when teaching large groups of students (100-150 students and more).The results reported in the Thompson study are, as far as timing of feedback is concerned, supported by Kulik & Kulik (1988).In a meta-analysis of findings on feedback timing and verbal learning they show that immediate feedback is more effective than delayed feedback in classrooms and when dealing with real learning material.Brown, Gibbs & Glover (2003) list a number of functions that the feedback teachers' writing on students' work may have, for instance identifying where errors have been made, demonstrating techniques or procedures the students may not have used appropriately or correctly, or engaging students in some thinking in relation to what they have written or presented.In order to decide which function is most important, one has to find out what types of feedback teachers actually give, and what students find most useful, they claim.The authors include and describe examples of assessment experience questionnaires and tutor feedback forms that may be useful in various settings.Gibbs (2002), in discussing conditions under which formative assessment supports learning, underlines that feedback needs to be quite regular, and in relatively small portions of the course to be effective.Feedback must be timely in that students receive it while it still matters to them, and it has to be appropriate to the purpose of the assignment and to its criteria for success.Brinko (1993) looks at feedback from the perspective of the teacher, and asks how feedback may improve teaching.Her conclusions are very similar to the ones described above.Feedback is more effective when it focuses upon behaviour rather than on the person, when it is given as soon as possible after performance, and when it contains a moderate amount of positive feedback with a selected and limited amount of negative feedback.
Knowing what we now know about the importance of feedback on students' learning, our main challenge is, probably, to develop structures to ensure that students attend to and act upon the feedback they actually get.This may, as Thompson (1994) points out, constitute a particular challenge in cases where one is faced with large groups of students.The present work is an attempt to indicate how feedback may be presented to large groups of students, and to investigate what effect written feedback may have on students' work.

Background of the study
For a period of three years, starting in Spring 2001, undergraduate students at a course in social-and community psychology (15 credits) were given the opportunity to exchange the traditional 4-hour written exam with portfolio assessment.The course ran once every semester for the three-year period, attracting large groups of students.Each semester, approximately 50% of the students accepted the offer.An account of the number of students who took part each semester, and the results they produced, may be found in Appendix Table A. As can be seen from this table, a total of 1021 students participated.
Based on a general understanding of the importance of feedback on learning, much effort was spent on providing the participants with written feedback.When developing the course we did not, however, discuss questions related to effectiveness of such feedback.No plans were, at the start, made to evaluate whether the students actually paid any attention to the feedback.As we went along, we became increasingly aware of the costs associated with this kind of work, an awareness that in Spring 2003 led to the following questions: does the feedback have any effect on the quality of the written work?Do students use the feedback, and do they actually profit from it?We decided to investigate the issue of feedback effectiveness as we set up a new course in Spring 2003.

The study The Spring 2003 portfolio
The portfolio consisted of three written essays and a written self-evaluation.Two of the essays were assigned to the students.The third had to be chosen from a list of 16 general topics made available to them at the beginning of the course.On this third essay the students had to specify both question and approach themselves.Before proceeding with writing the essay, their choice of question had to be approved by the lecturer.Students sent their suggestions via e-mail and received a reply within 24 hours.Each essay was to contain a minimum of 1,000 words, but not exceed 1,500 words. 1 A description was presented specifying important issues related to structure and content of the essays.The students were also given a detailed written description of the criteria on which the individual essays and the portfolio would be assessed, and told that they could work together with another student on one essay.Four criteria were described; focus, structure, use of sources, and language.The students were told that the essays should have a clear and well-defined focus, to be kept throughout the discussion.It should be well-structured, balancing details with overview, not giving too much attention to particularities.The students were advised to go to the library and use other sources than the proposed textbook.It was emphasised that they should keep citation to a minimum, but try to use the literature as a basis for their own discussion.We also made clear that misspellings, misuse of literature, and a narrative rather than academic style had to be avoided.All students were invited to a two-hour work-shop at the beginning of the course.In this workshop we used pieces of texts to illustrate the criteria that had been spelled out.We also commented on some preliminary texts produced by some of the students.Finally, the participants were asked to produce small pieces of texts and asked to discuss them with another participant.

Feedback
A standard feedback form (A 4) giving detailed information related to the four mentioned criteria, was used.Although no mark or grade was given on the individual essay, the person responsible for presenting the feedback indicated whether revision was necessary or not.This was done by ticking off one of three alternatives; 'needs extensive revision', 'needs some revision', 'ok as it is' at the bottom of the feedback form.All students were invited to send (e-mail) any queries they might have to me (the lecturer), or to the person whose job it was to present the written feedback.Many students took this opportunity, thus receiving additional feedback.The person responsible for providing the written feedback also acted as one of two examiners marking the portfolios at the end of the semester.The second examiner was external, appointed by the faculty of psychology.

Participants -procedure
In Spring 2003, 483 students signed up for the course, 269 of them exchanging the exam with our alternative.The first essay was introduced one week into the course and the students had to hand in a first draft two weeks later.Feedback was presented the following week.The second essay was presented five weeks into the course, and once again the students had to hand in a first draft two weeks later.Feedback followed the week after.The full portfolio was handed in towards the end of the semester, and was marked as one.The students were advised that the portfolio was not complete unless they handed in a written self-evaluation (500 words), in which they reflected upon their own learning during the course.
In order to investigate whether students did in fact act on the written feedback presented, the person presenting the feedback was asked to suggest a mark (not for the students to know) on the first essay, as it was handed in the first time.A copy of all feedback forms, marks included, was kept until the end of the course.A comparison was then made with the results of this informal marking with the final results on the portfolio as a whole.

Results
Figure 1 presents an illustration of the comparison between the informal marks and the final marks.The ECTS-marking scale was used.According to the qualitative description laid down by the Norwegian Council for Higher Education, an A is described as "excellent" and represents "An excellent performance, clearly outstanding.The candidate demonstrates excellent judgement and a high degree of independent thinking."B is described as "very good", C as "good", D as "satisfactory", and E is described as "sufficient".According to the qualitative description, an F represents "A performance that does not meet the minimum academic criteria.The candidate demonstrates an absence of both judgement and independent thinking."Students who receive an F have, in other words, failed.

Figure1. Comparing the (informal) marks on the first essay with final marks, Spring 2003. (Figures in percent).
As may be seen from Figure 1, approximately 20% of the students would have failed had the first draft been handed in as part of an exam -as the only essay and marked by this examiner only.This is, in fact, very close to the results traditionally found among students who sit for the exam at these courses (Raaheim, 2003).At the end of the course, however, less than two per cent of the participants fail.We also observe that there has been an upward shift with the marks showing a pattern that is close to a normal distribution.This may be taken to indicate that there is a general learning effect across the three essays.
Based on what we know about the positive effects of proper feedback, it seems fair to attribute this development, at least partly, to the fact that the students received feedback on two of the essays.

Discussion
Did the students profit from the written feedback presented to them?The results presented in Figure 1 indicate that this may in fact be the case.Admittedly, there are some methodological weaknesses.The first draft was not subjected to an objective assessment using independent assessors.Instead we relied on the judgements made by the person responsible for giving the written feedback.This person also served as examiner -together with an external examiner -in the final assessment of the portfolios.Besides, we did not ask the students how they used the feedback.This must be taken into consideration when interpreting the results.At the same time it is worth noticing that the person presenting the feedback is among the longest-serving professors of psychology in Norway, acknowledged as one of the nation's top scholars.When asked to present his view on the question he writes: "After having read more than two thousand essays in social psychology over a period of three years, I was very pleased by being able to conclude that the vast majority of the students clearly demonstrated how the written feedback led to a clear improvement in the end result." Over the six semesters approximately 90% of the students in the portfolio groups completed the course (ranging between 87 and 98), whereas the same was true for only 40% of the students in the exam groups.It was also found that students in the portfolio groups produced better results than the exam groups (Raaheim, 2003).The great difference in marks and failure rate between students in the portfolio groups and students who followed the regular programme may, of course, be explained in many different ways.One cannot rule out the possibility of a systematic difference, with the portfolio groups consisting of brighter students and/or more motivated students.It is also possible that time (to study) is an important element, as close to all students in the portfolio groups were full-time students, whereas this does not necessarily have to be true for the other groups.All of this can not, however, rule out the fact that nearly 20% of the first drafts in the portfolio group in Spring 2003 would not have passed, and that this is very similar to the picture found over the course of many years at this level among students who do not receive any feedback as they sit for the traditional exam.
The students in our group were allowed to co-operate with another student on one of the three papers.Very few students did, however, grasp this opportunity.Again, there may be many explanations.One may, for example interpret this as a general mistrust in the benefits of co-operation, based on real or imagined events.When asked at the end of the course why they did not choose to co-operate with another student, a very common reply was that they feared that the other person would not do his or her bit.Social loafing is a well-known phenomenon, and it would not be the first time that the fear of "free-riders" would prevent co-operative learning.If one wishes to introduce new ways of teaching and assessment within higher education, and if this implies that students have to work together in pairs or in groups, one does well to acknowledge that students, in general, have little experience in doing so.In order to achieve what Brown (2000) has called 'social labouring', students need guidance and training in working together.One may also choose to devise a system in which the group takes control and involves all partners, and where stronger students teach weaker students as shown by Bartlett (1995).
As Thompson (1994) argues, a system in which the lecturer presents extensive written feedback to the students may prove to be time-consuming, especially when faced with large groups.In our case we paid an external expert to do this, nearly tripling the costs as compared to the costs associated with the traditional exam.There is, however, much to be saved in time (and money) by refining the feedback procedure.In both our study and the one described by Thompson, students received extensive written feedback.In our case the feedback was presented on a standard feedback form and related to some prespecified criteria.As one reads many essays belonging to a particular course, one will, inevitably, experience that some mistakes and misunderstandings are repeated.Instead of supplying every student with an individual written feedback, one may use a feedback form with pre-specified categories and simply tick off for a particular category statement.In this way a lecturer may be able to give feedback to very many students in the course of relatively short time.However, the best way to handle this challenge is, probably, to develop a system in which students give feedback to each other.In the Norwegian context this certainly seems to be the correct way to go.A recent study shows that after the introduction of the Quality Reform in 2003, lecturers report that they spend much more time on giving students written feedback on individual papers than they did prior to the reform, and that they have less time to do research (Michelsen & Aamodt, 2006).Such peer feedback would have to be in writing and both receiver and provider should keep a copy of this in their portfolio.The portfolio would not be considered acceptable unless it included both reports.One might elaborate on this, and ask the students to document how they have considered the feedback provided by one or more of their peers.Such a system is well-founded on research, some of it referred to in this paper.There are some prerequisites; students would need to be trained in how to present feedback.This is especially important in introductory courses one might add, as research shows that students in advanced courses are more accurate assessors (Falchikov & Boud, 1989).It is also important that the criteria on which the essay is to be assessed are made explicit to all parties (Dougen, 1996).As is discussed by Orsmond, Merry & Spelling (2006), one ought to also be more careful in evaluating how effective the feedback really is.

Conclusions
In this paper we have tried to shed some light on the relationship between feedback and learning.Feedback is not always helpful to learning.If it is vague, general, focuses on the negative, arrives late, or is unrelated to assessment criteria, it does not do much good.The process of giving feedback to students is time-consuming and/or costs much money.In order to secure best value, assessment criteria have to be spelled out, and feedback must adhere to these criteria.As underlined by Orsmond, Merry & Reiling (2006), tutors ought to also evaluate how effective their feedback has been.
In our study students were supplied with feedback on a standard feedback form, and we argue that the feedback they received was helpful and increased the quality of the written work.We paid an external expert to provide the feedback.In the paper we discuss alternative ways of doing this.One alternative is to make use of the students themselves.In cases where a system of peer feedback is introduced, it is important that students are given training in how to present feedback.

Appendix
1 A full portfolio of 5000 words would thus be equivalent to the average essay produced at a 4-hour exam.