Student Feedback at CUED (2006)

We use a variety of feedback methods. Each have their strengths. All have proved useful. We are moving towards computer-mediated feedback because of the economies of scale (we have about 250 students per year, and 100s of lecture courses). We hope that the range of alternative methods and the general computer literacy of our students will avert sample bias.

Course-specific

• When practical experiments are set up, a feedback icon can be made available, allowing students to e-mail the experiment supervisor. Setting one of these up takes a few minutes.
• Lecturers are encouraged to offer the students ‘lecture feedback forms’
• The Interact Project developed a number of computer-based courses. Assessment was extensive and varied (video, debriefing, questionnaires, detailed observations of students, external assessors). When there was online documentation it was possible to find out what students needed to look up.

General

• A ‘Fast Feedback’ icon is on the standard desktop. This user-friendly front-end to the mailer lets students send e-mail straight to the teaching office. In November 1998 we updated this so that student had to select (using cascading menus) an appropriate subject title for their message. An “anonymised” copy of the message is sent straight to the staff member concerned. In 2003-2004 its average use was perhaps 1 or 2 times a day.
• The WWW-based help system has a number of pages where feedback is elicited. We also monitor how many times each help document is read so that we can concentrate our re-writing efforts effectively. Students can (and have) provided material – the way students help each other is often different to the methods and wordings staff might use.
• The ucam.eng.suggest newsgroup is used by some students to contact staff. The ucam.eng.students newsgroup is used by students to talk amongst themselves.
• The Staff-Student Joint Committee meet regularly. There is a postbox where students can leave messages for them. The SSJC also have an e-mail address and a Home Page.
• In November 1998 the online survey was upgraded so that students could complete it incrementally. A ‘Survey’ icon is on the standard desktop.


With 200-300 students in each year, organising a paper-based questionnaire becomes prohibitively time-consuming. Fortunately our students have at least some familiarity with computers, we have many computer terminals available to undergraduates, and we have on-site programmers. So in 1990 the Teaching Office suggested that a computer-based survey should be produced. We now use computers to get various types of input from students that in the past were done using paper. We haven’t yet reached ‘questionnaire fatigue’ levels, but response levels drop as the students get older.



All years are invited to do an online survey. It is easy to run – they type “survey” or click on the survey icon. For each course this asks about the quality of presentation and handouts, whether the work was too hard or too quickly presented, and whether the examples papers took too long. Then there are sections asking about coursework, supervisions and the course as a whole. Finally students have a chance to comment on whatever they like.

The results are processed into histograms that are widely distributed amongst academic staff. The comments are passed to the lecturers concerned via the Teaching Office.



A whole survey takes over 15 minutes to fill in, but students can fill it in a little at a time. It’s important to balance the need for extensive information with the need to make the questionnaire quick and easy to complete.


Input is treated as anonymous. This has many advantages. The disadvantage is that staff can’t reply to, or ask for more information from, the students.


We encourage participation by offering prizes, reminding students at lectures, reminding them when they log in, and putting up posters. Response rates have in general been falling






CUED online survey response rates

	1st	2nd	3rd	4th
1995	75%	57%	90%
1996	80%	48%	72%	50%
1997	49%	40%	48%	34%
1998	63%	56%	55%	48%
1999	43%	33%	27%	28%
2000	27%	21%	18%	16%
2001	21%	22%	18%	12%
2002	33%	22%	25%	20%
2003	22%	18%	17%	8%
2004	51%	37%	35%	30%

Implementation Notes

Our system is Unix-based. People have to login before they can use it. This made user-validation easy for non-WWW feedback mechanisms.

The first surveys were handcrafted using Xlib graphics. The Teaching Office provided the list of questions that they wanted asking. When details change little from year to year, little work is involved by the Teaching Office or the Computer Staff. With the change from a 3 year to a 4 year course however, it’s been harder to reduce the workload on the programmers. However, now that the syllabus has stablised we have developed other methods of getting student input with minimal programmer effort.

• Assessment of german/french ability is done using HTML’s forms facility. This lacks flexibility, but it may be the preferred option for short, simple questionnaires.
• The 1998 surveys were produced using an inhouse survey generator called surgen which along with the processur post-processor should considerably simplify future survey production.


• In January 2000 fffmaker was installed to automate the production of Fast Feedback programs and feedback pages within surveys.


• In May 2001 we installed a web-based version of the survey which is generated using a program called wwwgen.


• In September 2001 we added password-protect (rather than CUED-only domain protection) to the web-based survey.


• In September 2002 we introduced a web-based password-protected Fast Feedback facility.


• In 2004 more automation was brought in so that the Teaching Office could
  test, monitor and process the survey.