Last changed 28 July 2007 ............... Length about 3,000 words (22,000 bytes).
(This document started on 21 Jan 2007.) This is a WWW document maintained by Steve Draper, installed at You may copy it. How to refer to it.

Web site logical path: [] [~steve] [localed] [dropout] [this page]

Early intervention and supporting retention in level 1

By Steve Draper,   Department of Psychology,   University of Glasgow.

Contents (click to jump to a section)


This is a note about detecting level 1 students who seem to be slipping away, and the possibilities for improving student retention by interventions. The background concern prompting this discussion is that improving retention is a current university strategy objective; and Iain Allison (or here) has obtained LTDF funds to share information on students within a faculty.

The general question is whether we could and should do more to detect and act upon students who do not seem to be successfully engaged in a course early on, especially if this were to improve student retention rates. This note outlines current moderately proactive practice at this in one department, and then discusses what would be generally involved in such schemes. The main separate issues are: detection, what interventions could usefully be done, and whether the best locus of action is the course team or a central faculty unit.

Hitherto standard university operating policy has not required or been to keep tabs on students except at exams. This means that only at the end of semester exams in January, or even the end of session exams in June, must failing, or failing to attend, lead to action. By then the student is often in a position from which no recovery is possible. The general issue then is whether and how we could do better in any relevant sense of "better". Most courses have, or potentially could have, earlier indicators such as tutorial attendance.

Current (2006-7) practice in level 1 psychology

The practice on the psychology level 1 course, as it has been for a couple of years, is currently (2006-7) more active than the minimum, as follows. This deals in detail with semester 1 of level 1: it might be applied elsewhere too.

  1. What is tracked?
    1. Tutorial attendance: once per week from week 3 to 9 inclusive. Tutors return attendance to the course team within a day.
    2. Submission of coursework for credit: a deadline early November, another early Dec.
    3. Other coursework that could be acted on but isn't: lab quizzes (every 3 weeks).
    4. Gaining adequate marks on coursework.
  2. The trigger condition for followup: missing two consecutive tutorials, or being more than 3 days late with a single piece of coursework, or failing (grade less than D) a piece of coursework.
  3. The intervention.
    1. Who? A single designated course team member (one for 1a, another for 1b, team leader for 1d)
    2. When. Within 1-3 days for 1a, 3 days for 1b.
    3. How. Email. In (1a) it is semi-automatically generated by the attendance being entered in a database.
    4. What: a form email asking why they didn't show up / hand in the work. For failing grades, an email offering special advice and appointments.
    5. If there is no response, there is not necessarily any further action but the course team leader is informed, and any enquiries may be replied to with this information.

  4. How effective is it? The impression of the course team leader Margaret Martin is that those who have just gone missing without officially leaving stay that way or tell the department explicitly by way of reply. The course team now knows within a week when they dropped out of touch. For other students, it seems to prevent most quietly slipping behind, and gets them to return, to hand work in so they complete requirements even with some marks penalty.

More details

For coursework (1b), reminder emails are done manually after about 3 days at most (once a week, soon after the last tutorial slot). If not responded to, they used to be followed up by phone calls, but this became too onerous. Any response is forwarded to the team leader. If it is still not handed in after some time, more efforts are made to pursue the matter.

For 1a, tutors return attendance the same day, and the returns are entered by a secretary to a database. Then the designated staff member reviews the list of missing students, filters it, and presses a button so that the remaining names receive an automatic email. This tool (written in the department) saved a large amount of work.

A further feature that would be practicable (but is not currently implemented) would be to have any reply emails sent via an address that automatically records them against the student's ID. This would allow the database to have a field that shows the last time a student was "heard from".

Following up missing students by phone

In 2003-4 a study was done at Glasgow University on following up students who missed attendance by phone. (Jennifer B. Hume (2004) "Can a single telephone call prevent a student from dropping out?". This is discussed further here.) This was an attempt to replicate an unpublished study in the USA that had seemed to show dramatic improvements in retention by such means. It failed (when a careful re-analysis was done) to show any effect whatever on retention. However, contrary to the expectations of most staff, it did show that (at least when carried out in this study, by a student trained by and for the Nightline student telephone help line), students were either neutral or in very many cases grateful for this: they took it as a sign (an unprecedented sign) that the department actually cared about them. This is of course a significant issue in relation to improving Tinto's "academic integration" factor which is thought to predict student retention. Experience elsewhere also suggests that this is a common response.

The importance of this study is that it shows that a real possibility for practical action is to contact students about absences without this necessarily being perceived as intrusive: and that this is so using repeated calls to mobile phones or even leaving messages with parents. There are of course bad and good ways to approach this, but the study demonstrated that with care it is possible to do systematically with no adverse feelings at all.

A bad way to do it would be in a manner that was punitive, authoritarian, blaming, and aggressive. Obviously this would not contribute to academic integration. An outline of a script for a good approach is:

The general components of any scheme

Any scheme for intervening has to make specific choices within the following general schema.
  1. What is tracked? E.g. tutorial attendance, handing in coursework, marks given for coursework.
  2. The trigger condition: define the conditions for an intervention e.g. missing two consecutive classes, getting a mark below 40% for a piece of work.
  3. Who is responsible for the intervention (must designate a member of staff, or a piece of software to generate the action). Particularly if the phone is used, this has to be done at times of day when it is most likely to be answered, not when it is convenient for the staff member. A protocol for how many repeated attempts will be made and when needs to be adopted and followed. Although the total amount of work is small, it has to be given a high priority, and often other tasks must be arranged around this one, not vice versa. Experience shows that a course team leader, even when extremely interested in trying this out, is unlikely to be able to give this the priority it needs. Secretaries are often not at work at the hours most suited for this. In many ways, hiring students to do this can be the best solution.
  4. How soon it will be done. On a human level, the swiftness of response is a vital part of the message. It is not possible to convey concern if you appear to do nothing for weeks. Furthermore, if nothing happens when the student misses one class, they naturally assume that missing another will also not matter: promptness of response is part of the message.
    A second issue is not about communication but just that a student who is not learning is slipping away from a position from which they can recover: if they are to recover, the time at which they resume learning is an important factor.
  5. What will the intervention be? E.g.

What are the types of student case involved?

The retention literature gives various alternative lists of causes, although the main lesson is that multiple causes are significant in most cases. For present purposes, we want a categorisation relevant to the purpose of thinking through how early detection of at risk students might lead to actions that improve retention. In our experience of followup schemes, we found four kinds of case that appear when absence or missed deadlines are followed up:
  1. Hard core: the student has taken the decision to leave. Some will not respond to followup (depending on the communication method: obviously if they have left they are unlike to be reading university email), others will reply to say that they have decided to leave. They often say they didn't know who to tell, and are glad for the contact.
  2. Pastoral: the student is in crisis for any one of a number of reasons, but might stay and recover their academic position. Support and advice might make a difference in this. (Certainly there are students who go through such a crisis but stay, who later said that the support of a particular person e.g. an exceptional Advisor, was crucial.)
  3. Starting to slip away. Missing something for a small reason (e.g. a transport hitch, being ill for 3 days) can lead to missing the next thing for even less reason, and so on. Low key but prompt expressions of concern may remedy this in themselves.
  4. Special remediation. Particularly in courses where each week's work depends on mastery of the previous week's, falling slightly short quickly becomes irreversible failure. Targeted remedial activities might pay dividends.

Possible interventions for each type

Prospects for sharing information across departments (courses)

Sharing information does not seem relevant to (4) students needing remediation: this is wholly a departmental matter. For the "hard core" (1) already departed type, it is merely a very minor administrative convenience and one that can easily be handled by manual emails from whomever discovers their departure. For students who are starting to slip away (3), although knowing how they were doing on other courses might sometimes be relevant a) they may be avoiding one course but not others, so sharing wouldn't help; b) in any case, direct, prompt enquiries linked to specific courses seems the most likely remedy. Sharing information on this would only be important if a central service were provided for phone followups that all departments could draw on: i.e. the departments did the detection, but a hired independent set of people trained in counselling followed up by phone. Only for "pastoral" cases (2) of a student with a crisis not specific to one course, and possibly benefitting from pastoral intervention, might pooled information be valuable. In other words, sharing information is only likely to be worthwhile in connection with one of the four types of student cases.

Much of the potential information is tightly bound to individual course practice, is not immediately sharable, and is likely to change at short notice from year to year as course teams change their practices. This is true of detection opportunities e.g. tutorial attendance, of the meaning of each detected event (missing some things is much more important than others), and of the interventions e.g. remediation activities, chasing up missed coursework deadlines.

What information might be shared i.e. comprehensible and useful to those outside the course team?

Among the many problems are that to be useful the information must be up to date e.g. updated daily. This is too much work to do manually. But automation requires course teams to adopt other people's software. Yet the people with the most information to contribute have the least to gain from return information.

This sharing of information had better be extremely little effort for course teams, otherwise the benefit will not be sufficient to warrant it. The only type where pooled information might offer a substantial advantage in both detection and remediation is a "pastoral" problem.

One model might be to offer departments (course teams) services (such as a phone students service) and software that they can use for their own purposes to save work e.g. to send automatic emails triggered by a criterion on attendance records, to receive emails from students and record the most recent date one was sent, etc. That would be an incentive for them to participate, and would support better practice in the course teams where much of the beneficial work must in any case be located.

Possible good practice tactics

Web site logical path: [] [~steve] [localed] [dropout] [this page]
[Top of this page]