Peer feedback: the learning element of peer assessment
Published in Teaching in Higher Education, 11, (3), 279-290.
Ngar-Fun Liu and David Carless
Hong Kong Institute of Education; University of Hong Kong
This paper focuses on peer feedback in relation to assessment processes. It examines the
rationale for peer feedback, emphasising its potential for enhanced student learning. We draw on
relevant literature to argue that the dominance of peer assessment processes using grades can
undermine the potential of peer feedback for improving student learning. The paper throws
further light on the issue by drawing on a large-scale questionnaire survey of tertiary students
(1,740) and academics (460) in Hong Kong, supplemented by interview data. The findings
indicate that a significant number of academics and students resist peer assessment using grades
and that the majority report that students never or rarely grade each other in assessment
activities. This paper explores why there is resistance, in particular, by academics to peer
assessment and argues the case for a peer feedback process as an end in itself or as a precursor to
peer assessment involving the awarding of marks. It also recommends some strategies for
promoting peer feedback, through engaging students with criteria and for embedding peer
involvement within normal course processes.
Introduction
It is commonly accepted that there are two main purposes of assessment: a certification
(or summative) purpose and a learning (or formative) purpose. The first is usually
regarded as dominant, with students frequently being reported as driven by a natural
desire for high grades (e.g. Becker, Geer & Hughes, 1995), even when such instrumental
motivations may lead to adverse impacts, such as surface learning (e.g. Ramsden, 2003).
The authors are currently involved in a project which, whilst acknowledging the
dominance of the summative paradigm, seeks to place emphasis on the purpose of
assessment related to the promotion of learning. This initiative named the
learning-oriented assessment project (LOAP) aims to identify, promote and disseminate
good practices at the interface of assessment and learning in higher education in Hong Kong.
Our emphasis on ‘learning-oriented assessment’ (Carless, 2006) indicates that we believe the
learning aspects of assessment have considerable potential to be exploited more fruitfully and
the project promotes strategies such as peer and self-assessment, which we view as more
about learning than about assessment.
1
The term assessment is often interpreted as referring to marking, grading, measuring or
ranking and as a consequence peer assessment is regarded mainly as students giving
marks or grades to each other. However, peer involvement can be more than the teacher
sharing with students the responsibility of grading. A reliable assessment depends on
knowing what one is trying to assess and by what means one comes to an accurate
judgment. Once we begin asking questions about how assessors arrive at marks, we are
involved in a process of defining learning outcomes and the criteria for assessing those
outcomes. This intellectual engagement with outcomes, criteria and standards is at the
heart of student involvement in assessment and can lead to greater clarity about the
nature of high quality performance. Engaging learners in thinking about achieving
outcomes to certain agreed standards is a learning process and giving marks or grades is
only part of that process.
Before proceeding, we indicate how we use the terms peer feedback and peer assessment. By
peer feedback we mean a communication process through which learners enter into
dialogues related to performance and standards. Peer assessment is defined as students
grading the work or performance of their peers using relevant criteria (Falchikov, 2001).
So our distinction between the two terms is that peer feedback is primarily about rich
detailed comments but without formal grades, whilst peer assessment denotes grading
(irrespective of whether comments are also included). Whether grades are awarded or
not, the emphasis is on standards and how peer interaction can lead to enhanced
understandings and improved learning.
This paper synthesises relevant literature and draws on data from LOAP to make a case
for a peer feedback process as an end in itself or as a precursor to peer assessment using
grades. We suggest that peer feedback has greater potential for learning than peer
assessment, but that some combination of the two may be necessary for pragmatic
reasons. Our argument is developed in four stages: firstly, we state the rationale for peer
feedback processes; then we suggest that the literature on peer assessment has
over-emphasised discussion of the reliability of student grades; thirdly, we draw on
LOAP data and outline possible reasons for resistance to peer assessment; finally, the
fourth main section suggests some possible ways forward for the implementation of peer
feedback. Our discussion is stimulated by the Hong Kong context, but detailed
contextual background is not provided as we seek to raise arguments of a general nature
2
for consideration across international contexts.
Rationale for peer feedback in assessment processes
The conceptual rationale for peer assessment and peer feedback is that it enables students to
take an active role in the management of their own learning. It is an element of self-regulated
learning (Butler & Winne, 1995) by which students monitor their work using internal and
external feedback as catalysts. “Self-regulated learners seek feedback from external sources
such as peers’ contributions in collaborative groups” (Butler & Winne, 1995, p. 246). In their
model of formative assessment and self-regulated learning, Nicol & MacFarlane-Dick (2005)
also contend that by commenting on the work of peers, students develop objectivity in relation
to standards which can then be transferred to their own work.
The link between peer assessment and self-assessment is salient. Peer feedback can enable
students to better self-assess themselves as some skills are common to both peer and
self-assessment. Boud (1991) defines self-assessment as “the involvement of students in
identifying standards and/or criteria to apply to their work and making judgments about
the extent to which they have met these criteria and standards” (p.4). Self-assessment
can be enhanced by peer contributions which may take the form of questions, comments
or challenges which prompt one to reflect on what has been done (Boud, Cohen &
Sampson, 1999). Boud (1995) highlights the interplay of self- and peer assessment as
follows, “the defining feature of self-assessment is that the individual learner ultimately
makes a judgement about what has been learned, not that others have no input to it”
(p.200). Peers provide rich information which can then be used by individuals to make
their own self-assessments (Boud, 1995) and follow up with actions to improve their
work.
What other reasons are there to engage with peer feedback as a means to promote
learning? After all, there are many ways to enhance learning. The answer is a pragmatic
one that we suspect many students would find appealing. There is evidence that peer
feedback enhances student learning (Falchikov, 2001) as students are actively engaged in
articulating evolving understandings of subject matter. Peer feedback thus carries
potential for improved performance in high-stakes assessments, something obviously
highly valued by students.
3
A further practical reason for peer feedback is that students would receive more feedback
from peers and more quickly (see Gibbs, 1999, for an example) than when academics are
providing comments. With increasing resource constraints and decreasing capacity of
academics to provide sufficient feedback, peer feedback can become a central part of the
learning process, rather than an occasional option.
A further important reason for engaging learners with peer feedback is that learning is
likely to be extended from the private and individual domain, to a more public (i.e. to
one or more peers) domain. One important way we learn is through expressing and
articulating to others what we know or understand. In this process of self-expression, we
construct an evolving understanding of increasing complexity. One aspect of this process
is providing learners with opportunities to explore and articulate criteria and standards in
the context of working on specific assessment tasks. In order to clarify notions of quality,
learners need to analyse real, illustrative exemplars. This is where examining the work of
peers offers meaningful opportunities for articulating discipline-specific knowledge, as
well as criteria and standards. Once students are at ease with making their work public,
we could create conditions under which social learning might be facilitated. This is not
to say that making one’s work public may not sometimes be threatening, but there are
ways to minimise this risk which are discussed in the final section of the paper.
Focus on reliability of peer assessment
The existing literature on peer assessment has been dominated by studies of peer-tutor
grade correlations (Sluijsmans, Brand-Gruwel & van Merriënbor, 2002). For example,
Falchikov & Goldfinch (2000) carried out a meta-analysis of 48 quantitative peer
assessment studies that compared peer and teacher marks, demonstrating that students
are generally able to make reasonably reliable judgements. Stefani (1998) in her critique
of assessment practices reflects on the dominance of peer grading:
“Many academics became tied to the quantitative analyses of innovative assessment
procedures because of the extreme pressure to ‘prove’ that students could be as
reliable as ‘assessors’ as the tutors … what some staff seemed to be doing was
reducing the concept of student learning and student empowerment, to a series of
correlation co-efficients” (p. 343).
Yet, the focus (arguably overemphasis) on reliability of student grading is still prominent
as evidenced by further recent examples of measurement-focused forms of peer
4
assessment e.g. Miller (2003); Segers & Dochy, (2001).
Whilst acknowledging that establishing the reliability of student judgements is an
important issue, we feel that it is now well-recognised that students are reasonably
reliable assessors. We believe (like Stefani) that peer involvement in assessment
processes should be more than just grading peers’ work and comparing the scores with
those of the tutor. The literature also provides some support for a focus on peer feedback
rather than peer assessment. Sluijsmans et al., (2001) found that students felt
uncomfortable in awarding grades and preferred just giving feedback. Boud, Cohen &
Sampson (1999) reflect as follows, “the use of peer assessment in which students make
formal assessments of others within a working group can inhibit cooperation” (p.421).
They point out a potential contradiction between a collaborative learning process and
individual assessments which often carry a competitive flavour. Boud (2000) argues that,
“many forms of peer assessment are ineffective. These are processes in which peers are
used as surrogate assessors to generate grades” (p. 157). This generation of grades
belongs to the summative rather than the formative purpose of assessment. Brown, Bull
& Pendlebury (1997) argue that resistance by students to informal peer feedback is rare,
resistance to formal peer assessment for summative purposes is relatively more frequent,
based on three reasons: dislike of judging peers in ways that ‘count’; a distrust of the
process; and the time involved.
In sum, our belief is that whilst peer assessment involving grading has been much
discussed in the literature, in terms of learning, there are more substantive arguments in
favour of peer feedback rather than peer assessment. Falchikov, probably the most
prominent writer on peer involvement in assessment, seems to reach a similar conclusion
as evidenced by the evolvement of her practice towards peer feedback (Falchikov, 1995,
2001, 2005) rather than peer assessment.
Resistance to peer assessment
This section focuses on data collected from a large-scale questionnaire survey on
assessment beliefs and experiences returned by 1,740 tertiary students and 460
academics in Hong Kong (Liu, 2005). This survey is part of LOAP activities referred to
in the introduction. Here we use data from just the items most relevant to our discussion
to illuminate the issue. It is not our contention that our argument is driven by the data,
5
