Review by: Brahm, Taiga (2009-04-16)

Assessments of student learning, in particular, peer assessment have seen many developments in recent years. Peer assessment can be defined as a process through which groups or individual students rate their peers (van Gennip et al., 2009, 41). Peer assessment does not only include the evaluation of the outcomes but also of the process of learning. The authors of the paper reviewed studies on the effect of peer assessment on learning, taking into account the role of interpersonal variables and structural features influencing peer assessment. As interpersonal variables the authors identify psychological safety, value diversity, interdependence and trust. Psychological safety is defined as the perception that it is safe to take risks among peers. Value diversity can be described "as a difference in opinion of what a team's task, goal or mission should be (van Gennip et al., 2009, 43). With regard to interdependence, the authors differentiate between outcome and task interdependence. Outcome interdependence is seen as the belief of team members that their personal benefit depends on the success of the team, while task interdependence refers to the connectedness of the subtasks that make collaboration necessary. Trust is defined as the confidence in oneself to rate others fairly and to provide constructive feedback (p 44). In addition to the interpersonal variables, the authors identified structural features of peer assessment. These were grouped in three clusters:

• assessment description including the purpose, outcomes, place and time of the assessment
• the interaction, i. e. the directionality, privacy and kind of contact
• the composition of the feedback group, e. g. across different years of studies, abilities, and the number of people assessed resp. assessing

The main research question of the article is to what extent the outcomes of peer assessment on learning are related to interpersonal variables and to structural features of the peer assessment format (p 45).

The study was conducted as a literature review based on a sample of 15 articles which are listed in a table. They range from 1991 to 2006 and include various research designs. The studies were investigated with regard to interpersonal variables and structural features.

In terms of interpersonal variables, only 4 out of the 15 studies include interpersonal processes as variables under study. Among them, value diversity and interdependence did not appear at all, psychological safety was mentioned in one study and trust in three studies. However, these variables were measured but not related to learning benefits. A number of studies looked at structural features of peer assessment (cluster 2 and 3 of the clusters mentioned above). In the following, some selected results will be briefly summed up:

• With regard to the type of feedback three combinations of feedback could be identified: (1) reciprocal, public and face-to-face feedback, (2) mutual, public and face to face and (3) mutual, confidential and distance feedback. These combinations were almost equally divided among the articles. There was no direct connection of the clusters to the performance of the students.
• In terms of the feedback provider, two combinations of features emerged: (1) assessors and assessees have similar ability and both are individuals (instead of groups or pairs of students), (2) assessors and assessees have similar ability but the assessees are evaluated in groups or pairs. Again, no connection can be found between the outcome measure applied on the one hand, and the combination of structural features in the peer assessment process on the other hand (p 51).

In sum, the article aims at disclosing the effect of peer assessment on learning from a social perspective. It is based on a literature studies since the 1990s including 15 articles, thus, giving a good overview of the current state of the research. However, the selection of interpersonal variables seems to be quite random. It is based on three studies which are closely connected (in the sense that the latter are referring to the former), however, in the paper, it never becomes clear why these interpersonal variables and not others (e. g. cohesion of the students) were chosen (p 42). Additionally, the term trust does not seem to fit the interpersonal state that it is referring to. These seem to be more interrelated with self- or group efficacy, i. e. the confidence in oneself or the group of peers to fulfill the task at hand, here peer assessment.