An extended approach to adjust inconsistent minority peer- and self-assessment scores of teamwork using assessor’s reliability

Teamwork is generally assessed either solely by academic staff or by both academic staff and students themselves confidentially as well as collaboratively. Peer- and self-assessments have been used primarily to assess teamwork process and teacher assessment to assess teamwork product. Peer- and self-assessments are useful to elicit team members’ contribution towards teamwork and to convert team mark into individual marks, provided the scores are reliable (the extent to which the scores are consistent). However, not all peer- and self-assessment scores are reliable. Anecdotal and literature evidence suggest that there are several cases of inconsistencies in these scores. Individual contribution scores given by teammates to an assessee (including himself/herself) can sometimes vary significantly due to both intentional and unintentional reasons. Simply using total individual rating scores without considering an assessor’s reliability to estimate individual contribution factors can sometime results unfair grades and becomes hindrance to learning through teamwork. PURPOSE This study proposes an extended approach to adjust inconsistent and/or distorted minority peer and self-assessment scores of teamwork using standard normal probability concept. APPROACH In order to adjust inconsistent and/or distorted minority peer-and self-assessment scores of teamwork, an extended approach has been proposed. The approach uses the reliability of assessor’s scores of an assessee using standard normal probability curve. The evaluation of the extended approach is conducted by comparing with the existing approaches using two case examples of peer- and selfassessment of teamwork where minority team members’ scores are inconsistent. RESULTS The evaluation of the extended approach shows that the proposed method is superior to the available approaches in order to adjust inconsistent peer- and self-assessment scores for special cases where scores of minority team members are inconsistent. The extended approach helps both to automatically detect such scoring anomalies and to adjust the scores so that the fairer contributions to the teamwork would be obtained and utilised. CONCLUSIONS The extended approach is useful in that it helps both to automatically detect scoring anomalies and to devise the methods to adjust them. However, the approach does not address the issue of scoring inconsistencies by majority of team members as it uses average score as a basis for identifying inconsistencies. Moreover, the approach needs to be implemented in the real teamwork environment in order to identify the impacts of these scoring adjustments in teamwork process and teamwork product.