Related Experiment Videos
Asking for Less and Getting More: The Impact of Broadening a Rater's Focus in Formative Assessment
Walter Tavares1, Alexander Sadowski, Kevin W Eva
1W. Tavares is scientist and assistant professor, Wilson Centre, Department of Medicine and Post-MD Education, University Health Network/University of Toronto Faculty of Medicine, Toronto, Ontario, Canada, and clinician scientist, Department of Community and Health Services, Paramedic and Senior Services, Regional Municipality of York, Newmarket, Ontario, Canada. A. Sadowski is a research associate, Wilson Centre, University Health Network/University of Toronto Faculty of Medicine, Toronto, Ontario, Canada. K.W. Eva is senior scientist, Centre for Health Education Scholarship, and professor, Department of Medicine, University of British Columbia Faculty of Medicine, Vancouver, British Columbia, Canada.
Purpose:
There may be unintended consequences of broadening the competencies across which health professions trainees are assessed. This study was conducted to determine whether such broadening influences the formative guidance assessors provide to trainees and to test whether sequential collection of competency-specific assessment can overcome setbacks of simultaneous collection.
Method:
A randomized between-subjects experimental design, conducted in Toronto and Halifax, Canada, in 2016-2017 with paramedic educators experienced in observing/rating, in which observers' focus was manipulated. In the simultaneous condition, participants rated four unscripted (i.e., spontaneously generated) clinical performances using a six-dimension global rating scale and provided feedback. In three sequential conditions, participants were asked to rate the same performances and provide feedback but for only two of the six dimensions. Participants from these conditions were randomly merged to create a "full score" and set of feedback statements for each candidate.
Results:
Eighty-seven raters completed the study; 23 in the simultaneous condition and 21 or 22 for each pair of dimensions in the sequential conditions. After randomly merging participants, there were 21 "full scores" in the sequential condition. Compared with the sequential condition, participants in the simultaneous condition demonstrated reductions in the amount of unique feedback provided, increased likelihood of ignoring some dimensions of performance, lessened variety of feedback, and reduced reliability.
Conclusions:
Sequential or distributed assessment strategies in which raters are asked to focus on less may provide more effective assessment by overcoming the unintended consequences of asking raters to spread their attention thinly over many dimensions of competence.