This paper describes two case studies that examined statistical reasoning skills in an authentic environment in which learning occurred in small groups. Multiple forms of assessments were developed to obtain a detailed profile of reasoning demonstrated by individual students and groups of students on projects. Two conditions were developed by individual students and groups of students on projects. Two conditions were developed to exemplify assessment criteria on such projects: a library of exemplars condition and a text condition. Qualitative analyses of verbal protocols during group discussions and presentations indicated that the library of exemplars was effective in promoting (a) reasoning about data analysis and data presentation issues (b) planning and (c) aligning students ratings of projects with experimenter ratings. The importance of student participation and request for assistance in group situations was highlighted in this study. Moreover, information about the reasoning demonstrated by individual students is required to ensure that group assessments do not overestimate student performance.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education