Interpretation of three-dimensional versus two-dimensional data presentations

This study investigated the relative effectiveness of three-dimensional and two-dimensional presentations of data. Forty subjects (20 males and 20 females) participated in a 2-X-4-X-4-X-5 mixed factorial design with four levels of data format, four levels of questions, and five levels of trials. Two column charts drawn using different projection techniques (isometric or cabinet) were operationally defined as the 3-D data formats. A 2-D grouped column chart and a table of numbers were defined as the 2-D formats. Response time data and the number of correct responses were measured. Response time results indicated that overall subjects using the 3-D formats were significantly slower than subjects using the 2-D formats for three of the four question types. Further the question types appeared to vary in complexity regardless of the data format used. Response times for all data formats increased as question complexity increased; however, the increase was greatest for the 3-D formats. Response time decreased across trials for all data formats. Also, females took significantly longer than males using the isometric format; however, there were no other sex differences using any other data format. Correct response data indicated that accuracy was not significantly affected by data format although there was a trend towards poorer performance using the cabinet format. Subjects did make significantly more errors when the question type required them to perform minor calculations. The implications of the findings are discussed along with suggestions for further research.