The effect of the level of data presentation on visual analysts’ decisions.

Abstract

Single-case design (SCD) is frequently utilized in research and applied settings to evaluate the effect of an intervention over time. Once collected, single-case data are typically graphed and analyzed visually in both research and practice contexts. Despite the ubiquity of visual analysis in SCD, this analytic framework has often been critiqued due to findings of limited reliability across visual analysts. Recent research has identified that the way a graph is constructed may contribute to the limitations of visual analysis. The present study sought to evaluate the effect of visually representing multiple measurement occasions as a single data point (e.g., combining measurements taken daily into a weekly composite data point) on visual analysts' decisions regarding the magnitude of an intervention effect. Eleven participants viewed identical data sets, plotted to show different numbers of measurement occasions combined as a single data point, and provided ratings regarding the magnitude of intervention effect depicted within the graph. Results indicated a significant main effect, with data sets with higher levels of data combination being rated as demonstrating significantly larger intervention effects. The results of the study provide additional support for standardization of data presentation and graph construction within SCD in both research and practice contexts. Impact and ImplicationsThe manner in which a single-case graph is presented has an impact on rater interpretation of intervention effectiveness. Practitioners and researchers are encouraged to avoid combining multiple measurement occasions into single data points when doing so obscures or alters data characteristics such as trend and level changes of data.