Figure 2.

Hand scoring inter-rater reliability.

(A) % Identical observations for the comparison frames was calculated for each observer-observer pair. Darker colors indicate higher % identical observations. The red bar on the scale shows mean % Identical observations across all observer-observer pairs. (B) % Identical observations is broken down by # of behaviors present during a trial, ranging from 4 to 8. Each individual point indicates a single observer-observer trial comparison. Large Xs indicate the mean % identical observations for all observer-observer comparisons for each # of behaviors. (C) A confusion matrix shows all observer-observer judgment pairs for the comparison trials. Y axis is observer 1 and x axis is observer 2. % Behavior observations is plotted by row with black indicating 100%, white 0% and shades of gray in between. The number of behavior observations is indicated for each in parentheses. Asterisks indicate behaviors for which fewer than 10 instances of the behavior were observed.