Effects of Frequency of Data Collection and Graph Characteristics on Visual Inference

Abstract

Though visual inference is the most frequently used procedure in the analysis of behavioral and performance data, accuracy and reliability continue to be of concern. This study addressed the questions of whether different trends on graphs produced different judgments, and whether different frequencies of data collection produced different judgments. Also addressed was the question of whether there was an interaction between trend and frequency that influenced judgments of student performance. Results indicate that judgments tend to be accurate and consistent when performance data represent systematic improvement, but when graphed data represent a decrease in performance, no change, or highly variable performance, judgments tend to differ by frequency.