The Three Dimensions
Visual analysis examines three primary dimensions of graphed data: level (the average or typical value of the data within a condition), trend (the direction and rate of change over time within a condition: increasing, decreasing, or flat), and variability (how much the data points spread around the level or trend line). Changes in any of these dimensions between conditions may indicate a treatment effect.
Within-Condition Analysis
Before comparing across conditions, analyze the data within each condition. A stable baseline (low variability, no trend) provides the best foundation for comparison. If the baseline shows an improving trend, it is harder to attribute improvement to the intervention. If the baseline is highly variable, you need more data points to establish a stable pattern.
Between-Condition Analysis
Compare level, trend, and variability across adjacent conditions. Look for: an immediate change in level when the condition changes (immediacy of effect), a change in trend direction or slope, and a change in variability. The more dimensions that change, and the more immediately they change, the stronger the evidence for a functional relationship.
Practical Tips for Exam Questions
When shown a graph on the exam: (1) Identify the design type first. (2) Identify each condition and its phase change lines. (3) Look at the baseline data for stability. (4) Compare the treatment phase to baseline on level, trend, and variability. (5) Ask whether the data show replication (at least three demonstrations of effect). (6) Consider whether alternative explanations could account for the change.