Cleveland & McGill Perceptual Rankings

In 1984, statistician William Cleveland and psychologist Robert McGill published a landmark paper in the Journal of the American Statistical Association that fundamentally changed how we think about chart design — by proving, through controlled experiments, that not all visual encodings are created equal. Their study asked participants to judge quantitative values encoded in different visual forms and then measured how accurately people could decode each one. The results produced a clear ranking: position along a common scale was the most accurate encoding (think bar charts or dot plots), followed by position on non-aligned scales, then length, then angle and slope, then area, and finally color saturation and density at the bottom. This hierarchy — position beats length beats angle beats area beats color — became the empirical backbone of visualization best practices and is still taught in virtually every data visualization course today. Before Cleveland and McGill, chart design was largely guided by aesthetics and convention; after them, it was grounded in perceptual science. Their work showed that a pie chart, which encodes data as angles and areas, is objectively harder for humans to read accurately than a bar chart, which uses position — giving scientific weight to what many designers had suspected but couldn't prove. Cleveland went on to develop the dot plot as a superior alternative to the bar chart and published the influential books 'Visualizing Data' (1993) and 'The Elements of Graphing Data' (1985), further cementing the idea that visualization design should be driven by how human perception actually works, not by tradition or decoration.