🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
dual_axis_manipulation
Dual-axis manipulation uses two y-axes with deliberately chosen different scales to make visually unrelated or weakly correlated time series appear to move in lockstep. By freely choosing the scale, minimum, and maximum of each axis, any two series can be made to appear strongly correlated regardless of their actual statistical relationship.
A chart showing vaccine rates (left axis, 60-80%) and autism diagnosis rates (right axis, 1.0-1.8%) are both rising over the same period. The axes are scaled so the lines appear to track each other closely, implying a relationship that a scatterplot would show to be confounded by secular trends.
A cable news segment displays a chart with two lines: monthly unemployment claims (left axis, scaled 200,000–400,000) and a politician's approval rating (right axis, scaled 38%–42%). The axes are chosen so both lines appear to move in near-perfect lockstep, visually implying a tight relationship that is far weaker than it appears numerically.
A pharmaceutical company's investor presentation overlays drug trial enrollment numbers (left axis, 100–500 participants) with stock price (right axis, $12–$18) on the same time series chart. The scales are set so the lines appear to rise together dramatically, implying the enrollment growth is driving stock performance — a visual suggestion that the data does not actually support.
Binary (yes/no) questions an LLM must answer to identify this aspect:
Does the chart use two y-axes with different scales?
Type: binaryAre the scales chosen to make the two series appear to move together when they may not?
Type: binaryWould the apparent correlation between the series disappear if both were plotted on the same scale or as a scatterplot?
Type: binaryIs a causal or correlational interpretation being drawn from the visual alignment of the two series?
Type: binaryDual-axis manipulation uses two y-axes with deliberately chosen different scales to make visually unrelated or weakly correlated time series appear to move in lockstep. By freely choosing the scale, minimum, and maximum of each axis, any two series can be made to appear strongly correlated regardless of their actual statistical relationship.
Viewers naturally read visual proximity as evidence of correlation. The choice of axis scale is invisible — most viewers do not examine axis values carefully and assume the visual pattern reflects the data relationship.
Convert dual-axis charts to scatterplots or index plots to reveal the true correlation structure. Always check both axis scales and origins. Ask whether the data would be plotted the same way if the axes were swapped or rescaled.
Political advocacy groups routinely use dual-axis charts to imply causal relationships between policies and outcomes. The technique is considered a data visualization red flag by statisticians.
Use these tools to detect, analyze, or train this aspect.