Research Chart Axis and Denominator Checker

Interpret the output accurately

For the built-in fictional example, two bar heights appear to start at 70 and end at 72 and 78. Their heights above the displayed axis start are 2 and 8, a 4:1 visual-length ratio, while the underlying values are 78 / 72 ≈ 1.083. The absolute change is six units, and the relative change is 6 / 72 × 100 ≈ 8.33%. That does not mean the higher observation is four times as large. The tool separates these calculations and does not assert that any real paper is misleading.

When studying percentages, enter each numerator and denominator. A fictional 18 of 40 compared with 30 of 50 gives 45% and 60%: a 15-percentage-point gap and a 33.33% relative increase from the first rate. Those are not interchangeable statements. Read our research graph interpretation guide for questions about chart types, confidence intervals, missing data, visual scales and appropriate claims. The local CSV checker can flag structural data issues before a chart is built; the statistical test checker covers a different question about research design.

What an axis checker does not prove

The UK Office for National Statistics distinguishes line-chart positions from bar-chart lengths. A line axis need not always start at zero; the relevance of a range depends on what the chart shows. Its bar-chart guidance explains why a hidden or nonzero bar baseline can change perceived magnitude. Values must still come from the actual source, with comparable units and measurement definitions. No automatic numerical test can establish causation, statistical significance, honest sampling or an author's intention.