Bland-Altman Method Comparison Analysis
Also known as: Bland-Altman plot, limits of agreement analysis, method agreement analysis, Bland-Altman Uyum Analizi
The Bland-Altman analysis is a graphical and statistical technique for assessing agreement between two measurement methods applied to the same subjects. Introduced by J. Martin Bland and Douglas G. Altman in their landmark 1986 Lancet paper, it plots the difference between the two methods against their mean for each subject, and derives the bias (mean difference) along with limits of agreement (LoA) that capture 95% of differences in the population.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use the Bland-Altman analysis whenever you want to compare two continuous measurement methods applied to the same subjects and assess whether they agree closely enough to be used interchangeably. Four key assumptions must hold: the differences between methods should be approximately normally distributed (verify with a Shapiro-Wilk test or a histogram of differences); if the magnitude of differences changes with the measurement level (proportional bias), a logarithmic transformation should precede the analysis; when repeated measurements per subject are available, a mixed-effects or modified Bland-Altman approach is required to avoid underestimating variability; and critically, clinically acceptable limits must be defined before data collection, not derived from the data itself.
Strengths & limitations
- Directly addresses the question of interchangeability, which correlation coefficients cannot answer.
- The scatter plot immediately reveals proportional bias, outliers, and trends across the measurement range.
- Bias and limits of agreement are expressed in the original measurement units, making clinical interpretation straightforward.
- Universally recognised as the standard method for method comparison in clinical and laboratory research.
- Requires normality of differences; violating this assumption distorts the limits of agreement.
- Does not itself define what agreement is acceptable — clinicians must supply the tolerance threshold independently.
- Limits of agreement widen with small samples, making the analysis imprecise when n is below approximately 30.
- Standard Bland-Altman assumes independent observations; repeated measurements on the same subjects require a modified design.
Frequently asked
Why not just use Pearson correlation to compare two methods?
Pearson correlation measures the strength of a linear relationship, not agreement. Two methods can correlate perfectly (r = 1.0) yet differ by a constant or proportional offset throughout the measurement range. The Bland-Altman analysis detects such systematic biases and shows how large individual differences actually are.
What are 'limits of agreement' and how do I interpret them?
The limits of agreement (LoA) are placed at the mean difference ± 1.96 standard deviations of the differences. They estimate the interval within which 95% of the individual differences between the two methods fall in the population. Whether those limits are acceptable is a clinical or scientific judgement — you need to pre-specify the largest difference that would still allow interchangeability.
What should I do if the plot shows a fan-shaped pattern?
A fan shape (spread of differences increasing with the magnitude of measurement) indicates proportional bias — the two methods disagree more at higher values. In this case, apply a natural logarithm transformation to both sets of measurements before computing differences. The Bland-Altman analysis is then performed on the log-scale, and back-transformed ratio limits of agreement are reported.
How many subjects do I need?
The source registry recommends at least 30 subjects to keep the confidence intervals around the limits of agreement reasonably narrow. With fewer subjects the LoA estimates carry substantial uncertainty, which can lead to either incorrectly accepting or rejecting a method as interchangeable.
Sources
- Bland, J.M. & Altman, D.G. (1986). Statistical Methods for Assessing Agreement Between Two Methods of Clinical Measurement. Lancet, 327(8476), 307–310. DOI: 10.1016/S0140-6736(86)90837-8 ↗
- Giavarina, D. (2015). Understanding Bland Altman Analysis. Biochemia Medica, 25(2), 141–151. DOI: 10.11613/BM.2015.015 ↗
How to cite this page
ScholarGate. (2026, June 1). Bland-Altman Method Comparison Analysis. ScholarGate. https://scholargate.app/en/statistics/bland-altman
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Cohen's KappaStatistics↔ compare
- Equivalence Test (TOST)Statistics↔ compare
- Intraclass Correlation CoefficientStatistics↔ compare
- Paired t-testStatistics↔ compare
- Pearson CorrelationStatistics↔ compare