Bland-Altman Analysis

This Tech Tip demonstrates Bland-Altman analysis, a new feature in IBM SPSS Statistics Version 30. Bland-Altman analysis evaluates the agreement between two different instruments or measurement techniques. For example, you can assess how closely a new measurement technique aligns with an existing one. The resulting graph plots the averages of the two measurement techniques against their differences.  

IBM SPSS Statistics offers an integrated interface that allows you to conduct descriptive statistics, regression analyses, advanced statistics, and machine learning models. Its seamless integration with open-source tools and easily accessible help resources ensure that you are well-equipped for any data analysis task, helping you feel confident and capable. 

Bland-Altman analysis also quantifies the agreement between two measurements by constructing limits of agreement. These limits are calculated using the mean and standard deviation of the differences between the two measurements. This allows researchers to visually check the assumptions of normality concerning the differences. In the resulting graph, users can observe the limits of agreement and examine the scatter of points to determine if they are random or if any points fall beyond the agreement limits. A random scatter of points suggests that the measurements provide similar readings. 

To perform a Bland-Altman analysis, users can navigate to the Analyze menu and select Descriptive Statistics. The Bland-Altman Analysis option is located at the bottom of the Descriptive Statistics menu. Once the dialog box is open, select the variables to pair and set your Criteria, Print, and Plot options. Click OK to run the analysis and obtain the results. Users can also ask to print details of the analysis. This provides a table of the data showing the scores on each measure, the mean, the difference and the difference/mean.