Calculate the covariances. Use your statistical software to calculate a set of three covariance measures: the covariance between the items in the first trait, the covariance between the items in the second trait and the covariance between the items of traits 1 and 2. Most statistical software packages have a command that automatically calculates the covariance between two variables. For example, in the R software environment, the command is “cov(x,y).”
Calculate the sample standard deviations. Do this for both trait 1 and trait 2. Statistical software packages can do this for you. In R, the command is “sd(x).”
Find the correlations between the traits. Do this for all combinations of trait1/trait1, trait2/trait2 and trait1/trait2. Calculate these by using the formula cov(x,y)/[sd(x)*sd(y)]. Call the results Rxx, Ryy and Rxy for the combinations trait1/trait1, trait2/trait2 and trait1/trait2 respectively.
Compute the number that represents the test result. Multiply Rxx and Ryy. Take the square root of the result. Divide Rxy by the result. This final number will be between 0 and 1.
Interpret the finding and make a conclusion for the test. If the number is close to zero, this implies the common method bias is non-existent or low; you should conclude that common method bias does not affect your findings. If the number is close to 1, this implies that the common method bias is extremely strong; conclude that the use of a common method to test the two traits makes the results unreliable. If the number is high but not close to 0, such as in the range 0.5 to 0.8, conclude that common method bias is present but does not likely invalidate the study’s results.