EN
The article proposes a robust procedure for testing equality of covariance matrices, which may be viewed as a supplement to the survey comparability testing framework introduced by Jöreskog in 1971. The M test, though overly sensitive to departures from multivariate normality and outliers, was used to test his hypothesis. The simulation study showed that the proposed method approximately controls the type I error rate under departures from normality and in the presence of outliers. The proposed method was also used to test the comparability of cross-national Polish-American survey results.