Levene Test significant, Box´s M-Test not
Posted: Mon Sep 01, 2025 8:29 pm
Hi everyone,
i´m working on a mixed anova and did get some significant results for the LEVENE tests, regarding variance homogeneity. on the other hand, the covariance matrixes appear to be fine (Box´s M-Tests weren't significant). in my head it´d be more logical for them both to be significant / non-significant. Either way, can I take this as a good sign that my data has enough stability to be tested? Or would you still refrain from interpreting ANOVA results with the violated assumption of homogeneous variance?
Thank you for your expertise & opinion
Anni
i´m working on a mixed anova and did get some significant results for the LEVENE tests, regarding variance homogeneity. on the other hand, the covariance matrixes appear to be fine (Box´s M-Tests weren't significant). in my head it´d be more logical for them both to be significant / non-significant. Either way, can I take this as a good sign that my data has enough stability to be tested? Or would you still refrain from interpreting ANOVA results with the violated assumption of homogeneous variance?
Thank you for your expertise & opinion
Anni