Vuchkov I.Solakov E.2024-07-102024-07-102024-07-102024-07-101980-12-010006-344410.1093/biomet/67.2.489SCOPUS_ID:77956890050https://rlib.uctm.edu/handle/123456789/198The robustness to nonnormality of the F test for model adequacy depends essentially on the distribution of repeated trials over the support points of the design.The influence of the regression design matrix upon robustness is more limited. © 1980 Biometrika Trust.enArticle