I assume you are concerned about heterogeneity of variance. If so, please note that in version 25 of SPSS, a ROBUST sub-command was added to UNIANOVA. It offers five different methods for " the robust covariance matrix estimation". Perhaps one of those methods will suffice. You can read more here:
I assume you are concerned about heterogeneity of variance. If so, please note that in version 25 of SPSS, a ROBUST sub-command was added to UNIANOVA. It offers five different methods for " the robust covariance matrix estimation". Perhaps one of those methods will suffice. You can read more here:
Also, you could conduct an overall test for cell differences in SPSS by having each cell represent a group (e.g., the information in a 2 x 2 design can be re-expressed as a one-way design having four cells or groups), then running the Welch or Brown-Forsythe test in the one-way subprogram. The B-F seems a bit simpler than Welch and could be applied (with a little computation) to the results from spss, even using a factorial design:
Brown, M. B., & Forsythe, A. B. (1974). The small sample behavior of some statistics which test the equality of several means. Technometrics, 16, 129-132. doi:10.2307/1267501
In R, a common way to use a heteroscedasticity-corrected coefficient covariance matrix for anova is to use the Anova function in the car package with white.adjust=TRUE.
This link has a more complete discussion of approaches: https://stats.stackexchange.com/questions/91872/alternatives-to-one-way-anova-for-heteroskedastic-data .