It depends why you want an alternative. If it is because you don't like the big influence of big residuals, then you could use robust alternatives. If you just don't like the format of the output, then you could run your ANOVAs are regressions. If your dependent variables are measured with error, then you can use errors-in-variables approaches. etc.

IRT.

Hope this helps,

Matt

Unless you have normality issues or heterogeneity in your data then you need to check it very, the reason you bored of ANOVA technique may be becasue you run it with error consistently. However, If your data failed to satisfy parametric condition then the non-parametric alternative to ANOVA is Kruskal–Wallis test by ranks, Kruskal–Wallis H test, or Friedman's Test are fairly standard non-parametric statistic, even though the usage of these two statistics defends on your objective

If you have many groups, try a mixed effects (hierarchical/multilevel) model instead - estimate the main effect of the group as a random effect. This makes a much more sparse model and gives you an estimate how much variance is really in the grouping factor.

Also, in structural equation modeling, setting up a multigroup model, it can be handled elegantly if variances in the dependent variable vary between the groups, and also many dependent variables can be modeled at the same time.

It might be worth looking at an older book, Birkes and Dodge's ALternative Methods of Regression. It overs lots of alternatives to regression (and of course ANOVA is just a type of regression).

Thank you so much Ado Abdu Bichi

Thank you so much Daniel Wright