Dear researchers,
I am planning to conduct an experimental study on writing. There are 3 groups, 30 participants in each group.
The dependent variables are: complexity (measured based on 2 indices), accuracy (measured based on 2 indices), and fluency (measured based on 3 indices). In total, there are 3 main dependent variables or 7 dependent variables (7 indices).
There are 3 independent variables and 2 covariates.
As there are many assumptions that need to be met in running MANCOVAs (i.e, normal data distribution), what is the alternative to this test, especially for non-normal distribution of data?
Thank you very much for your help.
Dinesh Kumar
Nonparametric tests: Kruskal-Wallis test for independent variables and Friedman test for dependent variables can be applied as nonparametric equivalents of ANOVA. Generalized linear models (GLMs): These models can handle non-normal distributions and different link functions for the relationship between predictors and dependent variables. Bootstrapping: This resampling method can provide robust estimates of the effects and confidence intervals without the need for normal distribution assumptions. Transformations: Try transforming the data to meet the assumptions of normality and homoscedasticity before applying MANCOVA.
If MANCOVA assumptions are not met, consider these alternatives for non-normal distribution of data:
1 votes
1 thanks
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators
More Long Quoc Nguyen's questions See All
Similar questions and discussions