The sample is balance and unbalanced that mean if i have three group of respondents. which is Group A (33 respondents), Group B (57 respondents) and Group C (130 respondents). So how can i make a differences between 3 group of respondents which is unbalanced. Can i run the data by using Anova (para) or Kruskall Wallis (non-para)? Or its depends on result of normality test?