STEPS FOR CONDUCTING AN ANOVA FOR A COMPLETELY RANDOMIZED DESIGN

1. Check for randomization and independent samples

 

2. Check assumptions of normality and equal variance

3. Create an ANOVA table and carry out F-test

4. If you reject the hypothesis that the means are same:

  • Conduct a multiple comparison procedure.
  • Construct confidence intervals for mean(s)

5. If you fail to reject the hypothesis that the means are same then:

  • The treatment means are equal
  • The treatment means are not equal. Increase the sample size, consider other factor