Anova - Analysis of variances statistics cheat sheet for Professor Israni\'s Into to PDF

Preview

Summary

Analysis of variances statistics cheat sheet for Professor Israni's Into to Stats course...

Description

-Used for comparing means for more than two groups -Otherwise if you try to use IS tests with more than two groups the experiment wise alpha rate is inflated -ANOVE IS DESIGNED TO CONTROL TYPE 1 ERROR When there is no tx effect the variability between the groups sample means reflects the variability of the scores within the groups When there IS a tx effect then variability between the groups sample means reflects the treatment effect and the variability of the scores within the groups Variance between = variance within (due to ID) + treatment THUS WHEN THERE IS A TX THE VARIABILITY BETWEEN THE SAMPLE MEANS WILL BE GREATER THAN WHEN THERE IS NO TX EFFECT. Assumptions: 1. Additive treatment effect: variances of all the groups are equivalent 2. Test this with F-max if asked 3. Normally distributed outcome variable 4. Independent observations Steps: 1. Hypotheses: (null: mu1 = mu2 = mu3 etc , alt: At least one group is different from the others) 2. Obtain CV using alpha, DFb & DFw, and F distribution table 3. Calculate SSw and SSb 4. Use SS’s to calculate MS’s 5. Calculate F using MS’s 6. Conclusion: reject or fail to reject, say significant, F(DFb, DFw) = x, power, eta squared = SSb/SStotal Eta squared: Shows that the treatment found is of practical significance, eliminates sample size from the equation of calculating the treatment strength -Percentage that represents the proportion of variability that is attributed to the treatment conditions...