You need a Multivariate Analysis (MV) test to evaluate multiple comparisons. With five catagorical varying treatments/factors/levels to test, the result would be 15 tables using a t-test analysis. By using an MV test, there would only be one table to review.
Plus, every time that you apply a test, there is a chance that it is wrong. With 20 tests there are good odds that the statistical conclusions are a mistake, in at least one (a one in twenty chance).
ANOVA: Analysis of Variance looks for any statistical difference across groups. How much do the results deviate form the mean? (Note: variance is the square of the standard deviation). The "R" foundation for statistical computing (freeware) can analyze this to run the data, but you would drop the results into Excel to better visualize them.
The Tukey is always the post-hoc test after the answer is found. What this means is the you need to group similar results to more easily see where differences are.
No comments:
Post a Comment