ANOVA stats flashcard

Card Set Information

Author:
camturnbull
ID:
297017
Filename:
ANOVA stats flashcard
Updated:
2015-02-27 09:38:44
Tags:
Stats
Folders:
Stats
Description:
.
Show Answers:

Home > Flashcards > Print Preview

The flashcards below were created by user camturnbull on FreezingBlue Flashcards. What would you like to do?


  1. Who developed the ANOVA?
    R.A. Fischer
  2. When can an ANOVA be used?
    • In place of the T test for designs with 2 conditions 
    • Designs with more than 2 conditions 
    • Designs with more than one IV
  3. What can we view the ANOVA as?
    • A model fitting procedure 
    • AND a hypothesis testing procedure
  4. What is the SStotal?
    A measure of the variability for scores and residuals around a single mean
  5. What is the SSe?
    A measure of the variability of residuals and scores around the mean of each condition
  6. What is the SSmodel?
    The difference between the SStotal and SSe
  7. What does the SSmodel represent?
    The signal component in the data (the variability between the condition means)
  8. What are SSmodel and SSe also known as?
    • SSbetween (groups)
    • SSwithin (groups)
  9. What are the assumptions for the anova?
    The same as for any other parametric test (normality an homogeneity)
  10. What are the mean squares (MS) in anovas?
    Variances in the data (confusingly)
  11. How do we find the MS?
    Divide the SS with the degrees of freedom
  12. What is the DF total?
    Total number of scores in the experiment minus 1
  13. What is the DF model?
    • The number of experimental conditions -1
    • Experimental conditions = k
  14. What is the DF error?
    The degrees of freedom for the first condition minus the DF for the second one
  15. What is the F value?
    • The test statistic 
    • Named after Fischer (gay)
    • The ratio of the two variances or mean squares 
    • (MS model over MS error)
  16. What are the degrees of freedom for the F ratio?
    The DF for the model and error
  17. When the null is true, what is the expected F value?
    1.0
  18. When there is a high F value, what can be inferred?
    • An experimental effect is present 
    • This could be du to chance though
  19. How is the F-distribution different from the t-distribution?
    • It is positively skewed 
    • Only the upper tail is used when testing hypotheses (even in 2 sided tests)

What would you like to do?

Home > Flashcards > Print Preview