STA305 Pre-Midterm

Card Set Information

Author:
Mageros
ID:
296010
Filename:
STA305 Pre-Midterm
Updated:
2015-02-25 20:01:18
Tags:
UTM STA305
Folders:

Description:
STA305
Show Answers:

Home > Flashcards > Print Preview

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


  1. What is [Y]?
  2. What is [A]?
  3. What is [T]?
  4. What is SST?
    • [Y] - [T]
    • Total variation of individuals from the grand mean.
  5. What is SSE?
    • [Y] - [A]
    • The variation of individuals from group means.
  6. What is SSA?
    • [A] - [T]
    • The variation of group means from grand mean.
  7. What are the 5 Assumptions for doing a t-test?
    • 1. Both groups are independent from each other.
    • 2. Observations within groups are independent.
    • 3. Groups follow a roughly Normal distribution. If not, try non-parametric method like Mann-Whitney test.
    • 4. (Population) variance of each group equal. If not, use Satterthwaite's approximation for conservative results.
    • 5. Equal sample sizes preferred. If not, less powerful.
  8. What is the power of a test?
    • It is the probability of (correctly) rejecting the null, when the alternative is true. 
    • It is the complement of the probability of making a Type II Error.
  9. What is  equal to?
  10. In contrasts, what is t equal to?
  11. With equal sample sizes, the standard error has what form?
  12. With unequal sample sizes, the standard error has what form?
  13. What is a confidence interval for ?
  14. How do you apply Bonferroni Correction? How do you find the F-critical value?
    • Multiply the p-value by the number of tests done. If exceeds 1, leave it as 1.
    • To find the F crit, look in the F-table using df(k-1, kn-k) and divide your alpha by the number of tests you are doing.
  15. How do you apply the Sidak-Bonferroni Correction? How do you find the t crit?
    • The formula is  and is more powerful than Bonferroni.
    • Use the Sidak-Bonferroni t-distribution, and locate your alpha level. The number or tests are on the columns, and df is equal to kn-k.
  16. How do you apply Scheffe's method? When is it most useful?
    • To use Scheffe's method, 
    •  where df = (k-1, kn-k) respectively. 
    • It is most useful when you want to test a lot of contrasts ahead of time without specifying.
  17. How do you apply the Dunnett method? When is this useful?
    • where df = (k, kn-k), respectively.
    • It is useful when comparing means of treatment groups to the control. The groups in which means differ by more than this number are said to be significantly different.
  18. How do you apply Tukey's HSD method? When is it most appropriate?
    • Where  is found in the Studentized Range Statistic, and a is the number of means (or groups, k). Now, use (k, kn-k).
  19. Which method is best for testing all pairwise comparisons?
    Tukey, but remember, the sample sizes should be nearly equal.
  20. Which method is best when you want to want to  do tests that go beyond pairwise comparisons and you can specify them all in advance?
    Bonferroni or Sidak-Bonferroni.
  21. Which method is best if you want lots of special contrasts but cannot specify them all in advance?
    Scheffe.
  22. Briefly define each type of statement.
    • Analytic propositions
    • Logical Positivist Statements
    • Dogmatic Claims
    • Aesthetic/Ethical Claims
  23. What does a scientific claim need to be?
    Falsifiable, proven true or false by experiment. Hypothesis is formed and either accepted or rejected.
  24. Briefly define the following types of variables:
    • Categorical (each level is a factor in this category)
    • Ordinal
    • Interval
    • Ratio
  25. What are nuisance variables?
    Anything that influences the response variable other than the treatment condition
  26. What can we do about nuisance variables?
    • Control.
    • If we can control a nuisance variable and keep it constant throughout all treatment levels, then we should. It is no longer a problem variable anymore.
  27. When is blocking used?
    For nuisance variables that cannot be controlled but can still be observed, we can use Blocking to make sure each group has an equal amount of each
  28. When is randomization used?
    For nuisance variables that cannot be controlled orobserved, we depend on Randomization to spread these variables out evenly
  29. When can you draw causal links?
    Only properly controlled, double-blind, randomized experiments can determine causation
  30. What is the individual ?
  31. What is the pooled ?
  32. What is  equal to?
  33. What is  equal to?

What would you like to do?

Home > Flashcards > Print Preview