29 Replies to “Psychology 3256 (Winter 2009) Review Notes”

  1. Hey, just wondering what everyone else got for the yates order review question with the three variables? the last line of mine was sv=s(BAC) and df= (n-1)bac=40 (I only used 2 for a because there was 2 levels of a within each level of b) is this what everyone else got?

  2. And the answer to the second is:
    B 1
    S(B) 8
    A 1
    AB 1
    AS(B) 8

    right?

    Are the answers for the other review questions posted anywhere?.. I’m not that familiar with how to use this site.

  3. For the regression model question I chose Y=X1X2 because it accounted for nearly 50% of the variance and it had a C(P) value of 3.1 which is so close to 3. X3 correlates with X1 and X2 because their R2 values don’t add up. What does everyone else think??

  4. @Melanie yes, that would be the one I would pick. The three variable model is not bad, but frankly I mean, no c orrelation and a nice low CP, I would go with the one you picked.

  5. Hi

    I’m working on filling in the table, and I just have a quick question. I know that the rows, columns and IV are equal; are they also equal to the residual? If not how to do fin the value?
    Thanks!

  6. so…for the latin square example of the df table I got that p=3, N=9, and n=1. For the final F scores i got Rows F=2, Columns F=5 is this what other people got also?

  7. Melanie, I got p=3 too, but for the F scores, if I’m not mistaken, the only one you’d be interested in would be the IV (which was already given to us), you wouldn’t test the other two. Also I got N-1 = 8 is that what you meant?

  8. I got the same as Melanie for the F scores (rows = 2 and columns = 5), and I got the same as Amy for N-1 = 8….so Dave what is correct??

  9. Hi Dave, this is what I got for the second hierarchical design you posted… is that right?

    A = 2
    B(A) = 3
    S(AB) = 54
    C = 1
    CA = 2
    CB(A) = 3
    CS(AB) = 54
    Total = 119

  10. The key thing about residual plots is that they should be randomly distributed around the x axis, there should be no pattern at all.

  11. Dave,

    I can’t remember this Latin Squares shit and I can’t help Dana … how did you get 2 df for rows, columns, and IV? It’s friggen bugging me that I can’t remember this!!

    Thanks

    We know it’s p-1 blah blah … and (p-1)(p-2) for error … a little help for my sanity please

  12. Hey Dave…just wondering about not having F scores for the rows and colums in a latin square. Is that because they are nuisance variables and we choose the levels, is this why there is no error term to test them?

  13. Hey Dave…just wondering about not having F scores for the rows and colums in a latin square. Is that because they are nuisance variables and we choose the levels, is this why there is no error term to test them?

    The residual plots that you posted, it looks like only the first one shows an unbiased variable. The other plots all look like you can predict the residual based on X increasing or decreasing? Is that what to look for?

  14. @Melanie yes, there is a problem with those other three and you are dead on. As far as testing the rows and columns of Latin square designs, they are nuisance variables that are considered random factors, so, they have no error terms with the requisite expected value to do a test, and well, who cares. Kind of like subjects in any other design, or blocks in a randomized block.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.