Friday, February 6, 2015

Power and effect size: Lecture related material

Files for Power and Effect size lecture

Material backing up a lecture on Power and Effects size to 4th year students.


Lecture slides (1 Feb 2017): 
- PowerPoint file "PSYC402H-ES-power-2017.pptx"

Links

FAQs

Difficulty downloading the above files?
Please note section below.


Power analysis for moderation?
G*Power does not cover moderation directly. However, moderation is a regression, thus estimate N in moderation by using:
  1. The G*Power regression (R2 increase) option.
    • You would have at least three predictors 
      • IV1
      • IV2 (or MO [moderator])
      • IV1xIV2 (the interaction between IV1 and IV2)
  2. Rule of thumb N = 50 + 8k
  3. Rule of thumb N = 104 + k
Where: k = predictors and N = total participants.


Out of 1, 2, and 3 use the one that produces the larger N


Power analysis for mediation
G*Power does not cover mediation directly. Estimate N in mediation using these options:
  1. Using regression (change from zero) in G*Power
  2. N = 50 + 8k
  3. N = 104 + k
Where k = predictors and N = total participants.


Out of 1, 2, and 3 use the one that produces the largest N.


Difficulty downloading the above files
  • Try right-clicking the link and try any "save as" or "save link" options. 
  • If you are using Chrome and are logged in to Google you might like to click the file (Google will open the file as a preview) then you should have options at the top of the window to download the file.

No comments:

Post a Comment