Week 9 Statistical inference and effect sizes Flashcards
What do you do if p>0.05?
Fail to reject the null hypothesis
What do you need to get a significant effect when the real effect is very small?
A very large sample
What are the key sections in a report?
Background
Methods (if you have been given them)
Results
Conclusion
In a report what goes into your background?
What is the study looking at?
What is your hypothesis (null and alternative)
What should the methods include?
Number and information of participants.
What they are doing
Tests used (analysis)
Assumptions needed plus what you used to test for the assumptions
What do you need to include in the results?
The assumption of normality was met
Mean values and SD
Mean difference and CI
T-test values
What do you say in your conclusion?
The main finding of the study
Whether you reject or fail to reject the null hypothesis
A basic summary saying what the finding mean
What is effect size?
An objective and standardised measure of the magnitude of an observed effect.
Basically how significant an effect is
What are the measures of effect size?
Cohen’s d
Glass’ g
Pearson’s correlation coefficient (r)
If the effect size is r=0.10 what is the size if the effect?
Small, 1% of variance
If the effect size is r=0.30 what is the size if the effect?
Medium, 9% of variance
If the effect size is r=0.50 what is the size if the effect?
Large, 25% of variance