Week 7/8 - Parametric Statistics Flashcards
Random sampling error
Random factors cause a difference between sample and population (fix by increasing sample size)
Non-experimental t-tests
Cross-sectional, observation
Experimental t-tests
Normally distributed, control vs experimental groups
Random variation
Expected difference in values
Systematic variation
Selection biases (sample not random)
Systematic error
Bias causes difference in sample from population (fix by using representative sample)
T-test
2 groups, scale DV
Inferential statistics
Shows if difference is due to IV or chance
Region of rejection
Reject H0 if z-score is in top and bottom 2.5% of distribution
P-value
If t-value is high, p-value is low
Paired t-test
1 sample group does test twice
Repeated measures t-test
See if 2 measures of DV differ significantly (for pre/post tests)
Statistical power increases when…
- low SD
- high group difference
- big sample size
Effect size
Magnitude of effect
Cohen’s D
D>0.2 small while D>0.8 is large