Control Test
Control Test - textbook - RD
How much data to collect:
Power analysis:
- r (expected correlation)
- alpha (Type I error ratem typically 0.05)
- beta(Type II error rate, typically 0.20)
We do not need full power, but comment on lack of power In limitations.
Threats to External Validity
Threats to external validity are expeected to be talked about
- History, maturation, mortality, testing-participants get familiar with tests
More on the textbook 170-171
Reminder: check our data, more on ppt
Grading notes
Categorial
Nominal like gender, number are no meaningful order
Quantative: Number is meaningful statistically
Basic Tests:
Get tools online to calculate for us
- T-test - compare means two groups (try to break the hypothesis)
- Assumption, Hypothesis: H0: u1=u2
- ANOVA - not tell which is different.
- Chi-square