Week 6 - F Distributions Flashcards
Sampling Error
Random samples from a population do not have the same mean even though the null hypothesis is true
F-Distribution
The distribution of F when the null hypothesis is true (F should equal 1)
Errors in Hypothesis Testings
Type I error rate
Type II error rate
Type I Error Rate
You think there is an effect, but there isn’t (false positive)
- represented as alpha
Type II Error Rate
You think there isn’t an effect, but there is (false negative)
- represented as beta
IG ANOVA Assumptions
- DV should be measured on a metric scale
- Independence of observations
- Normality of distributions
- Homogeneity of variance
Independence Assumption
States that it is not possible to predict one score in the data from any other score
How can the independence assumption be met in a between group design?
Adequate experimental design
Normality Assumption
States that the samples are drawn from normally distributed populations for each level of the IV
How can you see whether normality assumption is breached
- Inspect frequency histograms
- Compute skewness and kurtosis statistics
What to do with outliers?
- remove them from data
- transform data to remove the influence of outliers
- use a non-parametric test
- bootstapping techniques
- run analysis with and without outliers and see if they affect your results
Homogeneity of Variance Assumption
Ensuring that variance within each treatment conditions/groups is similar
How to deal with breaches of homogeneity assumption
- if there are equal group sizes and breach is minor run ANOVA as is
- lower alpha level to control for impact on type I error
- use alternate statistical test which does not have homogeneity assumption
- transform data to remove heterogeneity
- perform a robust test
- computer intensive methods
Data transformations
Involves performing an identical mathematical operation on all the scores
