Revision Concepts Flashcards
What is a Type 1 Error?
- a level
- When a result is significant we could either be correct or making a type 1 error
What is a Type 2 Error?
- b level
- When a result is not significant we could either be correct or making a type 2 error
What is Power?
1 - beta
probability of finding a significant effect when one exists in the population
What is t?
observed difference / expected difference
If t(15) = 2.3, p = .000, what does the value of 15 represent?
df (degrees of freedom)
In a one-sample t-test of effect sizes, what does the point of estimate stand for?
Cohen’s d
What is a small effect size?
(Cohens d, Persons r, Persons r2, eta-sq n2)
Cohens d = 0.2
Persons r = .1
Persons r2 = .01
Eta-sq n2 = .01
What is a medium effect size?
(Cohens d, Persons r, Persons r2, eta-sq n2)
Cohens d = 0.5
Persons r = .3
Persons r2 = .09
Eta-sq n2 = .06
What is a large effect size?
(Cohens d, Persons r, Persons r2, eta-sq n2)
Cohens d = 0.8
Persons r = .5
Persons r2 = .25
Eta-sq n2 = .14
What sample size will normally ensure that the sampling distribution of the means is normal?
30
- should also look symmetrical
What is the equation to test the assumption of normality?
Take each statistic for skew and kurtosis and divide it by its standard error.
- If this value is between ±1.96 then it can be considered normal.
- Aim for the number to be insignificant
What follow up test is used to test a prior prediction?
planned comparisons (contrasts)
If the number of comparisons is fewer than the number of levels in a planned comparison, what alpha level can we use?
a = 0.5
If the number of comparisions is more than the number of levels in a planned comparison, what alpha level can we use?
Bonferroni adjustment a = 0.25
What is the best method to use when running planned comparisons when you’re only interested in comparing 2 levels at one time?
Estimated Marginal means