Chapter 12: Inferential Statistics Flashcards
Which of the hypotheses do we want to reject: null or experimental?
Null
Why is it difficult to prove that the experimental hypothesis is true?
We don’t know all of the cases. But- by rejecting (null hypothesis) all of the cases, - the experimental hypothesis is true.
If the null hypothesis is true, what does this mean as far as a difference between groups?
If the null hypothesis is true, there is a difference between groups.
Draw the Type I Error/Type II Error Table.
See lecture slides.
If one makes a type I error, what is he doing?
He is rejecting the null hypothesis when the null hypothesis is true (which means he is stating his intervention is effective, when it is not effective).
Ethically, which type of error is more of a problem: I or II?
Type I Error - because you make a client or entity spend their money or time by saying an intervention is effective.
If one makes a type II error, what is he doing?
Stating that intervention is effective when it is not. Stating intervention is effective when it is not effective.
What does .05 mean, the alpha level?
5/100 times it occurs because of error, meaning 95 times the intervention works because it is truly working.
What type of p-value indicates strong research?
The lesser the p-value, the stronger the research.
If a p value is > .05, what does this mean with regard to the null hypothesis?
You have to accept the null hypothesis. Means the intervention is NOT working.
If the p-value is < .05, what does this mean with regard to the null hypothesis?
You have to reject the null hypothesis. Means the intervention IS working.
What is statistical significance?
Deals with the p-value.
Tells you there is a difference but doesn’t tell you how much of a difference.
What is practical significance?
Tells you how different or how big the difference is - is there a big difference or a small difference.
What is Cohen’s d?
A measure of practical significance. Also known as effect size.
Cohen’s D guidelines
List the steps to follow for testing the differences between means (4).
- How many independent variables are there?
- At the groups related?
- Compute descriptive stats.
- Do a test of statistical significance.
T-Test: I.V.s and Groups
T-Test:
1 I.V., 2 groups
One-Way ANOVA
1 I.V. and 2+ groups
Factorial ANOVA
2+ I.V.s (Factors)
Two-way ANOVA
2+ I.V.s
Multivariate Analyisis of Variance (Manova)
Two or more I.V.s
What does it mean for groups to be related or not related?
Not related: Between-groups design
Related: Within subjects design
What types of testing do you use in a between-groups design (not related)?
- Independent t-test
2. Between measure of ANOVA
What types of testing do you use in a within-groups design (related)?
- Dependent (paired) t-test
2. Repeated measures of ANOVA
What does a t-test serve to do?
Compares the means of two different groups
How big does a sample size need to be for an independent samples t-test?
> 30
What is the main characteristic of a paired t-test?
Repeated measurement on the individual.
ANOVA characteristics
Extension of an independent samples test
Compares the means of groups of independent observations
Can compare 2 or ore groups
One way ANOVA
2+tx levels or classifciations
