Meta Analysis and hypothesis testing

Question 1

Why do Meta Analysis?

Answer 1

Power: More studies reduces inconsistency in measuring effect size
Theory testing: Does effect fit predictions of a theory
Moderating factors: Which experimental variables influence effect size

Question 2

How to conduct a meta analysis?

Answer 2

1. pick a question
2. select studies
3. calculate summary effect size and heterogeneity
4. check for publication bias
5. check for moderators

Question 3

Large samples may result in significant effect but effect may be small.

Answer 3

Large effect –> don’t need so much samples and it gets significant quicker

Question 4

What does it mean by introducing noise to study?

Answer 4

Introducing confounding variables to study (to control)

Question 5

What prevents us from getting true score?

Answer 5

systematic errors and random errors

Question 6

How can we reduce systematic and randoms errors?

Answer 6

Pilot study, increase sample size, repeating measure and training experimenters

Question 7

What happens when we use multiple t test (more than 2 groups)?

Answer 7

May get false positive, therefore use ANOVA

Question 8

What is the purpose of experimental research?

Answer 8

MaxMinCon
Maximizing experimental variance.
Minimizing error variance
Controlling extraneous variance

Question 9

Other names of the tests (t test, f test, correlation coefficient, multiple regression)

Answer 9

1. one sample t test : standard error of mean
2. independent t test: standard error of difference
3. Analysis of variance F test: mean square error
4. correlation coefficient: standard error of estimate
5. multiple regression: standard error of regression coefficient

Question 10

What is error of measurement?

Answer 10

Difference between true score and observed score.

Question 11

How can we improve quasi experiments?

Answer 11

1. pre and post test
2. removing and reinstating treatments
3. adding control or comparison groups
4. reversing treatments
5. adding replications