# Effect Size & Power

?
As pooled standard deviation gets smaller, the effect is likely to be what?
More significant: smaller deviation from mean scores of each group indicates a more significant effect.
1 of 25
How does effect size counter the limitations of null hypothesis significance testing? give four reasons.
Null hypothesis significance testing 'all or nothing' approach, p >.05 is arbitrary, two groups will always differ to some extent , biased by sample size.
2 of 25
What is the pooled SD?
weighted average of standard deviation of each group.
3 of 25
What kind of effect size is used in most experimental studies?
cohen's D
4 of 25
A big mean difference between groups and small pooled SD infer what?
A large effect size.
5 of 25
Effect Size D's are...
.2 Small, .5 medium, .8 large
6 of 25
Effect Size R's are...
.1 Small, .3 Medium, .5 Large
7 of 25
Effect size r is used predominantly in what kind of studies?
Correlational
8 of 25
Effect size will increase when (2 reasons)
There is less variability in the data, there is a larger difference between the means of the groups.
9 of 25
When sample size increases, effect size what?
does NOT neccessarily increase effect size but it is likely that there will be a more significant effect as the sample size increases.
10 of 25
p value is driven by what?
sample size and effect size.
11 of 25
what is power?
the ability to detect a significant effect when a significant effect exists.
12 of 25
B is what? therefore, power is?
The opposite of finding a significant effect. Power is 1-B
13 of 25
What is a type 1 error?
Falsely accepting H1
14 of 25
What is a type 2 error?
Falsely rejecting H1 (accepting H0)
15 of 25
Power calculations can be deciphered a priori for what purpose?
To help calculate the sample size required for a significant effect.
16 of 25
How can power calculations be used post hoc?
To identify how powerful the study was.
17 of 25
If sample sizes are different, what must be calculated?
harmonic mean
18 of 25
for a priori, how do we know an effect size before a study?
we don't, but we can look at other similar studies, calculate the effect size for these and then average them
19 of 25
The average effect size of previous studies can be used in current studies how?
for efficiency- inferring adequate sample sizes.
20 of 25
What else can increase power?
Similarity, for example using a within subjects design.
21 of 25
A larger sample size?
Does not change the effect size and increases the chance of getting a significant effect.
22 of 25
What combination increases the likelihood of detecting a significant effect between groups?
Large effect size and large sample size.
23 of 25
In experimental designs, which factors increase power?
Large effect size, large sample size, within subjects design.
24 of 25
What is absolutely IMPERATIVE when calculating effect size and power?
checking if the sample sizes are equal or not! If they aren't in effect size, we use the alternative formula. If they aren't in power calculations, we need to determine a harmonic mean to decipher power.
25 of 25

## Other cards in this set

### Card 2

#### Front

How does effect size counter the limitations of null hypothesis significance testing? give four reasons.

#### Back

Null hypothesis significance testing 'all or nothing' approach, p >.05 is arbitrary, two groups will always differ to some extent , biased by sample size.

### Card 3

#### Front

What is the pooled SD?

### Card 4

#### Front

What kind of effect size is used in most experimental studies?

### Card 5

#### Front

A big mean difference between groups and small pooled SD infer what?

## Similar Psychology resources:

See all Psychology resources »See all Statistics resources »