W8 - Effect Size Flashcards
What do inferential statistics allow us to do?
Make conclusions about the pop based on sample statistics
What can be HIGHLY influenced by the sample size
The p-value
Therefore whether the data shows it as being statistically different or not/
Explain sample size + statistical significance
Very large sample size = Possible to get a significant effect when the real effect is very small
What does a larger sample size result in for the representativeness of the sample of the population
Larger sample size = More representative the sample of the pop
Larger sample size = More representative the sample of the pop
What follows?
Lower risk of sampling bias = ⬆️ reliability from a small effect
Why is an effect size created?
To determine whether something is PRACTICALLY significant + meaningful in the real world.
What can effect size facilitate with?
Comparison of effect across studies
What is an estimate of effect size?
Correlation coefficient (r)
What can Cohen’s D be used for?
To look at whether the difference between the 2 means of 2 groups is meaningful or not.
When would you use Cohen’s D?
When focus is on difference between 2 means
Cohen’s D equation
D = (M1-M2) / SD
What can the SD in the Cohen’s D equation be?
Pooled = SD in total sample
OR
SD of control group / condition
In what ways can we interpret the Cohen’s D size?
In terms of small, medium or large effect
Small effect size for the r value and the D value
r = 0.1
D = 0.2
Medium effect size for the r value and the D value
r = 0.3
D = 0.5
Large effect size for the r value and the D value
r = 0.5
D = 0.8
What is another way to calculate an effect size?
By converting a t-statistic into a Pearsons correlation coefficient (r)
Equation to calculate effect size by converting a t-statistic to Pearsons correlation coefficient
r = square root of ( t^2 / t^2 + df)
Provide reasons why its important to report effect size (3 marks)
Indicates whether the effect matters in the real world
Objective + standardised measure of magnitude of effect
Facilitates comparison of effect across studies
When looking at SPSS data, what table would you look at to describe the relationship between the 2 variables?
What data would you provide in your answer?
Correlations table showing:
Pearsons correlation
Sig. (2 tailed)
N
i.e answer = Moderate to strong significant positive relationship, r = 0.480 and p<0.05
How would you answer this question?
What is the regression equation for these variables?
Y= a + bX
Then work it out
(Look at the coefficients table)
