WEEK15: Correlations Flashcards
What is a limitation of correlational analysis?
Correlational analysis can not infer casual relationships
What is the correlation test called?
Equivalent non parametric test?
Assumptions for parametric test/ non parametric (data, distribution)
Pearsons r
Spearmans rho
Parametric- interval or ratio data, normal distribution
Non-parametric- ordinal data, does not assume a normal distribution
What are the two values of a correlation analysis?
What do they mean?
+1 = perfect positive relationship
-1 = perfect negative relationship
What is a:
positive relationship?
negative relationship?
Positive relationship = high scores of X go with high scores of Y
Negative relationship = high scores in X go with low scores in Y
Boundaries for positive and negative correlation coefficient
Cohen
none = 0
Weak = 0.1 - 0.2
Moderate = 0.3 - 0.4
Strong - 0.5 - 0.9
Perfect = 1
What is the correlation coefficient a direct measure of?
Correlation coefficient is the direct measure of effect size
What do two variables share if two variables are correlated?
Calculation? units? What is this known as? How can this divide be represented?
On SpSS where is the correlation coefficient shown? What is the range for this value?
When two variables are correlated, it is assumed that they ‘share’ variance
To calculate the amount of variance they share: square the correlation coefficient (r) and multiply by 100 (Share is expressed in %) This divide can be represented using a Venn diagram
Example: r = 0.60, so 0.6 x 0.6 = 36 x 100 = 36% (coefficient of determination)
On SPSS the correlation coefficient is on the Pearson corre row ( this value should be in the range of -1 and +1
What variable is used to calculate the correlation coefficient?
not?
The calculation for df and N?
The sample size is used to calculate correlation coeffienct not the degrees of freedom
To calculate sample size: df = N - 2 ( so plus 2 on df for participants)
How is a pearsons r reported?
r(x)=xxx, p = significance
What kind of relationship does a correlational analysis show?
What scale do the variables need to be measured on for a correlational anaylsis?
For a correlational analysis variables need to be measured on a continuous scale.
Linear relationship
What is a linear regression line? (the connection between?)
Also can be described as?
A linear regression line shows the connection between scattered data points eg line of best fit
How to calculate the variance not accounted for in the correlation coefficient (unique variation)
100 - correlation coefficient squared
When is a correlation coefficient significant?
What is the p-value?
For the result to be significant the P value (probability: one or two-tailed) has to be below 0.05
