STATS2- correlation & pearson's r

what is a correlation

relationship between 2+ non-independent variables

what does correlation measure

covariance: extent to which change in 1 variable is associated with predictable changes in another variable

4 features of correlational research design

• measures if 2 variables are related

• can’t infer causality

• produces positive or negative correlations

• eg: linear correlations

what does it mean when you have high co-variance

if scores for 1 variable change then the scores for the other variable changes in a predictable way

total covariance

sample covariance

pearson’s r

how to work out standard deviation from variance

sd= square root of variance

values for perfect positive/negative & no correlation

• perfect positive correlation= 1

• perfect negative correlation = -1

• no correlation = 0

values for strong positive/negative correlation

0.9-0.7 or -0.9 - -0.7

values for moderate positive/negative correlation

0.6-0.4 or -0.6 - -0.4

values for weak positive/negative correlation

0.3-0.1 or -0.3 - -0.1

4 steps for pearson’s r test

1. hypothesis: x correlation between variables

2. collect data & calculate pearson’s r

3. calculate p value

4. reject or fail to reject null hypothesis

   • pearson’s > critical value = reject null

what to do with negative r values when looking at critical values table

ignore -