Week 6: Correlaton Flashcards
What is a correlation?
A correlation measures the strength and direction of a linear relationship between two continuous variables
What are the key assumptions for Pearson correlation?
- Random sample
- Continuous data
- Paired sample data
- Independence of observations
- Variables approximately normally distributed
- Linear relationship
- Absence of outliers
How do you calculate a Pearson correlation coefficient?
<correlate>
or
<pwcorr>
</pwcorr></correlate>
What does a correlation coefficient of -0.4304 indicate?
A moderate negative linear relationship between the variables
When should you use Spearman rank correlation instead of Pearson correlation?
When the data do not meet the assumptions for Pearson correlation, such as when the variables are not normally distributed or the relationship is monotonic but not linear
What is the main difference between Pearson and Spearman correlations?
Pearson measures the strength of a linear relationship, while Spearman assesses monotonic relationships using ranked data
How do you compute Spearman correlation?
<spearman>
</spearman>
What is the relationship between R2 from regression and Pearson correlation?
The R2 value is the square of the Pearson correlation coefficient, showing the proportion of variance explained by the linear model
What does an R2 value of 0.1765 suggest in regression analysis?
About 17.65% of the variation in the dependent variable is explained by the independent variable
Why is it important to visually inspect a scatterplot before analysing correlation?
To confirm the linearity of the relationship and check for outliers or patterns that could violate correlation assumptions
What is the H0 in correlation analysis?
There is no association (correlation) between the two variables
How can correlation results be misinterpreted?
Correlation does not imply causation; a significant correlation only indicates a statistical relationship, not a causal one
