Correlations

Question 1

Why is covariance used?

Answer 1

-In experiments, we often manipulate IVs like this to test effects on a DV
-Sometimes, this isn’t possible/ethical to do so and so the solution is covariance

Question 2

How to look at statistical covariance?

Answer 2

-Want to see how related 2 data points/variables are and how they change together
-If variables increase and decrease in tandem then covariance is positive
-If one variable tends to increase when other decreases, the covariance is negative
-EQUATION ON POWERPOINT

Question 3

What’s a problem that can occur?

Answer 3

-It’s difficult to compare scales when they are completely different number e.g. 0-100 and 0-250
-This means that standardisation needs to occur, using correlation

Question 4

What is Pearson’s Correlation Coefficient?

Answer 4

-Parametric (normally distributed data)
-Tells us how much 2 variables are related
-Take the covariance for each group and then dividing this by the standard deviation for each group = creates an r value which is a correlation co-efficient
-Always a value between -1 and +1

Question 5

When to use Spearman’s Rho?

Answer 5

-Non-parametric version of correlation coefficient test
-Used with ordinal or categorical data
-Data is ranked and then correlations are calculated based on ranks
-Always a value between -1 and +1
EQUATION ON POWERPOINT

Question 6

Summarise the difference between covariance and correlation

Answer 6

-Covariance = relationship
-Correlation = analysis

Question 7

Summarise the difference between Pearson’s and Spearman’s

Answer 7

-Pearson’s = parametric
-Spearman’s = non-parametric