Correlation and Regression

Question 1

Correlation

Answer 1

A measure of the linear relationship between two variables. The correlation value characterizes the size and nature of the relationship.
Cannot infer causation from correlation, just nature and size of the relationship, and if they are related.
Can test the correlation between multiple variables at the same time.

Question 2

Spurious relationship

Answer 2

A linear relationship between variables that is not logical, and usually due to a coincidence or the presence of a third common cause (confound variable). Cannot be determined by a statistical test.

Question 3

r (correlation coefficient)

Answer 3

A means for determining the degree of linear relationship between two variables. NOT a ratio statistic

Question 4

Explained variance (r squared)

Answer 4

An indicator of the proportion of variation in one variable (y) that is shared, or can be predicted or explained when the x values are known. Cannot indicate the direction of the relationship.
Also known as the coefficient of determination.

Question 5

Regression

Answer 5

A calculation the provides a linear equation the predicts values of y for given values of x.
Used to describe the relationship between two variables mathematically and to enable researchers to make predictions.
Doesn’t usually predict the exact value of y, but reasonably close.

Question 6

Regression Equation

Answer 6

The equation that best fits the scatterplot.