Correlation Coefficient: Understanding and Calculating Relationships Between Variables

Correlation Coefficient

Introduction to Correlation Coefficient

The correlation coefficient is a statistical measure that quantifies the direction and strength of the linear relationship between two quantitative variables. It is commonly denoted by r.

• Definition: The linear correlation coefficient (r) measures the direction and strength of correlation between two variables.

• Direction: The sign of r indicates the direction of the relationship:

  • Positive r: As one variable increases, the other tends to increase.

  • Negative r: As one variable increases, the other tends to decrease.

• Strength: The magnitude of r (how close it is to -1 or 1) indicates the strength of the relationship:

  • Strong correlation: Points are tightly clustered around a straight line.

  • Weak correlation: Points are more scattered.

  • No correlation: Points show no apparent linear pattern.

Range of r: $-1 \leq r \leq 1$

Interpretation scale:

Value of r	Interpretation
r ≈ 1	Strong positive linear correlation
r ≈ -1	Strong negative linear correlation
r ≈ 0	No linear correlation
0.5 < |r| < 0.8	Moderate correlation
|r| < 0.5	Weak correlation

Visualizing Correlation

Scatterplots are used to visualize the relationship between two variables. The clustering and direction of points indicate the type and strength of correlation.

• Example: Three scatterplots with different values of r:

  • r = -0.94: Strong negative correlation (points closely follow a downward-sloping line).

  • r = 0.13: Very weak positive correlation (points are scattered, little linear pattern).

  • r = 0.44: Moderate positive correlation (points loosely follow an upward-sloping line).

Important Note: The slope of the "best-fit line" does not affect the value of r; r only measures the strength and direction of the linear relationship.

Interpreting Correlation Coefficient in Context

Given a value of r, you can match it to a scatterplot that best represents the relationship:

• Example: If r = -0.82, the scatterplot with a strong downward trend (negative slope) best matches the data.

• Application: In marketing, a positive r (e.g., r = 0.89) between advertising budget and sales revenue suggests that higher advertising budgets are associated with higher sales, though not perfectly.

Calculating the Correlation Coefficient

The formula for the sample correlation coefficient is:

$r = \frac{\sum (x_i - \bar{x})(y_i - \bar{y})}{\sqrt{\sum (x_i - \bar{x})^2 \sum (y_i - \bar{y})^2}}$

• xi and yi are the individual sample values.

• \bar{x} and \bar{y} are the sample means of x and y.

Alternatively, calculators and statistical software can compute r directly from data sets.

Using a Calculator to Find the Correlation Coefficient

Most graphing calculators (e.g., TI-84) have built-in functions to compute r:

• Enter data into lists (e.g., L1 and L2).

• Use the LinReg or CALC menu to calculate r.

• Example: For test scores vs. time studying, entering the data and running LinReg yields r = 0.34 (weak positive correlation).

Tabular Data Example

Consider a scientist measuring the speed of sound at different altitudes:

Altitude (thousands of feet)	Speed of Sound (ft/sec)
0	1120.4
1	1116.4
2	1112.5
3	1108.6
4	1104.7
5	1100.8
6	1096.9
7	1093.0
8	1089.1
9	1085.2

To determine if there is a correlation between altitude and speed of sound, enter the data into a calculator or software and compute r.

Summary Table: Correlation Coefficient Properties

Property	Description
Range	-1 to 1
Direction	Sign of r (+ or -)
Strength	Magnitude of r
Linear Relationship	r measures only linear association
Unitless	r has no units

Key Points

• Correlation coefficient (r) quantifies the direction and strength of a linear relationship between two variables.

• Values of r close to 1 or -1 indicate strong linear relationships; values near 0 indicate weak or no linear relationship.

• Scatterplots are useful for visualizing correlation.

• Calculators and software can quickly compute r from data sets.

• Correlation does not imply causation; other factors may influence the relationship.