Skip to main content
Indietro

Correlation and Regression: Study Notes for Introductory Statistics

Guida di studio - Note intelligenti

Appunti personalizzati basati sui tuoi materiali, ampliati con definizioni chiave, esempi e contesto.

Correlation and Regression

Section 9.1: Correlation

Correlation is a fundamental concept in statistics that describes the relationship between two variables. Understanding correlation helps us determine whether changes in one variable are associated with changes in another. This section covers the types of correlation, how to measure correlation, and the distinction between correlation and causation.

  • Correlation: A relationship between two variables, represented by ordered pairs (x, y).

  • Independent Variable (x): Also called the explanatory variable; plotted on the horizontal axis.

  • Dependent Variable (y): Also called the response variable; plotted on the vertical axis.

Scatter Plots

A scatter plot is a graphical tool used to visualize the relationship between two quantitative variables. Each point represents an ordered pair (x, y).

  • Positive Linear Correlation: As x increases, y tends to increase.

  • Negative Linear Correlation: As x increases, y tends to decrease.

  • No Correlation: No apparent relationship between x and y.

  • Nonlinear Correlation: Relationship exists but is not linear.

Example: GDP and Carbon Dioxide Emissions

Scatter plot of GDP vs. CO2 emissions for 10 countries shows a positive linear correlation: as GDP increases, CO2 emissions tend to increase.

Example: Exercise Hours and GPA

Scatter plot of weekly exercise hours vs. GPA shows no linear correlation: exercise hours do not appear related to GPA.

Example: Geyser Eruptions

Scatter plot of eruption duration vs. time until next eruption shows a positive linear correlation: longer eruptions are followed by longer intervals until the next eruption.

Correlation Coefficient

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

  • Sample correlation coefficient: r

  • Population correlation coefficient: \( \rho \)

  • Range: \( -1 \leq r \leq 1 \)

  • Interpretation:

    • r = 1: Perfect positive correlation

    • r = -1: Perfect negative correlation

    • r = 0: No linear correlation

    • Values closer to 1 or -1 indicate stronger correlation

Formula for r:

Example: Calculating r for GDP and CO2 Emissions

Using the formula and data, the calculated r suggests a strong positive linear correlation.

Using Technology to Find Correlation Coefficient

Statistical software and calculators can quickly compute the correlation coefficient.

  • Excel: Use =CORREL(A1:A25,B1:B25) to find r for eruption durations and times.

  • TI-84 Plus: Linear regression output provides r and r2.

  • StatCrunch: Directly reports the correlation between variables.

Excel calculation of correlation coefficientTI-84 Plus linear regression outputStatCrunch correlation output

Testing the Significance of the Correlation Coefficient

After calculating r, it is important to determine if the observed correlation is statistically significant. This can be done using critical values from a table or by performing a hypothesis test.

  • Critical Value Table: Compare r to the critical value for the given sample size and significance level.

  • Significance: If |r| > critical value, the correlation is significant.

Example: Using Table 11 for Old Faithful Data

With n = 25 and a 5% significance level, the critical value is 0.396. Since r = 0.9787 > 0.396, the correlation is significant.

Hypothesis Testing for the Population Correlation Coefficient

A hypothesis test can formally assess whether the sample correlation coefficient provides enough evidence to conclude that the population correlation coefficient is significant.

  • Null Hypothesis (H0): \( \rho = 0 \) (no correlation)

  • Alternative Hypothesis (H1): \( \rho \neq 0 \) (correlation exists)

  • Test Statistic: t-distribution with n - 2 degrees of freedom

Example: t-Test for Correlation Coefficient

For n = 10, calculate t and compare to critical values. If t falls in the rejection region, reject H0 and conclude significant correlation.

Correlation vs. Causation

It is crucial to understand that correlation does not imply causation. Even if two variables are strongly correlated, it does not mean that one causes the other.

  • Direct cause-and-effect: x causes y

  • Reverse cause-and-effect: y causes x

  • Lurking variables: Other variables may influence both x and y

  • Coincidence: The relationship may be due to chance

Summary Table: Types of Correlation

Type

Description

r Value

Perfect Positive

As x increases, y increases in a perfectly linear fashion

1

Strong Positive

As x increases, y increases; points closely follow a line

0.81

Weak Positive

As x increases, y increases; points loosely follow a line

0.45

Perfect Negative

As x increases, y decreases in a perfectly linear fashion

-1

Strong Negative

As x increases, y decreases; points closely follow a line

-0.81 (inferred)

No Correlation

No relationship between x and y

0.04

Additional info: Strong negative correlation r value inferred as -0.81 for completeness.

Pearson Logo

Study Prep