Correlation and Regression Demonstration
The purpose of this demonstration is to give you an intuitive concept of correlation
and regression. The control displayed below consists of 25 randomly selected data
points. Each data point is represented by a circle. The line represents the linear
regression equation through the points. The Pearson's product moment
correlation coefficient and linear regression equation are shown below the scatterplot.
The scale of the scatterplot is from 0 to 215 on each axis.
You can click and drag each of the data points. As you move a data point around,
observe what happens to the correlation coefficient, the linear regression equation, and
the regression line.
Here are some things to try:
- Make the correlation coefficient as close to +1 as you can. Mentally note the
orientation of the regression line.
- Make the correlation coefficient as close to +1 as you can, but make the regression
line different than it was before.
- Make the correlation coefficient as close to 0 as you can. How is the regression line
oriented now?
- Make the correlation coefficient as close to -1 as you can.
- What happens to the correlation coefficient and regression line as
you move a single point far away from the rest of the points? Such a
point is called an outlier.
- What can you say about the scatterplot as the correlation coefficient increases from
0 to 1?
- What does the regression line look like when the slope is close to +1? What does it
look like when it is close to -1?
- What does the regression line look like when the intercept is close to 0?