Regression Demonstration

Click on the first button to get scores for 50 individuals.
For your convenience,these scores will be put in a table at the bottom of this page and sorted from highest ot lowest.

Get test results for 50 individuals and sort them from high to low (Get the First Measurement)

As you look at these scores, realize that many of these scores will be inaccurate due to random error.
Put another way, scores you get for each individual will often differ from their true score because the score has been pushed up or down by random error.
Specifically, the "true" score for each these 50 participants is 50.

Click on the second button to retest those 50 individuals.
Note that, because, like all tests, this test is not perfectly reliable, an individual's score on the second test will not be the same as that individual's score on the second test.

Retest everyone (Measure everyone again)

Also, note that there will be a pattern in how scores differ from test to retest--for individuals with extreme scores.
Specifically, extremely high scorers do not, on average, score as high on the retest.
Similarly, extremely low scorers do not, on average, score as low on the retest as on the original test.
This pattern will be more obvious once you click on the "Calculate Change" button below.

Calculate Change

Note that the change scores from the first measurement to the second measurement for the high scorers tend to be negative, indicating that participants are scoring lower on the retest.
Note that the change scores from the first measurement to the second measurement for the low scorers tend to be positive, indicating that participants are scoring higher on the retest.
Note that, for participants who score near the average, there isn't much of a pattern in how participants' scores change from test to retest.