1. Go see this colour personality test.
2. Take it twice; once for real, and again just clicking random squares.
3. You'll get two personality descriptions. Without reading them (do your best), copy and paste both descriptions into two separate text files, being sure to label which one is the real you, and which one is bogus.
4. Also copy both descriptions together into a third file.
5. Now in this third file, mix the two descriptions together so you don't know which statements come from your real test and which from the bogus one.
5. Mixed them up enough? Okay, now one by one, decide if each statement was a 'hit' or a 'miss'. If a line contains more than one statement, break them up.
For example, this line is actually three statements.
May be physically unwell, in need of gentle handling and considerate treatment.
6. Now that you've judged all the statements, it's time to reveal! Open up the two other files, and find out how many hits and misses there were in your real and bogus descriptions.
For me, the bogus one had 8 hits, 10 misses. The real one had 11 hits and 5 misses. Ooo, looks good from here, right? but it could be a fluke.
7. We need one more step. The Chi-square test is a test for statistical flukiness. Go to the Web Chi-Square Calculator, make a 2x2 table, plug your numbers in (so, for instance, I had 8 and 10 in the first row, and 11 and 5 in the second row), and hit calculate.
Here's what I got:
Degrees of freedom: 1
Chi-square = 2.02972709551657
For significance at the .05 level, chi-square should be greater than or equal to 3.84.
The distribution is not significant.
p is less than or equal to 0.20.
When it says "The distribution is not significant," it means there's no significant difference between the real and bogus test.
8. Post your results in comments. What did you get?
9. Read about the Forer Effect.