Research design explained | Empirical evidence activity

Kelley (2004) recommends using the Monty Hall Dilemma (also called the “Let’s Make a Deal” Dilemma) to show that the empirical approach can be superior to relying exclusively on intuition. If you are unfamiliar with the Monty Hall dilemma (made famous by Marilyn vos Savant in her Parade Magazine column), go to any of the following websites:

  • a short (6 minute) video. You may want to show the first 90 seconds that shows the problem, but wait before showing the rest of the video (the part that explains the solution)
  • an interactive demonstration of the Monty Hall problem
  • (a nice applet)

Note that the password-protected Instructor’s website for Research design explained has a link to the authors’ “plug-and-play” site for this lab: instructions, handouts, and data analysis tools.

Kelley reports that students are convinced that changing from their original guess will not increase their chance of winning the prize. He notes that even when he tells students why changing from their original guess improves their odds of winning, most are still unconvinced. To see that changing really does matter, he has the class do a study to determine whether it is better to stay with one’s original guess or to change.

To approximate his procedure, give each student copies of the two data sheets at the end of this file. Then, have all the students use either a random numbers table or the Randomizer ( to select a winning door for each trial. Next, have students form pairs. One member of the pair is the experimenter (host) and collects data from the other member (the contestant/participant). After the experimenter/host has run both the “stay” and the “switch” conditions, the experimenter becomes the participant and the participant becomes the experimenter.

If you want to speed things up (the demonstration can take between 10 and 40 minutes), you could have students use only 20 trials in each condition. If you wanted to speed things up even further, you could have students do the task on the web at one of the sites below, all of which offer simulations of the Monty Hall Dilemma:

To see whether the difference in strategies was statistically significant, students could plug their data into

Stay Strategy

Trial / Winning door / Contestant’s choice / Door opened by host / Got car or goat
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

Proportion of wins = Cars = ____

Attempts 30

Switch Strategy

Trial / Winning door / Contestant’s first choice / Door opened by host / Contestant’s final choice / Got car or goat?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

Proportion of wins = Cars = ____

Attempts 30