Tactile Simulation

We will perform a tactile (i.e. "hands on") simulation of the experiment.

Set Up Cards and Simulate:

Mark 11 cards as "success" and 13 as "failure", shuffle them well, and randomly deal out 12 cards to represent the students assigned to group A.  Count the number of "successes" in your randomly generated group A.

Quick Question!  What are you assuming about the effect of the observer's interest when you randomly distribute the "successes" among groups A and B?  

How many of these 12 students randomly assigned to Group A are successes?  Enter the number of successes here.   Is this result as extreme as in the actual sample (i.e. three or less successes)? 

Repeat this a total of five times, recording your results in a table like this:

Repetition #

1

2

3

4

5

"successes" assigned to group A

 

 

 

 

 

as extreme as actual sample?

 

 

 

 

 

Combine your results with the rest of the class, forming a dotplot of the number of successes randomly assigned to group A.  In addition to creating the dotplot by hand, we can also create it using tools such as Webstat or Minitab.  See the Introduction to Minitab sheet for how to use Minitab.  Below are directions for how to use Webstat.

Dotplot creation using Webstat 

Think About Your Results:

Once you have created a dot-plot, please answer the following questions:

  1.  How many repetitions of the experiment were performed by the class as a whole?  How many of those gave a result as extreme as the actual sample (3 or fewer successes in group A)?  What proportion of the repetitions is this?
  2. Remember that your random shuffling and dealing assumed that the observer’s incentive had no effect on the participant’s performance.  Based on these simulated results, does it appear that it is very likely for random assignment of "successes" to produce a result as extreme as the actual sample when the observer has no effect?
  3.  In light of your answer to the previous question, considering that the actual sample is what the researchers found, would you say that the data provide reasonably strong evidence in support of the researchers’ conjecture?  Explain.
  4. If it had turned out that only 1 of the successes had been in group A (and 10 in group B), would that sample result have provided strong evidence in favor of the researchers’ conjecture?  Explain, based on the result of your simulation.