Experimental Design Applets


The applets in this section demonstrate basic issues of experimental design. The Poor Experimental Design ignores randomization rules and allows for increased experimental error. The Improved Experimental Design offers improvement over the first design by adding randomization and reducing experimental error.

Both applets require the input of several participants. The purpose of the applets is to test the reaction times between a participant's dominant and non-dominant hand. The basic outline for the experiment is as follows:

  1. The participant clicks the Start! button.
  2. There is a three second delay before the clock starts so that the particpant can switch to the specified hand.
  3. When the clock starts, one of the squares on the grid will change colors.
  4. The participant is to move the mouse to the colored square and click the square as quickly as possible.
  5. The same participant is then to switch hands and run the experiment again.
  6. The exercise allows twelve participants before being required to reset the applet.
  7. The stored data can be compiled and analyzed using either a paired t-test (Improved Experiment) or a pooled t-test (Poor Experiment) when at least two participants have run the experiment.
The experiments differ in the following manner.
  Poor Experimental Design Improved Experimental Design
Layout 3 x 3 grid Equidistant from the center
Start button Right of the grid Center of squares
Hands Right/Left Dominant/Nondominant
Randomization Right hand always go first Randomizes which hand goes first
Test Pooled t-test Paired t-test
Run Order Haphazard Randomized

View Instructions for this applet

View JSE article for this section


Open in new window   Open in current window
Copyright © 2000 C.M. Anderson-Cook , S.S. Dorai-Raj
<DOWNLOAD> | <Statistical Java>