Steps in Setting up an Experiment

· Create a hypothesis

· Determine how to set up contrasting treatments for the experimental and control groups.

· Determine how measurements will be made. Are they valid and reliable?

· Select subjects

· Run experiment

Classical experimental design (often called Randomized Controlled Trial or RCT)

One experimental group
One control group
Randomized selection of subjects for both groups
Pretest, Stimulus, Posttest for experimental group (Stimulus is also sometimes called treatment or intervention. )
Pretest, Posttest for control group
The Pretest and Posttest are done on the dependent variable

Experimental group	PreE	Stimulus	PostE
Control group	PreC		PostC

Calculate strength of effect due to stimulus

Diff_exp = Posttest_exp – Pretest_exp

Diff_control = Posttest_control – Pretest_control

Effect = Diff_exp – Diff_control

An Inadequate research design

Experimental group

Stimulus

PostE

There is no pre-test and no control group. We look at the results after the application of the stimulus or treatment.

Quasi-experimental designs

1 group, Pre and Post-test

Experimental group

PreE

Stimulus

PostE

This is a big improvement over the previous design. We can now measure differences.

2 groups, Post-test Experimental group and Post-test control group

Experimental group		Stimulus	PostE
Control group			PostC

We compare two groups that are similar. One has the stimulus and the other does not. For example, the experimental group goes through a learning program, the other does not. Differences in the post test results are assumed to be due to the stimulus.

1 group, multiple pre-tests and multiple post-tests

Experimental group

PreE

Stimulus

PostE

There is no control group. Multiple pre-tests and post-test are carried out in order to determine if changes occurred only after the experimental stimulus was applied.