Internal vs. External Validity

What is internal validity?

Extent to which the experiment shows the changes in the behavior are due to the independent variable and not the result of uncontrolled or confounding variables 

Examples of confounding variables:  

-Starting medication

-Change in home life (divorce, death, new baby, move, etc)

-Change in school/therapy life (new therapists, clinicians, new building, new school, etc.)

What is external validity?

The degree to which a study's findings have generality to other subjects, settings, and/or behaviors

-Will this study prove effective if a different population of participants is used?

-Will this study be effective if used with a different type of behaviors?

If the study was done in a clinic, will it be effective if conducted in a school classroom setting?  What about in a home environment?

Resources:  Cooper J.O, Heron T.E, Heward W.L. Applied behavior analysis (2nd ed.) Upper Saddle River, NJ: Pearson; 2007.

Testing a Definition

So now that you have learned the characteristics of a good definition, how do you test to see if your definition is actually good?

Cooper, Heron, and Heward describe a few questions and answers to determine if you have a good definition.

1.  Can you count the number of occurrences?

-You should be able to answer "Yes"

2.  Will a stranger know what to look for based on the definition alone?

-You should be able to answer "Yes"

3.  Can you break the target behavior down to smaller,  more specific components?

-You should answer "No"

-The target behavior should already be broken down into the most specific components as possible

Resources:  Cooper J.O, Heron T.E, Heward W.L. Applied behavior analysis (2nd ed.) Upper Saddle River, NJ: Pearson; 2007.

Reversal Design

Fourth Edition Task List

Section 1

B-04

Reversal Design

(Also known as ABA design)

How does it demonstrate experimental control?

Demonstrates experimental control by repeatedly introducing and withdrawing the experimental control.

Also, confounding variables are unlikely to repeatedly occur at the same times as the experimental variable

The following visual represents a typical ABA or Reversal Design. 

A= Baseline

B=Intervention

A= Baseline

So, the reversal design sounds fantastic.  What is the catch?

You cannot use the reversal design with variables that are irreversible or it is unethical to return to baseline.

Examples of when reversal designs would NOT work...

If you teach a child how to read words, it is irreversible, so a reversal design would not work.

If during baseline, the student knew how to read 3 words.  Then you implement the intervention and now the student can read 12 words.  You cannot reverse back to baseline because even if you remove the intervention the student has been taught to read 12 words and they will continue to read 12 words. It is irreversible.

If a child is engaging in head-banging during baseline 30 times per minute.  A doctor has determined that they are very close to going blind in one eye due to the intensity of the head-bangs. Therefore, if during the intervention you reduce the head-banging to 1x per hour, it would be unethical to remove the intervention and allow the child to head-bang up to 30x per minute.  In this situation, a reversal design would not work. 

Resources:  Cooper J.O, Heron T.E, Heward W.L. Applied behavior analysis (2nd ed.) Upper Saddle River, NJ: Pearson; 2007.