Friday, 26 May 2006

The next time you stereotype, think external validity.

External validity is a measure of how readily youc an generalise a statement or research. High external validity implies you can make a very broad statement.

Technically we would say you have high external validity if you are basing your statement on a large number of observations. For example, if the caravan you buy from a pikey doesn't make it out the gate, and its the only thing you have ever bought from a pikey, and you claim that all Pikeys are thieving dishonest con artists, your statement has very low external validity.

On the other hand a consumer report which collected a sample of 1,000 customer satisfaction reports from people who bought caravans from pikeys would have a much higher external validity.

So the next time you stereotype or generalise (as we are all [especially me] prone to do), think of exactly how many observations you have and consider your external validty.

3 comments:

Rhian said...

Goiz! You reminded me of Dad when I read this post.

Anonymous said...

does that work if you transpose bus for caravan, pikey for fuller?

D. said...

Yes, yes it does.