Key Takeaways

• Measurement Error is anything you measure that you didn't mean to measure.  Error makes it harder for you to detect real differences in the thing you are trying to measure. All measures include some error.  Good measures have less error than bad measures. 

• Here are some rules of Thumb that help reduce error in your measures:

(1) Where possible, measure behaviour directly.  If you can't do that, ask about behaviours.

(2) Avoid questions that involve a lot of interpretation from your respondent. 

(3) Use the simplest language you can.

(4) Only ask one thing per question.  

(5) Only ask questions people know the answer to.  Ask about their own experiences, not your research question!

• You can sometimes tell if some of your measures aren't measuring what you thought they would by (1) testing them with a small group ahead of time, (2) measuring the same concept in more than one way and (3) making sure all your data is consistent.