Mr. Marsh has received criticism that his most recent study lacks reliability. He is concerned on how he might might improve future studies

If he uses pilot versions of a measure, might it improve reliability? If not a pilot version, what else could he use? Discuss three other ways that he could improve reliability.

What will be an ideal response?

Yes, a pilot version can improve reliability, but it does take more time. He also could find measures from past research and build on them with appropriate citations.

Clearly conceptualize - clear and unambiguous definitions of concepts - it is better when a researcher defines a concept to eliminate "noise." Increase the level of measurement - higher or more precise measurement is better. Use multiple indicators - two or more indicators of the same concept are better than one. Using more than one measure is a widely accepted principle of good measurement. Use pilot studies and replication - trying pilot versions of a measure can improve its reliability.