PSY 361

Homework #2, Chapter 4

 

1) The classical test score theory assumes that error is random or systematic?

 

2) Reliability is expressed as the variance of the true scores over what?

 

3) Explain how a split-half reliability coefficient and a KR20 coefficient/coefficient alpha are related?

 

4) If you are working with difference scores would you expect to have higher or lower reliability compared to the reliability of the two scores you are using to calculate the difference scores?

 

5) If you wanted to increase the reliability of a measure would you decrease or increase the number of items on the measure?