Discussion Questions

Discuss the various characteristics of norms. What makes for an appropriate norm group? Assume you were developing an entrance test for an auto mechanics school. What type of norm group would be most appropriate for this test?

Considering a normal distribution of scores, which of the following are farthest from the mean:

For the following decisions, choose which type of measurement, norm-referenced or criterion-referenced, you believe most appropriate:

1. admission to your university honors program
2. state licensing/registration as a medical doctor, barber, and psychologist
3. minimal competency exam for graduation from high school
4. program planning for remedial instruction in reading
5. grade in your psychological measurement course.

Drawing upon the discussion of correlation in chapter 4, choose the most appropriate type correlation coefficient (Pearson, Spearman, Rho, etc.), and diagram what the scatter plots might look like for the following examples:

1. the relationship between chronological age and intelligence
2. height and weight
3. male/female and weight
4. male/female and dead or alive
5. tall/short and pass or fail a physical fitness test
6. thin/heavy and GPA.

Discuss your experience using introductory psychology students as subjects in an experiment. What type of a sample is this? What sources of error may have entered into the study?

Give examples of various tests you have taken in school or employment situations. Classify them as either norm- or criterion-referenced and tell what characteristics of the test lead to your classification decision. Describe how the tests were appropriate or inappropriate for the purposes of testing and the particular setting as well as how the results were used.

What are the various types of decisions that would benefit from a regression analysis? Types of decisions might include admission to your department's graduate program in psychology or your college/university undergraduate program, and employment decisions. How would the regression analysis be conducted? Differentiate regression from multiple regression. Are more predictors better? Why or why not?