I. Measurement Issues

A. Behavior can be described in a variety of ways…

B. Constructs and the problem of measurement

1. Hypotheses and constructs (ideas).

2. operational definitions and conceptual definitions

Discussion Questions:

1) Name some constructs that cannot be seen but that are important to our everyday lives. How would you go about operationally defining these in an experiment? How does an operational definition limit our understanding of a construct?

2) Consider the following hypotheses. How would you define each of the constructs specified in the hypotheses? Design (briefly) a study to test the hypotheses.

a) Television viewing causes increased aggression with peers

b) Highly active infants are more likely than average or low active infants to be uninhibited with their peers when they enter school.

c) With age, children are better able to use facial cues to identify emotions in others.

C. Scales of measurement

1. Nominal (or categorical)

2. Ordinal

3. Interval

4. Ratio

D. Measurement Error

Sources of measurement error

II. Evaluating our measures

A. Reliability

1. Test-retest

2. Split-half

3. Inter-observer agreement

How to increase?

B. Validity

1. Face validity

2. Construct validity

3. Criterion validity/Predictive validity

III. Discussion

A. For each of the following measures, determine whether it is measured on a nominal, ordinal, interval, or ratio scale: (Why?)

Body temperature
Sexual orientation
The number of times a baby smiles in 5 minutes
The order in which 150 runners finish a race

B. DISCUSSION: Measuring “social interaction”

IV. Making decisions about data collection

A. WHERE will we conduct our research?

1. Control versus “naturalness”

2. What is my question?

3. How do I wish to generalize my results?

B. HOW will we collect our data?

what techniques best fit my question?

V. Typically used methods

A. Observation

1. participant-observer or objective observer?

2. disguised vs. undisguised?

3. observational methods

· narrative record

· time-sampling/checklists

· event-sampling

· ratings

4. Inter-observer agreement as a measure of reliability

B. Report measures - Interviews and Questionnaires

1. Who should respond?

2. Structured or unstructured?

3. Problems with report measures

4. Writing questionnaires

Discussion

1) In your group, design a questionnaire to assess people’s self-esteem. First, define self-esteem (conceptually), then create a 10-item questionnaire that assesses it. Be sure to use the principles described in chapter 4 of your text.

2) Trade questionnaires with another group. What strengths and limitations do you see to the new questionnaire? Discuss how you might use an interview to assess self-esteem. What are the advantages and disadvantages of using a questionnaire vs. interview to assess something?

3) When are report measures more appropriate? When is observation more appropriate?

C. Performance Measures

D. Physiological measures

Back to top