I. Variables

A. Types

1. Independent variables vs Dependent variables (in general)
2. Predictor variables vs Criterion variables (in regressions)
3. Subject/Demographic/Personal variables
3. Confounders, Mediators, and Moderators

B. Operational definitions = definition of construct made by explaining how construct will be measured

C. Relationships between (correlations!)

1. Types of relationships

a. Positive linear
b. Negative linear
c. Curvilinear
d. No relationship

2. Nature of correlations

a. Assess direction and strength (r) of the relationship
b. Cannot tell us about causal factors

 

II. Types of research questions

A. Descriptive: How does being at war affect children's behavior?
B. Associational: What characteristics of war are associated with changes in behavior?
C. Directional: Is watching news coverage of war associated with greater aggression in children?
D. Causal: Does war cause children to be more aggressive?

 

III. Creating an instrument/measure/test

A. Start with theory (e.g., biological theory of sex differences in aggression)

1. Theories should be logical, internally consistent, and falsifiable
2. Understand theory's assumptions (e.g., androgen increases aggression)
3. Develop operational hypotheses that ought to be true if theory is true
4. Theories can be disproven and supported but not proven

B. Design or select the measuring instrument

1. Strive for validity and precision

a. Face
b. Content
c. Criterion validity (Divergent and Convergent)
d. Predictive validity

2. Test for reliability (go over True Score Theory)

a. Test-retest (Is test consistent across time?)
b. Inter-rater (Do different raters/scorers agree with each other?)
c. Alternate form (Do multiple versions yield same information?)
d. Internal consistency (Do different items on a scale measure same construct?)

3. Minimize error

a. Systematic error

1. Measure or try to control for Social Desirability
2. Use deception in obtaining informed consent
3. Use unobtrusive methods of data collection (e.g., physical trace, archive)
4. Experimental or statistical control

b. Random error

1. Standardize procedure
2. Experimental or statistical control

 

IV. Small group project:

A. Pick a construct you want to measure quantitatively
B. Create a 6-10 item scale, which will accurately and reliably measure this construct
C. Describe how you would test the reliability and validity of your scale