LECTURE 7.2


DESCRIPTIVE METHODS

I. Observation: To describe behavior and get hypotheses

A. Observation should be

1. Systematic: no memory

2. Objective: "just the facts"

a. The value of instruments

b. The value of checksheets and training (Content analysis)

3. Nonreactive: observer should not affect observed

B. Observation common in biology, sociology, ethology

C. Naturalistic versus participant observation

D. Difficulties with observation

1. Sometimes impossible

2. Sometimes impractical

3. Sometimes unethical: subjects' rights to privacy

II. Case study

A. Learn from unique or extreme

B. Problems with drawing conclusions from case studies

1. Internal validity: Lack of comparison group

2. External validity

III. Archival data

A. Nonreactive

B. May not be accurate: Rape reported to police not the same as actual rapes.

C. Accuracy may change over time as a function of changes in record keeping:

1. Pre 1960, Cruelty only reason for divorce

2. Computerization of records

D. The problem of group data

E. The problem of poor measures

F. An illustration of the value of archival data: David Phillips (1979) work on the effects of publicizing a suicide on actual suicide

1. Discounted seasonal factors (time-series design)

2. Found only happened in area where study publicized

3. Car accident victims died faster after suicide story (hitting gas rather than brake)

4. Single fatality car accidents increase after publicized suicide, multiple fatality crashes after publicized murder-suicide

5. If suicide victim young, young die in car crashes; if old, old die in crashes

IV. Ex post facto data

A. The problem of after the fact

B. The problem of studying subject variables

1. Often, your way is only one way of classifying differences between subjects, differ in many other ways as well (e.g., neurotic vs. non-neurotic might also be low SES vs. high SES, introvert-extrovert, etc.)

2. Difficulty of getting good sample from each group

3. Difficulty of arguing that measure is equally valid for each group

C. Subject variables are intrinsically interesting (depressed vs. nondepressed, male vs. female, etc.)

D. Interactions involving subject variables are more revealing than main effects

V. Tests

A. Assessing validity

B. Ethical considerations


Back to Chapter 7 Main Menu