Outline

Introduction to correlation.

Characteristics of correlations

Where and Why they are used

Problems with Correlational Research

Correlations are used to measure and describe relations between two variables, e.g. relation between Grade and time to complete an exam (see graph).

Figure 16-1 (p. 522)
The relationship between exam grade and time needed to complete the exam. Notice the general trend in these data: Students who finish the exam early tend to have better grades.

No attempt is made to control or manipulate variables.

Two scores are required for each participant.

CORRELATIONS MEASURE THREE CHARACTERISTICS OF THE RELATION

Direction: Whether it is positive (e.g. scores in IV₁ increase as scores in IV₂ increase) or negative (e.g. scores in IV₁increase as scores in IV₂ decrease).

Form: Whether the relationship is linear or not.

Degree: How strong the relationship is.

WHY WE USE CORRELATIONS

In order to make predictions

In order the demonstrate the validity of measures

In order to determine reliability

In order to validate predictions made by theories

PROBLEMS WITH CORRELATIONAL RESEARCH

· Direction of cause and effect: We do not know which variable caused the other.

· Third-variable problem: Some uncontrolled variable may be responsible for the observed relationship between the variables.

Because of these problems correlational research cannot be used to establish causal relationships among variables.

The tendency to interpret correlations in terms of cause and effect is a common error.

1. Hippocrate’s Good News Survey

People who often ate Frosted Flakes as children had half the cancer rate of those who never ate the cereal.

2. Children who took vitamins are more than twice as likely to use marijuana later and cocaine.

3. People who had routine physicals in the previous 3 years were twice as likely to report high blood pressure and cholesterol levels.

4. As the quality of a state’s day care programs goes up so does the reported rate of child abuse.