Research Methods
for the digitally inclined
by
Stephen R. Schmidt

Dispositional Variables

When properties of the subjects are treated like an independent variable

I. Logic of Dispositional Research

II. Examples of Dispositional Research

III. Problems with Dispositional Research

IV. Possible Solutions to the selection problem

V. Developmental Research

I. Logic of Experiments

A. Major components of Between Subjects Designs:

1. Hold all factors other than the I.V. constant
2. Random selection of subjects to conditions

However, what if we want to study a variable that we cannot really manipulate?

B. Dispositional Variables:

Differences between populations or sub populations that become the topic of investigation in and of themselves.

II. Some Examples

I.Q. levels
clinical population differences
social class or economic differences
racial differences
sex
sexual preference
age

A. Handedness Example

Porac & Corenís (1981)

Coren (1989)

Conclusion: Left handed people are more accident prone, and thus die at a younger age. Can you think of alternative explanations of these results?

B. Fitness Example

Deaths per 10,000 people/year

Deaths per 10,000 people/year

Conclusion: Increasing your level of exercise will reduce your risks of heart disease and cancer. Can you think of alternative explanations of these results?

III. Problems with Dispositional Research

A. Selection Problem:

Basic assumption of between subjects design is called into question. We can no longer assign subjects to conditions randomly, so how can we insure that the different groups are the same?

B. Nature of the comparison:

What are you trying to conclude?

Example: male versus female mathematical abilities

IV. Possible Solutions

A. Between subjects matching:

Try to equate the groups on other important variables.

- May not be possible.
- You may not know what the important variables are.

B. Use a within-subjects design.

Try to test the person under more than one value.

Examples:

pre versus post menses
pre- versus post menopause
alcoholic versus per alcoholic
follow subjects as they age

V. Developmental Psychology

A. Nature of developmental change

- not just age, but a group of variables correlated with age.

1) biological change

muscle coordination
hormonal changes (puberty, pregnancy, menopause)
changes in brain structure

2) Learning

instrumental and classical conditioning
new behaviors acquired
language learning

3) Experience:

Normative Age graded experiences

Examples:

age 6 go to school
age 16 allowed to drive
age 21 allowed to drink

In developmental research, we are interested in the development or changes in structures and behaviors of individuals as a result of these factors.

B. Two major methods

Cross sectional designs
Longitudinal designs

1. Cross sectional designs

a) Definition:

Different groups of individuals are selected at the same point in time to represent different developmental levels.

Example: DeLoache (1987) compared

21/2 and 3 year olds in a model task.
Toy hidden is model, child then asked to find the toy in the real room.

Results:

b) Advantages of Cross-sectional research

1) Can obtain a wealth of developmental information in a short period of time.
2) Only need to test each individual once, so there are no effects of repeated testing.

c) Disadvantages

1) Problem of equating the groups on important variables
2) Separation of developmental processes from cultural or historical differences

ìcohort effectsî generation of people raised in the same ìworldî or ìclimateî share experiences.

Examples:

W.W.II cohort
Vietnam
ìComputer Ageî

An example of a cohort problem: Intelligence as a function of age (Jones & Kaplan, 1945).

Age

Does intelligence start dropping off in the mid 20's?

The cohort problem:

2. Longitudinal Designs

a) Definition:

Follow the same group of individuals over an extended period of times and measure changes in performance.

Example: Eisdorfer & Wilkie (1973)

Gave a full scale WAIS four times to individuals over a 10 year period.

Eisdorfer & Wilkie2nd group

b) Advantages of Longitudinal Designs

1. Holds important subject variables constant

2. May detect effects not visible in cross-sectional designs

c) Disadvantages of Longitudinal Designs

1) Attrition: (Duke study)

Observation Date            Number of Subjects
I       1955                          256
II      1959                          192
III     1964                          139
IV     1966                          110

2) Effects of Repeated Testing

3) Cohort Effect threatens generlizability

4) Practical issues

- inflexible
- slow to provide answers
- publish or perish?

3. Cross-Sequential Designs

A. Definition:

Combines elements of longitudinal and cross sectional designs. Several different developmental groups are followed over a period of time.

Example 1: Schaie & Strother (1968)

Mental tests given to groups ranging in age from 20 to 60, and each group followed for 7 years.

Example 2:

Dennis (1966) (see also Cole, 1979)

Followed publication records of individuals in a number of fields over the course of their lives.
Included artists and scientist that lived to 79 or older.

Social Science group (history, philosophy)
Science (Biology, Chemistry, mathematics)
Artists (drama, liberal writing, poetry)

Dennis (1966) results:

Percentage of Work Produced During Each Decade of Life