Transcript Funder B - Appalachian State University

Methodology: The Heart and Soul of
Psychology
1.
What makes a
psychologist different
from anyone else who
studies thoughts,
feelings, and behavior.
Only one thing, our
research methodology
2. A Protozoan Tale
Reliability and Validity
Reliability-The tendency of an instrument to
provide the same measurement on repeated
occasions.
Validity-The degree to which a test measures
what it is purported to measure.
Measurement Error
Measurement Error-Deviations from the true score
due to extraneous influences.
Five Ways To Reduce Error
1. Double check data entries
2. Standardize instructions
3. Measure significant attitudes and behaviors
4. Aggregation
5. Detect social desirability and lack of motivation
Double Check Data Entries
1. Even a few errors can greatly distort results.
2. The student with the 11200 SAT.
3. Confusing missing data and 0 on a computer
program.
Standardization of Instructions
1. All participants get the same instructions
2. Use of computers and other technological
devices
3. Train your investigators well
Measure Significant Attitudes and
Behaviors
1. In general, the more significant the attitude the
more reliable over time.
2. You can ask an attitude that may not have
existed until you brought it up: Mayor’s tax plan
Aggregation
1. Combining different measurements, usually by
averaging. Random errors cancel out.
2. Rolling dice
3. Taking heights
Measure Significant Attitudes and
Behaviors
1. In general, the more significant the attitude the
more reliable over time.
2. You can ask an attitude that may not have
existed until you brought it up: Mayor’s tax plan
Validity
1. The extent to which a test measures what it
purports to measure.
2. Construct-An idea about a psychological
attribute that goes beyond what an instrument
measures (e.g., intelligence, silliness).
3. Construct validation-Establishing the validity of
a measure by comparing it to a wide range of
other measures (e.g., Survey of Academic
Orientations (SAO))
Generalizability
The degree to which a measurement can be found
under diverse circumstances, such as time, context,
subject population, etc.
1.
2.
3.
4.
5.
Gender bias
No-shows
Cohort effect
Economic diversity
Ethnic diversity
Gender Bias
The sexual distribution of the sample should
mirror the sexual distribution of the population.
1. Combat soldiers.
2. Retention of undergraduates at a predominently
female institution
No-Shows
1. Persons who agree to participate in an
experiment but do not do so.
2. Research indicates that no-shows are different
than people who show up for experiments.
3. High degrees of non-participants impact the
generalizability of the study
Cohort Effect
1. The tendency for people who lived at one time
to be different than people who lived at a
different time (Will our results generalize across
time?)
2. Would you get Milgram’s results today?
3. Support and interest in American education
across generations
Ethnic Diversity
1. To what extent are ethnic groups appropriately
represented in a sample (are minorities over or
under represented?)
2. Assuming all minorities are the same
a. Kiowa and the Cheyenne
b. Catholics and Protestants, Sunni and Shia
Economic Diversity
1. College students are often the sample of
convenience
2. College students tend to come from middle or
upper middle class families
3. Most people primarily associate with persons of
their own economic level (Distinguished Lecture
Series: What to do about the poor?)