Psych Research Methods Final

Selection effects

preexisting group differences

Maturation

natural change over time

Regression to the mean

extreme scores move toward average

Attrition

dropout is systematic

History

external event affects participants

testing/instrumentation

measurement changes

observer bias

researcher influence

Demand characteristics

participants guess hypothesis

Placebo effects

expectations influence outcomes

Reasons for null results:

• no effect

• weak manipulation: effect is small —> insufficient power (low N)

• poor measurement

• too much within-group variability (noise)

between-group variance

signal (good)

within-group variance

noise (bad)

High noise can hide…

real effects

design fix: comparison groups…

increase differences

design fix: random assignment…

to reduce noise

design fix: double-blind procedures and placebo…

eliminates biases

A researcher tests a stress-reduction program. Participants improve over time, but there is no control group.

What is the biggest threat?

A. Selection

B. Maturation

C. Instrumentation

D. External validity

B. Maturation

increased internal validity means what for external validity?

decreased external validity

A study compares two groups, but one group had higher ability before the study began.

What is the issue?

A. Regression
B. Selection effects
C. Attrition
D. Testing

B. Selection effects

Participants score at the very top of a measure regardless of condition.

What problem is this?

A. Floor effect
B. Ceiling effect
C. Confound
D. Attrition

B. Ceiling effect

“Factor” =

Independent variable

in a 2 × 4 factorial design how many IVs?

2 IVs

in a 2 × 4 factorial design how many cells?

8 cells total

Interaction effects

a difference in differences. Clues: “it depends” or “especially for”

eg. “exercise improves performance only for well-rested participants.

Graph interpretation: parallel lines

no interaction

graph interpretation: non-parallel lines

interaction

graph interpretation: crossing lines

strong interaction

where to find design info: the factorial notation (eg. 2×2×4) appears in what section?

the methods section

A study finds that caffeine improves performance, but only for well-rested participants.

This indicates:

A. Main effect
B. Interaction
C. Confound
D. Null effect

B. Interaction

In a graph, two lines are parallel across conditions.

This suggests:

A. Strong interaction
B. No interaction
C. Weak manipulation
D. Sampling bias

B. no interaction

A study is described as a 2 × 3 design.

How many cells are there?

A. 2
B. 3
C. 5
D. 6

D. 6

Quasi-experiment definition:

• no random assignment

• researcher lacks full control

what is the key issue with quasi-experiments?

there is a selection issue?

trade offs of a quasi-experiment

• decreased internal validity

• increased external validity

Interrupted time series (quasi-experimental research design)

• measure repeatedly before and after an event

• used for policies, disasters, real-world changes

Small-N designs

• focus on individuals

• strong internal control

• use reversal (ABA) designs

limitations of a small-N design

typically there are no inferential statistics

A researcher compares two existing groups without random assignment.

Main concern?

A. Attrition
B. Selection effects
C. Instrumentation
D. Demand characteristics

B. Selection effects

A study measures outcomes before and after a major policy change.

This is:

A. True experiment
B. Interrupted time series
C. Reversal design
D. Cross-sectional

B. interrupted time series

Behavior improves when a treatment is introduced, worsens when removed, and improves again when reintroduced.

This supports:

A. Weak validity
B. Selection bias
C. Causality
D. External validity

• reversal (ABA) design

• shows high internal validity

• C. Causality

direct replication =

same methods

conceptual replication

different methods, same theory

meta-analysis

• combines results across studies

• estimates overall effect

external validity

ask:

• can this generalize to:

  • other people?

  • other settings?

  • other cultures?

WEIRD samples. What’s the problem?

not representative of the global population

theory-testing mode =

internal validity

generalization mode

external validity

A study is repeated using identical methods and similar participants.

This is:

A. Conceptual replication
B. Direct replication
C. Meta-analysis
D. Extension

B. Direct replication

A study tests the same theory but with a different population and different measures.

This is:

A. Direct replication
B. Conceptual replication
C. Confound
D. Quasi-experiment

B. Conceptual replication

A study includes only college students from one university.

What is the main limitation?

A. Internal validity
B. Construct validity
C. External validity
D. Statistical validity

C. external validity