Pay 203 Exam 4

Within-subject design

Each subject receives every level of the independent variable at different times; compare each subject to themselves

Repeated-measures design

A type of within-subject design in which participants’ behavior is assessed multiple times in each condition and randomization of exposure to levels of the independent variable is not possible

Sequence effects

Order effects; experiencing one condition affects performance in a subsequent condition in within-subject designs

Progressive effects

A type of sequence effect; performance changes gradually accumulate over conditions (practice, fatigue)

Carry-over effects

A type of sequence effect; one condition influences performance on a subsequent condition

Counterbalancing

Arranging experimental conditions to minimize extraneous factors and reduce carryover/order effects

Longitudinal design

Study of variables in the same individuals over a long period of time

Attrition

Differences due to unequal dropout rates between groups

Factorial design

Design with 2+ independent variables where all levels are combined

Main effect

Differences between levels of one independent variable averaged over others

Interaction

Effect of one independent variable depends on level of another

Between-subjects factorial design

Each subject is exposed to only one condition

Within-subject factorial design

Each subject is exposed to all levels of each independent variable

Mixed factorial design

At least one between-subjects and one within-subjects factor

PxE factorial design

One manipulated (environmental) and one non-manipulated (person) variable

Repeated measures factorial design

Groups tested repeatedly; within-subjects factor; counterbalancing not possible

Level

A condition in an experiment

Factor

An independent variable

Relational research

Observes relationships between variables without manipulation

Contingency analysis

Determines if one binary variable depends on another binary variable

Correlational research

Examines the degree and nature of relationship between variables

Correlation coefficient

Numerical index of linear relationship (-1 to +1), commonly Pearson’s r

Direction of correlation

Indicated by sign (+ or -)

Degree of correlation

Indicated by value (weak: .10–.29, moderate: .30–.49, strong: .50–1.00)

Directionality problem

Unclear whether A causes B or B causes A

Third variable problem

Two variables are related due to a third variable

Spurious correlations

False correlation due to third variable or coincidence

Fallacy of affirming the consequent

Assuming that observing an effect proves a specific cause

Zero correlation due to nonlinearity

Variables are related but not in a linear way

Zero correlation due to truncated range

Limited variability prevents detecting correlation

Independent variable

Manipulated variable; suspected cause

Control variable

Variable held constant to remove its effect

Subject variable

Participant characteristic not manipulated but may influence results

Positive correlation

Variables increase/decrease together

Negative correlation

One variable increases while the other decreases

Zero correlation

No linear relationship

Quasi-experiments

Studies with at least one non-manipulated variable

Observational research

Describes behavior without manipulation

Archival research

Uses previously collected data for new purposes

Naturalistic observation

Observing behavior in natural environment

Participant-observer research

Researcher joins the group being studied

Behavior checklist

List of predefined behaviors used in observation

Partial concealment

Participants know observation occurs but not exactly what is recorded

Observer bias

Researcher interprets events based on expectations

Subject reactivity

Participants change behavior because they know they’re observed

Demand characteristics

Cues that influence participant behavior in a study

Posttest only design

Measure taken only after treatment

Pretest-posttest design

Measure taken before and after treatment

Pretest-posttest nonequivalent control group

Groups not randomly assigned; both measured before/after, only one gets treatment

Selection bias

Differences due to preexisting group differences

Small-n design

Single-case or small participant designs

Stable baseline performance

Behavior remains consistent before intervention

A phase

Baseline phase without intervention

B phase

Treatment phase with intervention

Withdrawal design

Treatment is applied, removed, and possibly reapplied to test effects

Multiple baseline design

Intervention applied sequentially across behaviors

Alternating treatments design

Rapid alternation of treatments to compare effects

Changing criterion design

Gradually changing behavioral requirements to assess intervention effects

One-shot case study

Single group measured once after treatment

Whole-interval recording

Record behavior only if it occurs for entire interval

Partial-interval recording

Record behavior if it occurs at any point in interval

Momentary time sampling

Record behavior at the end of interval

Planned activity check

Record number of individuals engaged in behavior at interval end

Survey

Collects data from a sample to generalize to population

Psychological assessment

Evaluates behavior and characteristics for diagnosis/treatment

Open-ended questions

Questions allowing free, detailed responses

Closed questions

Questions with fixed response options

Random assignment

Randomly assigning participants to conditions

Probability sampling

All individuals have a nonzero chance of selection

Random sampling

All individuals have equal chance of selection

Stratified sampling

Random sampling within homogeneous subgroups

Cluster sampling

Random sampling within heterogeneous groups

Convenience sampling

Sampling those easiest to access

Purposive sampling

Sampling based on expert judgment

Quota sampling

Sampling until specific subgroup numbers are reached

Snowball sampling

Participants recruit other participants

Subject/participant representativeness

Lack of random sampling may bias results

Demographic information

Population characteristics (age, gender, income, etc.)