Study Guide Test 3 Review

Factor

thing being studied

Levels

Different conditions or values assigned within a factor in an experiment. (conditions/groups)

Between groups factor

A design where different participants are assigned to different levels of the factor.

Within groups factor (repeated measures factor)

A design where the same participants are exposed to all levels of the factor.

Longitudinal study

A research design that follows the same subjects over an extended period.

Carryover effects

When the effect of a treatment persists and affects subsequent measures.

how are carryover effects solved?

Carryover effects are typically solved through counterbalancing, randomization, or washout periods between conditions to minimize their impact on experimental results.

Complete counterbalancing

all possible treatment orders

how do you determine orders in complete counterbalancing

for n conditions calculate

n x (n-1) x (n-2) …. x 1

example:

2 conditions (A, B):

2! =2 ×1 =2

3 conditions (A, B, C):

3!=3×2×1=6

Independent variable with 2 levels

There are 2 possible orders in complete counterbalancing.

Independent variable with 3 levels

There are 6 possible orders in complete counterbalancing.

Incomplete counterbalancing

A method where only a subset of all possible orders is used.

Latin Square design

A counterbalancing approach that ensures each condition appears once in each position. (n x n)

can you counterbalance a longitudinal design?

No, because age/maturation/etc are the independent variable

developmental design strategies

cross-sectional designs, longitudinal designs

cross-sectional designs (definition/drawbacks)

These use different age groups to show a change over time during development.

drawbacks: vulnerability to selection/recall bias

Cross-sectional developmental studies suffer from cohort effects, but longitudinal studies do not.

longitudinal design drawbacks

dropout

Internal Validity

When an experiment is constructed so that extraneous factors are sufficiently controlled, we can conclude that changes in the dependent variable are due solely (maybe) to the differing levels of the independent variable.

Potential Problems with Dependent Variables

Ideally you want your dependent variable to be able to capture a range of performance.

– Floor and ceiling effects

– Restricted ranges

confounding variables

external factor that affects both dependent and independent variables, introduces error and bias.

What to do about confounding variables

1. Hold them constant, or match your groups/levels.

2. Make sure that their influence is not systematically

occurring in one or more of your groups/levels. In

other words, try to randomly assign the influence of

the confounding variable across the levels of the

experiment. In theory, this type of approach should

balance out the effect of the confounding variable.

• Problem: You need enough subjects to be able to use

this approach.

Statistical Control of Potential Confounding Variables

Variables that we know will affect the measurement of the dependent variable, removed influence via statistical control (covariates)

Increasing Generalizability

Increases External Validity

Quasi-Independent Variables

anything that cannot be directly controlled and manipulated by the researcher.

• can be subject variables

Natural treatments

Events in the real world.

• Major historical events

• Natural disasters

• Changes in rules/regulations

Before and After Designs

These designs do not include a comparison group. just show differences, especially in photos.

The Interrupted Time Series Design

simplest type of design includes a pretest and

posttest.

Pre-test → Intervention/Treatment → Post-test

Ways to Improve Before and After Designs and Designs where you cannot counterbalance

Multiple Group Before and After Design (showing that results occur across multiple groups)

Repeated Measures Panel Design (many pre/post tests are examined)

Multiple Baseline Design/Staggered Start Design

Age Effects

How your age might influence behavior independent of generational and/or shared historical events.

Cohort Effects

effects that can be attributed to when an individual was born.

Period Effects

istorical influences on behavior that affect a variety of ages

• (e.g., 9-11, David Bowie’s death)

Cohort effects vs Period Effects

Cohort effects apply to a specific group of people (usual defined by age), while period effects/events affect everyone who was exposed to the event (but maybe not in the same way)

Case study

An in-depth analysis of a single subject or group, useful for exploring complex issues.

AB study design

A design with a baseline (A) and a treatment phase (B), but can have limitations in establishing causal relationships.

Withdrawal or reversal design

A method where the treatment is applied and then removed to observe changes in behavior.

Multiple baseline design

A design that staggers the introduction of treatment across individuals or settings to demonstrate effects.

Vividness effect

A phenomenon where more vivid or emotionally charged information is better remembered.

Placebo effect

When participants experience a perceived benefit from an inert treatment.

Confirmation bias

The tendency to check for information that supports one’s beliefs while ignoring contradictions.

Survey advantages

Including efficiency and the ability to collect data from a large population.

Survey drawbacks

Potential issues with self-report accuracy and sample bias.

Sampling frame

A list from which a sample is drawn for a survey.

Probabilistic sampling

Sampling where each member of the population has a known chance of being selected.

Non-probabilistic sampling

Sampling where not all members of the population have a chance of being selected.

Simple random sampling

Everyone in the population has an equal chance of selection.

Systematic random sampling

Selecting participants based on a fixed interval from a randomly chosen start.

Stratified random sampling

Population divided into subgroups, and random samples are taken from each.

Cluster sampling

Dividing the population into clusters and randomly selecting entire clusters.

Convenience sample

Participants selected based on their easy availability.

Survey open-ended questions

Allow respondents to answer freely, providing richer data but can be hard to analyze.

Forced Choice question

A survey question format where respondents must choose between options.

Likert scale

A rating scale that measures attitudes or opinions on a level of agreement.

Semantic differential scale

A scale measuring the meaning of concepts across bipolar adjectives.

Context effects

Changes in survey responses based on the order or context of questions.

Socially desired responding

Answering survey questions in a way that is perceived favorably by others.

Double negative question

A question that contains two negatives, leading to confusion.

Single negative question

A question with one negative which can lead to misinterpretation.

similarities/differences between longitudinal and within subject studies

Differences:

longitudinal studies follow a set group of people over a long period of time. Within subject studies compare different conditions within the same subjects at a single point in time.

Similarities:

Both study designs aim to understand changes and effects over time and conditions.

Small n designs

Research designs that focus on a small number of participants; often criticized for low generalizability.