Chapter 10: Using between-subjects and within-subjects experimental designs

factorial design

incorporate two or more independent variables in a single design

main effect

Refers to whether or not statistically significant differences exist between the levels of an independent variable in a factorial design

interaction

in a factorial design, an interaction occurs when the effect of one independent variable depends on the level of another independent variable

meaningfulness

Validity speaks to if we can draw any conclusions about our construct of internal from our measurements

-Meaningfulness speaks to what kind of conclusions we can draw

independent variable

• The factor of interest to the experimenter, the one that is being studied to see if it will influence behavior. 

• In experimental research, this is the variable that is manipulated. 

• The value is determined by the experiment, not the participant or subject. 

dependent variable

• Variable measured (observed and recorded) in the study.

• Value is determined by the behaviors of the subject and may depend on the value of the independent variable. 

• Those behaviors that that are the measured outcomes of experiments. 

confound

Two variables that combine in such a way that the effects of one cannot be separated from the effects of the other.

control variables

A potential independent variable that is held constant during an experiment

Examples

Time of day, temperature, time of last meal, knowledge of subject.

between-subject design

Each treatment is administered to a different group of subjects.

within-subject design

A single group of subjects is exposed to all of the treatments.

single-subject design

Subjects are randomly assigned to different treatment groups.

randomized two-group design

Randomized groups design that includes only two groups.

• Very simple design that is simple to conduct.

• Requires relatively few subjects.

• No pretesting or categorizing of subjects needed.

multiple group design

• Additional levels of the independent variable can be added to form a randomized multigroup design.

parametric design

Manipulating the independent variable quantitatively

nonparametric design

Manipulating the independent variable qualitatively

multiple control group design

• which includes a number of control groups, is a variant of the randomized multigroup design.

matched group design

• Matched sets of subjects are distributed at random, one subject per group, into different treatment groups.

• Used when subject characteristics correlate strongly with the dependent variable.

• Reduces error variance.

• Allows you to control subject variables that may mask the effect of the independent variable.

carryover effects

• Exposure to a treatment affects performance in a subsequent treatment.

counterbalancing

 Presenting various treatments of the experiment in a different order for different subjects.

matched group

A single group of subjects is exposed to all of the treatments.

anova

• Assume for a moment that all 3 conditions come from the same population.

• That means they would have the same population means: μ = μ_A =  μ_B =  μ_C

• The same holds for population variance:    σ = σ_A =  σ_B =  σ_C