Analysis of Variance: ANOVA Designs
Types of Experimental Designs Analyzed by ANOVA
Various types of experimental designs can be analyzed using Analysis of Variance, commonly known as ANOVA.
This guide discusses ANOVA designs and defines key terms utilized in the analysis.
Key Terms and Definitions
Independent Variable:
A variable manipulated by the experimenter.
Example: In the smiles and leniency case study, the effect of four types of smiles (neutral, false, felt, miserable) on leniency shown to a person was investigated.
In this context, the type of smile is the independent variable.
Factor:
Used synonymously with the independent variable in ANOVA design.
In the smiles and leniency case study, the type of smile is also termed a factor.
Levels of a Factor:
The number of different states or conditions of the factor being studied.
In this case, there are four levels as there are four types of smiles compared.
Types of ANOVA
One-Way ANOVA:
Conducted when there is only one factor, as demonstrated in the smiles and leniency case study.
Two-Way ANOVA:
Conducted on experiments that involve two factors.
Example: The obesity and bias case study examines the influence of two factors:
The weight of the woman sitting next to the applicant.
The relationship between the applicant and the woman on job qualification ratings.
Between Subjects vs. Within Subjects Factors
Between Subjects Factor:
When different subjects are assigned to different levels of a factor.
Comparison is made between different groups of subjects.
Example: In the smiles and leniency study, four levels of the factor (type of smile) were represented by four distinct groups of subjects.
Within Subjects Factor:
When the same subjects are tested under each level of the factor.
This implies that comparisons are made within the same group of subjects, typically in repeated measures format.
Example: In the ADHD treatment study, subjects received each of four dosage levels (0.15, 0.30, 0.60 mg per kilogram).
Thus, there was only one group of subjects, with comparisons made across conditions within this single subject group.
Complex Experimental Designs
Often, studies may involve more than one factor.
Example: A hypothetical study examining the effects of age and gender on reading speed testing boys and girls aged 8, 10, and 12.
This results in six distinct groups as combinations of age and gender are tested.
The analysis in such scenarios would also utilize Two-Way ANOVA.
Factorial Design:
When all combinations of the different levels of factors are included in the study.
The design in the age and gender study can be described concisely as an “Age 3 by Gender 2 factorial design.”
Here, ‘3’ represents the three levels of age and ‘2’ reflects the two levels of gender.
Complex designs can contain multiple factors and may integrate both between and within subjects variables.