Tuesday, 8 September 2015

Types of ANOVA



One-Way Analysis of Variance: This is the same as One-Factor ANOVA; the analysis includes only one factor (the thing that you believe is responsible for the variability) (e.g., dormitory).
Factor: dormitory     Levels: Moore, Steinbright & Butz
Hypotheses: as shown above
Two-Way Analysis of Variance: This is the same as Two-Factor ANOVA; the analysis includes two factors (e.g., dormitory and gender).
Factor: dormitory     Levels: Moore, Steinbright & Butz
Factor: gender          Levels: female & male
Hypotheses: 3 sets (1 for dormitory; 1 for gender; 1 for the interaction
                     between dormitory and gender)
Three-Way Analysis of Variance: This is the same as Three-Factor ANOVA; the analysis includes three factors (e.g., dormitory, gender, and class).
Factor: dormitory     Levels: Moore, Steinbright & Butz
Factor: gender          Levels: female & male
Factor: class             Levels:  freshman, sophomore, junior, senior
Hypotheses: 7 sets (don't worry about these, and those for four factor
                      ANOVA and beyond)
Completely Randomized One-Way ANOVAOne factor is used and for each level, observations are collected via independent random selection. This is the simplest type of ANOVA, one of the more commonly used analyses, and the one that we will be using in this class.    In other words, you decide on your levels (or categories or groups) and then you randomly select observations to be included in your samples (your typical approach to collecting data).

Useful Terminology

Factor: The factor is the thing, force, application, classification or criterion that you believe is responsible for the variation in the data. For example, one might believe that the choice of dormitory may have an impact on loudness.  In this case, the factor would be the dormitory.  Other examples: amount of drug administered; gender; geographic location; college campus; profession; ...

Level: The term, level, pertains to each type within the chosen factor. Other terms used in place of level are category and group.
For example, if the factor is dormitory, then we have three levelsMooreSteinbright, and Butz.  If on the other hand, amount of drug administered is the factor, then the levels might be: 0 mg, 10 mg, and 20 mg.

Reference
http://www2.cedarcrest.edu/academic/bio/hale/biostat/session23links/anovabasics.html

No comments:

Post a Comment