Analysis of Variance (ANOVA)
ANOVA compares data for different groups or for the same group for different measures to see if there is difference (Monash, n.d). If there is difference, it measures the amount of difference. For example, a nurse may wish to test the theory that drinking warm water helps obese people to lose weight. To do so, she sets two groups of participants where she ensures that one group takes 8 glasses of warm water every day per person and the other takes an equal amount of cold water. Physical activity and eating for the two groups should be maintained equal and the nurse collects daily weight for the two groups. The nurse then uses ANOVA to compare data for the two groups to see if weight loss is difference across groups and in doing so sees if weight loss is related to whether a person drinks cold or warm water.
- It calculates mean for each group known as group means
- It calculates overall mean, which is mean for the two groups combined
- It calculates total deviation for every person’s weight from the person’s respective group mean known as within group variation
- Next, ANOVA calculates the variation of each group’s mean from the overall mean known as between group variation
- Finally, ANOVA calculates and gives the ratio of between group to within group variation known as the F ratio. If F ratio is significant, it shows there is significant difference between the groups, which shows that drinking warm water significantly affects weight loss. If F ratio is not significant, it shows that drinking hot water has no significant impact on weight loss and the theory does not hold water.
Various available software packages perform ANOVA, some of which include SPSS, SAS and minitab.
References
Monash (n.d). Analysis of Variance (ANOVA). Retrieved from http://www.csse.monash.edu.au/~smarkham/resources/anova.htm
