Analysis of Variance or ANOVA is a useful statistical method that helps in assessing a continuous dependent variable with respect to three or more two categorical independent variables consisting of more than two categories. The essence of performing the ANOVA is to find variability of scores among independent variables, or if they are significantly different. The null hypothesis (Ho) is that the means of all the independent variables are equal, and the alternative hypothesis (Ha) would be that the means of the independent populations are not equal. ANOVA is not capable of identifying where the differences among the independent variables lie; whether it is between the first and second variable or the first or fourth variable, etc (Pallant, p. 242).
A simple yet useful application of ANOVA is in the area of sales. Monitoring the income of sales departments is relevant in knowing the effectiveness of promotions and services. The income would be the dependent variable. The independent variables would depend on the kind of analysis that the marketing researcher would like to focus on. An example would be the determination of the effect of the location of a book in a bookstore in terms of volume of sales. Three or more locations could be thought of, like (1) near the entrance gate, (2) on an elevated area of the bookstore, (3) near the exit door of the bookstore, etc. Data on the volume of sales must be obtained, and the statistical treatment may already be implemented, together with the underlying assumptions in order to perform ANOVA.
Again, if it will be proven by ANOVA that the mean sales among the various locations are different, meaning the Ho is rejected, further statistical techniques should be used to identify where these differences lie. But the marketing researcher would already have a statistical proof that the locations of that certain book affect its sales volume and may have further investigations as to where new books could be placed in order effectively hasten or increase the popularity of the book and later on reflect a high volume of sales. If the Ho is accepted, efforts and money will not be wasted anymore if changing the current arrangement of the bookstore will not be done to improve the volume of sales of that book and other books of interest because the mean volume of sales are not equal.
Reference
Pallant, J. F. (2007). SPSS Survival Manual: A Step by Step Guide to Data Analysis Using SPSS. Retrieved May 4, 2009, from http://books.google.com/books?id=j1Q7H46y1rYC&pg=PA242&dq=analysis+of+variance&lr=&as_drrb_is=b&as_minm_is=0&as_miny_is=2007&as_maxm_is=0&as_maxy_is=2009&as_brr=3&as_pt=ALLTYPES#PPA242,M1
