What is ANOVA? • The test of hypothesis concerning the equality of two population means makes use of both the Z and t tests. • However, if there are more than two populations, the test for the equality of means could be carried out by considering two populations at a time. This would be a very cumbersome procedure. • One easy way out could be to use the analysis of variance (ANOVA) technique. The technique helps in performing this test in one go and, therefore, is considered to be important technique of analysis for the researcher. What is ANOVA? • The basic principle underlying the technique is that the total variation in the dependent variable is broken into two parts—one which can be attributed to some specific causes and the other that may be attributed to chance. • The one which is attributed to the specific causes is called the variation between samples and the one which is attributed to chance is termed as the variation within samples. • Therefore, in ANOVA, the total variance may be decomposed into various components corresponding to the sources of the variation. • For eg. the sales of chairs could differ because of the various styles, and sizes of stores selling them. What is ANOVA? • In ANOVA, the dependent variable in question is metric (interval or ratio scale), whereas the independent variables are categorical (nominal scale). • If there is one independent variable (one factor) divided into various categories, we have one-way or one-factor analysis of variance. • In the two-way or two-factor analysis of variance, two factors each divided into the various categories are involved. Completely Randomized Design in a One-way ANOVA • In this design, there is one dependent variable and one independent variable. • The dependent variable is metric (interval/ratio scale) whereas the independent variable is categorical (nominal scale). • A sample is drawn at random from each category of the independent variable. • The size of the sample from each category could be equal or different. Completely Randomized Design in a One-way ANOVA In order to test the null hypothesis, H0 : μA = μB = μC = μD against the alternative hypothesis H1 : At least two means are not equal (Treatment means are not equal) The results could be summarized in the table below called One-way ANOVA. Randomized Block Design in Two-way ANOVA • The main limitation of the completely randomized design is that all extraneous variables were assumed to be constant over all the treatment groups which may not be true. • There may be extraneous variables influencing the dependent variable. • In the randomized block design it is possible to separate the influence of one extraneous variable on a particular dependent variable, thereby providing a clear picture of the impact of treatment on test units. Randomized Block Design in Two-way ANOVA • Assume that the sales of a product is influenced not only by the price level (low, medium and high) but also by the size of the stores. The price level is assumed to be independent variable whereas the size of the stores are extraneous variable and could be treated as different blocks. • The hypothesis to be tested are the following: I Price H0 : μA = μB = μC (Assuming three price level) H1 : At least the two means are not same. II (Blocks or Stores) H0 : ν1 = ν2 = ν3 (Assuming three different sizes of stores) H1 : At least two means are not same. Randomized Block Design in Two-way ANOVA • The above results could be presented in a two-way ANOVA table as below:
For a given level of significance, the computed F corresponding to
treatments and blocks could be compared with the corresponding tabulated values to accept of reject the two hypothesis. Factorial Design • In factorial design, the dependent variable is the interval or the ratio scale and there are two or more independent variables which are nominal scale. • It is possible to examine the interaction between the variables. If there are two independent variables, each having three cells, there would be a total of nine interactions. • The hypothesis to be tested would be the following: Factorial Design Row-wise: H0 : Average of rows are equal. H1 : Average of rows are not equal Column-wise: H0 : Average of columns are equal. H1 : Average of columns are not equal Interaction: H0 : Average of all interactions between rows and columns are equal. H1 : Average of all interactions between rows and columns are not equal
The ANOVA table for factorial design can be computed in the usual manner to test the various hypothesis stated earlier using F statistic.