How to Solve an ANOVA

Formulate the research hypothesis. If, for example, a researcher wanted to study the effect of phone color on the speed in which operators at a call center answer the phone, the researcher would test three or more different groups of operators. The first group would get red phones, the second group would get blue phones and the third group, yellow phones. The question of interest to the researcher would be if the average time to answer the phone was significantly lower in one of the groups as opposed to the other two.

Enter the data using a column labeled Operator, Color and Answer Time. Run the ANOVA test by selecting this option from the main menu of the software package chosen.

Check the printout for the F-ratio. This is the inferential statistic calculated by the ANOVA test. Next, locate the p-value in the printout. If the p-value is less-than or equal to .05, the value of the F-ratio you obtained for this ANOVA is significant. An F-ratio that is significant means that the mean of one of the groups is significantly different from the others.

Look at the means calculated by the computer program to ascertain which of the three groups was significantly different from the others. In this example, if the mean for the red phone group was 8 seconds, the blue phone group was 12 seconds and the yellow phone group was 15 seconds, you would conclude that phone color had a significant effect on speed-to-answer and that using a red phone produced a significantly better result than using a blue or a yellow phone.