Recently, I consulted a student in ANOVA, analysis of variance. What is an ANOVA? The term sounds like a show on PBS or something. (Okay that was corny). There are a few things about the ANOVA which you must understand once and for all. If you have an independent variable with more than two levels- it triggers the need for an ANOVA. By levels I mean categories. So if you have (for instance) race as an independent variable, it will have more than two classifications as answer choices. You must recognize what you have as independent variables. When it comes to hypothesis testing, you really want to know if the different classifications yield similar means. If the different classifications yield the same mean then all the data comes from the same population. Remember the premise about the mean statistic. It’s not just another ‘pretty face’ -lol, If your dependent variable is normally distributed with no skewing or funny business- the mean represents your data. If a particular category in your independent variable has a statistically significant mean that is different from the other classifications- it may come from another population all together. That is why you conduct an ANOVA. Therefore, we are left with the null hypothesis as follows:
Null hypothesis: µ1=µ2=µ3….=µn where ‘n’ denotes the number of groups in your independent variable.
Alternative hypothesis: at least one of the means are different
It is important to note that the ANOVA will NOT locate where the differences are…think of the ANOVA as a nosy neighbor who gossips but won’t disclose the details. She or he will not tell you which group(s) are different- just know that at least one of them are….lol
If you enjoyed this blog there are moore to follow.-Amy
