ONE-FACTOR ANOVA
ANOVA single factor tests for difference means for k related sets of measures on continuous data.
Let y1, y2, .... be one data series with mean mux.
Let x1, x2, .... be a second data series with mean muy.
Let z1, z2, .... be a second data series with mean muz.
and so on.
Each series is an independent set of measures on the same object.
Then we consider the test of
H0: muy = mux = muz = --
versus
Ha: at least one of muy, mux, muz, ... not equal to 0
As an example, consider the same data as in Excel:
Difference in Two Means
Use Tools | Data Analysis | ANOVA: Single Factor
Select the input range to be the two series (list price and sale price)
To get a p-value = 0.704987335.
This is the same result as using a test for difference between unpaired
means with equal variances,
such as by using Tools | Data Analysis | t-test: Two-Sample assuming
Equal Variance.
TWO-FACTOR ANOVA
One can extend one-factor ANOVA to two-factor ANOVA. See a textbook.
REGRESSION FOR ONE-FACTOR ANOVA
One can use regression to implement one-factor ANOVA.
Let there be k series: series 1, series 2, ..... , series k.
Stack all the series into one long column (along with an indication
of which series it comes from).
Create (k-1) dummy or indicator variables D2, ..., Dk
where D2 = 1 if series 2 and 0 otherwise, D3 = 1 if series 3 and 0
otherwise, and so on.
Regress the long stacked series on a constant and k separate regressors.
Then test H0: using the F-test for the overall fit of the regressors
D2, ...., Dk.
This is the p-value given at the end of the ANOVA table that comes
with regression output.
One can also use regression to implement two-factor ANOVA.
For further information on how to use Excel go to
http://www.econ.ucdavis.edu/faculty/cameron