Amir H. Ghaseminejad, explains how we can judge if the means of three normal populations have changed as a result of treatments. By taking samples and dividing a measure of bias between means of samples by the a measure of random error, we have a statistic that has F distribution. If the populations are not different as a result of treatments the F statistic will be close to 1. and if the F statistic from samples is bigger, which according to F distribution based on correctness of NULL hypothesis has little chance to happen, Then may be not should be rejected.
Thanks a lot for uploading this!
ampikine 10 months ago