The type-I error rate for a single statistical test, that is for the decision process about a single null hypothesis. It is also called the individual error rate.
Experiment-wise type-I error rate
The type-I error rate that is associated with a hypothesis that is rejected if any of several null hypotheses are rejected. It is also called the family error rate of the simultaneous error rate.
Two statistical tests are independt if the two test statistics are independent ransom variables. This will be the case if the direction and magnitude of sampling error in one data set provides no information about the direction or magnitude of sampling error in the other data set. Notice that this is a property of the data sets, not of the test procedures.
two hypothesis are independent if knowing whether or not one of them is true provides no information about whether or not the other one is true.
Multiple comparison test
used to compare the means of two groups after you have rejected the null hypothesis in an ANOVA with more than two groups. A multiple comparison test differs from a two-sample t-test in that the experiment-wise error rate is controlled for a set of several such tests.
Unplanned multiple comparison test
A multiple comparison test is unplanned if the specific pair-wise comparisons that are made are selected after looking at the data.
Planned multiple comparison test
A multiple comparison test is planned if the specific pair-wise comparisons that are made are selected before looking at the data.
An ANOVA design in which each group has the same sample size.
(MS) is a sample variance, i.e., an estimate of a variance
Sum of squares
(SS) is a sum of squared vales that, when divided by its associated degrees of freedom, yields a mean square.
Degree of freedom
Equal to the number of independent parameters that can be estimated from that data set. The degrees of freedom associat4ed with a SS are equal to the degrees of freedom that are available in the data set from which the variance will be estimated.