Degrees of Freedom

Degrees of Freedom is an integer which represents amount of information. In a set of n experiments, there are n degrees of freedom (df) When the average is computed, it uses up one degree of freedom leaving n - 1 df in the deviations: Y-average.

When there are r replicates of a trial, the are r(df) in those replicates. The average has one df; then the standard deviation (s) contains (r - 1) df. When several s values are pooled, the pooled s contains the df which is the sum of the df of the contributors.

Each b coefficient in a linear model from a good design contains one df.

When confidence limits are calculated, the t multiplier is looked up for the df of the s.