Estimating variance (4 of 4)
Next chapter: Confidence intervals
The number of
degrees of freedom
an estimate is based upon is equal to the number of independent scores
that went into the estimate minus the number of parameters estimated
en route to the estimation of the parameter of interest. In this case,
there are N independent scores and one parameter (μ)
is estimated en route to the estimation of the parameter of interest,
σ² . Therefore the estimate has
N-1 degrees of freedom. The formula for s² can then be written
as:
where df = N-1. Naturally, the greater the
degrees of freedom the closer the estimate is likely to be to σ².
Next chapter: Confidence intervals