From the autocovariance function of residual errors in least-square curve fits, one can
estimate the random error present in experimental data, along with the relative proportion of
systematic to statistical error. The methods outlined here have important implications for
teaching error analysis.