OLS regression estimates rates of change in conditional means.
The permutation testing approaches are identical to those used
for LAD regression, and are described in Kennedy and Cade (1996)
and Anderson and Legendre (1999). The test statistic is similar
in structure to that for LAD regression, except for OLS T_obs
equals (sum of squared residuals for
reduced parameter model - sum of squared residuals for full parameter model) / sum of squared residuals for full model.
Generally called indirectly via the lad function call with the option OLS set equal to TRUE.
References
Anderson, M. J., and P. Legendre. 1999. An empirical comparison of permutation methods for
tests of partial regression coefficients in a linear model. Journal Statistical Computation and Simulation62, 271--303.
Kennedy, P.E., and B.S. Cade. 1996. Randomization tests for multiple regression. Communications in Statistics - Simulation and Computation25, 923--936.