SEQUENTIAL CONFIDENCE INTERVAL FOR TIlE REGRESSION COEFFICIENT

SUMMARY The object of the present investigation is to consider a robust procedure for the problem of providing a bounded-length confidence interval for the regression coefficient (in a simple regression model) based on Kendall's (1955) tau. The problem of estimating the difference in the location parameters in the two-sample case may be viewed as a special case of our problem. It is shown that the estimate of the regression coefficient based on Kendall's tau [cf. Sen (1968)], as extended here in the sequential case, possesses certain desirable properties. Comparison with the procedure based on the least squares estimator [considered by GIeser (1965) and Albert (1966)] is also made.