Estimation of confidence intervals in measurements of triple-gauge boson couplings

Estimation of triple-gauge couplings at LEP2 is performed within the theoretical models which impose different constraints on the measured values. Estimation of confidence intervals is delicate due to a presence of model excluded regions of parameter space. In the paper a comparison of widely used classical central confidence intervals and the ones, based on the likelihood ratio ordering, is presented. The advantage of the latter method is that it always results in non-empty confidence intervals and properly takes into account a possibility of measurements outside the interval of model-allowed values, thus giving correct coverage for any possible measurement outcome.