Tests of irregularly sampled stochastic time series for AGN

Since most observational material relevant for the variability of Active Galaxies is not available in regularly sampled lightcurves, such as have been collected by the AGN WATCH, using more data, to obtain reverberation results for more objects, one will have to rely on irregularly sampled lightcurves. Previous analyses of irregularly sampled observations suggest that it is possible to obtain results from such data sets. To evaluate the statistical significance of such determinations, we constructed simulated observational data of a stochastic process with random sampling, different transfer functions and noise conditions. These allow to estimate the statistical significance of cross-correlation calculations (CCF) on irregularly sampled data, as compared with the known input process. The results suggest that, in these conditions, the CCF is sufficiently robust to allow delays to be determined, but a careful error treatment is required.