On selection effects in pulsar searches

Abstract

The problem of selection effects as a major source of error in statistical studies of pulsar data is examined. A previously neglected selection effect is found which is a function of the period P of the pulsar. It is shown that short-P pulsars are more difficult to detect than longer period pulsars, particularly if the dispersion measures of the short-P pulsars are high. In addition, a declination-dependent selection effect is found in the II Molonglo Survey (II MS), and some evidence is presented for the existence of both selection effects in the pulsar data from this survey. The implications of these additional selection effects are examined for the 'injection' of pulsars whereby pulsars seem to switch on only at longer P. The II MS data is used to calculate that the observability of pulsars with P between 0.0 sec and 0.5 sec is about 18% less when the new selection effects are included than with previous calculations. The mean correction is 6% for P between 0.5 sec and 1.0 sec. It is concluded that injection is not qualitatively affected by these corrections.