A REASSESSMENT OF SOME POLARIZATION STANDARD STARS

There have been recent reports suggesting that several polarized standard stars exhibit variability. In the investigation here, we concentrate on the work of Bastien et al. [AJ, 495, 900 (1988)] who claim that the majority of their measurements reveal instability of the catalogued standards (11 stars out of 13). We criticize their statistical assessments and comment on the shortcomings in their data and analyses. Firstly, it is demonstrated that their test based on comparisons of internal and external variances does not give rise to criteria for assessing variability. Secondly, their normality tests based on kurtosis and skewness are also shown to be fallacious and that in any case they lack rigor. Problems with the homogeneity of the observations are discussed with respect to dead-time corrections, instrumental polarization and modulation efficiency, all of which may give rise to variable systematic errors within the pooled data. For four stars which have sufficient homogeneous measurements, tests previously used on catalogues of unpolarized stars have been adopted to investigate the data and only one star (HD 111613) provides firm evidence for variability.