Volume 50 number 1 (2022)
Pre-whitening is a commonly used procedure in the frequency analysis of variable star light curves when multiple pulsation frequencies are active. However, each pre-whitening cycle introduces a new frequency into the data, often resulting in statistically significant frequencies appearing in the analysis which do not correspond to a pulsation within the star. In this paper we examine the effectiveness of a simple modification, permitted by some frequency analysis software packages, which we call restricted range analysis. We show that while restricted range analysis may considerably reduce the number of spurious frequencies generated it does not eliminate them. Thus each reported significant frequency must still be checked against the periodogram to confirm it matches a clear feature within it. Further, restricted range analysis appears to not detect some frequencies and these must be searched for in the periodogram. We conclude that while restricted range analysis is useful, it does not overcome all the problems associated with pre-whitening in frequency analysis.