Power Spectrum Estimation II. Linear Maximum Likelihood

This second paper of two companion papers on the estimation of power spectra specializes to the topic of estimating galaxy power spectra at large, linear scales using maximum likelihood methods. As in the first paper, the aims are pedagogical, and the emphasis is on concepts rather than technical detail. The paper covers most of the salient issues, including selection functions, likelihood functions, Karhunen-Loève compression, pair-integral bias, Local Group flow, angular or radial systematics (arising for example from extinction), redshift distortions, quadratic compression, decorrelation, and disentanglement. The procedures are illustrated with results from the IRAS PSCz survey. Most of the PSCz graphics included in this paper have not been published elsewhere.