Quote:
|
Originally Posted by iantresman
Critical Examinations of QSO Redshift Periodicities and Associations with Galaxies in Sloan Digital Sky Survey Data (2005) Su Min Tang, Shuang Nan Zhang
Just reading through the paper above, the authors write (page 8):
"After smoothing off the sharp edges in the lowest and highest redshifts, a periodicity around Δz= 0.67 is detected in the full sample of SDSS QSOs, as shown in Fig. 9; however a periodicity of Δz= 0.67 ± 0.05 or any other frequency is not found in the 2dF QSOs, as shown in Fig. 10." But conclude the section by writing:
"In sum, there is no evidence for intrinsic periodicity in redshifts of QSOs." Can anyone help explain why the sum is no periodicity, even though they found a periodicity?
I think M.B. Bell, and D. McDiarmid explain it in their paper, " Six Peaks Visible in the Redshift Distribution of 46,400 SDSS Quasars Agree with the Preferred Redshifts Predicted by the Decreasing Intrinsic Redshift Model" (2006), but I don't understand it.
Regards,
Ian Tresman
|
Well, the authors themselves explain it, in the very same paragraph as the quote you selected!
Quote:
|
Such a difference between these two surveys is not surprising since the redshift-dependent spectroscopic completeness is relatively flat in 2dF (Croom et al. 2004), while in SDSS the spectroscopic completeness varies drastically at some redshifts (Richards et al. 2002). It is therefore improper to use all QSO redshifts in SDSS to probe any intrinsic periodicity without addressing selection bias. To further investigate whether such a periodicity around Δz = 0.67 in SDSS QSOs is spuriously produced by various incompleteness as function of redshift, we select a high-completeness sub-sample of 23,109 QSOs with mi < 19 and z < 2 in SDSS DR3, and three sub-samples containing QSOs in low-completeness regions. As shown in Fig. 11, no periodicity is found in the high-completeness sample where the power spectrum is consistent with a continuously ascending curve due to the low frequency component of the redshift distribution, whereas in different low-completeness samples, strong periodicity always appears, but with different peak locations (0.88 in (b), 0.67 in (c) and 0.74 in (d)). This should be a strong indicator that the peaks in low-completeness samples are caused by different selection effects in different samples. In sum, there is no evidence for intrinsic periodicity in redshifts of QSOs.
|
In short, the apparent periodicity arises from selection effect (the bane of an astronomer's life).
The Bell and McDiarmid paper is a valiant attempt to tease out a periodicity in the SDSS QSO z data, and address the Tang and Zhang analyses. I feel this second paper falls short of achieving its stated objective, not least because it addresses selection effects inadequately (I also think their statistical analyses are less robust than Tang and Zhang's, but haven't looked at it in sufficient detail to say for sure).