Aroet,
It might be worth watching this (which I think was recorded much later than 2005).
This opening screen says that its from 2000.
I think I'd feel more comfortable with the conventional view, if someone actually explained how it is that Halton Arp (in that lecure) presents one piece of evidence with a probability of being due to chance of 1 in 10^7 (start listening at about 9.00 mins - but ideally listen to it all).
Also note comments at 21 mins in! Listen to it all Arouet - it isn't too technical. Notice also some of his asides about his social interactions with his colleagues!
I listened to the two bits you noted, but honestly, for me to even begin to try and figure this stuff out I need written sources, with citations. While I still won't be able to understand all the technical niceties, it at least allows me to track sources, seek different views, etc.
For example, he states that a certain paper of his got not attention. Without a citation, I can't even make a rudimentary investigation as to whether that is true. Regarding the 1 in 10M stat - we don't have nearly enough information in that short clip to evaluate that.
As I'm not knowledgeable enough to evaluate these claims on my own, my approach in these types of situations is to try and find a variety of sources that discuss it, and try and determine the gist of the debate. I find in lectures people aren't usually as specific as they are in papers as well.
I suspect Tang, Shuang and Zhang did not make use of all the detail that HA has available.
It's also possible that the reverse is also true. Here is how they describe their methods:
In an attempt to resolve these issues, we turn to the Sloan Digital Sky Survey (SDSS) (and also 2dF QSO Redshift Survey (2QZ) occasionally) to carry out this study, which have the largest homogeneous sample of data as well as the spectroscopic sub-classification of galaxies.
They also cite 3 of Arp's papers.
In looking at the papers that have cited Tang et. al, it appears that Arp addressed it in:
THE 2dF REDSHIFT SURVEY. I. PHYSICAL ASSOCIATION AND PERIODICITY IN QUASAR FAMILIES
This seems to be Arp's primary issue with that paper:
Tang & Zhang (2005) make it clear that periodicity results are null when using well-controlled samples that are up to 15 times larger than previous studies. We maintain that the primary difficulty in analyzing redshift fields based on an ejection hypothesis is simply the overlap that exists between fields of quasars that are within reach of two or more galaxies. Our quasar family detection constraints overcome this difficulty and verify the quasar periodicity signal for the 2dF data set.
There's a curious section that I'm not sure if I'm reading correctly where I'm not sure if it looks like they are determining what will be their primary data set post hoc? Maybe someone can clarify:
The constraint set produces fairly tight envelopes for all three fs data sets. The envelope for the fs = 0.9 data set is particularly tight all the way down to |zv|max = 0.010. We conclude that the increased robustness of that data set increases the number of unique quasar associations and reduces the number of false positive detections. We therefore consider the tests run against the fs = 0.9 data set to represent our primary results.
It appears that Arb acknowledges Tang's findings that the large sample shows null results, but came up with certain "constraints" that he feels necessary which produces results more closely in line with his hypothesis. He suggests certain other testing that should take place.
[/quote]
Notice also that Halton Arp was a sufficiently famous astronomer that a whole set of galaxies are named after him - Arp nnn - he isn't some sort of amateur!
I didn't suggest he was. Of course neither are his critics. It is not unusual for professional scientists to have different views that they champion. We shouldn't confuse being wrong with being incompetent. Scientists pursue all sorts of dead ends. That part of the process.