Arp objects, QSOs, Statistics

Dancing David said:
test test
Click to expand...

Now why don't you edit your original post by sticking what you posted to get these links working into the proper spot in that original post. Then folks won't have trouble reading what I wrote. And then you can try to actually address what I posted. Or not, as the case may be. :D
 
Oh, hello, Master of Snot, why have you answered the actual point of the thread, i am planning to respond to you, but please do go on and show your lack of critical thought, have you actually tried to address sampling bias and a posteriori statistics, there are ways to control for sampling bias.
 
http://www.internationalskeptics.com/forums/showpost.php?p=3563781&postcount=102
Arp hasn't claimed to be able to tell you the real distance to every object in the universe, RC. Only the mainstream has made that claim. Arp just maintains that some objects which you claim to know the distance to are not at those distances based on a wide range of observations and calculations of probabilities ... which you and David have specifically ignored or dismissed with nothing more than handwaving. And his theory (actually Narlikar's and Hoyle's) provides the theoretical basis for explaining why redshift of these objects might not correspond to distance ... but rather to the time since they were first created.
Click to expand...
Really? Tell us how dark matter explains that quasar that appears to be in front of a galaxy (NGC 7319). Tell us how dark matter explains the highly unlikely association of high and low redshift objects along that filament coming from NGC 7603. Tell us how dark matter explains the unlikely association of high and low redshift objects with respect to NGC 3628 and its features. Tell us how dark matter explains similar unlikely associations around GC 6217, NGC 470/474, NGC 3516, NGC 5985. Tell us how dark matter explains why in each of the cases the redshift of the high redshift objects near the low redshift object decreases as one moves away from the low redshift object. Tell us how dark matter explains the positional alignment of the various objects in the Local Group with respect to the major object in the group, M31? What's dark matter got to do with any of them?
Click to expand...
 
David, using insulting pet names for other members is a violation of the user agreement. I know because I got a warning for it.

I advise you to try and be civil.
 
Well, while this is off topic, I do believe that everybody does what they do because somehow they get something from it. What I don't understand, is why people keep doing something they don't enjoy. That makes no sense.
 
Poor Arp!

Dancing David,

In your long posts at the start of this thread (thanks, it was a lot of work bringing that stuff in, and a great way to begin this discussion), one of them was quoting a fairly detailed discussion about one of these groups of objects where there were some apparently statistically relevant "alignment" of QSO-like objects with he axis of a nearby bright galaxy.

While I tend to agree that Arp's statistics overall may be flawed, perhaps some of these objects have been ejected from Seyfert cores? Perhaps some of these observed redshifts are not cosmological?

I am not saying that the Big Bang didn't happen, or anything like that, but perhaps there are some weird things going in in the AGN's, and firing off some high velocity clumps of matter is not out of the question?

Aren't some of these objects in possession of strange spectra?

Stranger things have happened..............
 
QSO density around bright galaxies

In my copy of "The Arp Atlas of Peculiar Galaxies: A Chronicle and Observer's Guide", the whole first half of the book is devoted to Arp and his off-main ideas.

One thing that seems undeniable is the QSO density aroundn bright galaxies. It appears that many bright galaxies have in their neighborhood more QSO's than one would expect.

For example, the authors cite a paper by Burbidge, Burbidge, Arp and Zibetti, "QSOs Associated with M82", of which one conclusion about QSO density around M82 was:

QUOTE: "These correspond to densities of 30, 41 and 51 per square degree respectively. Such densities are to be compared with those obtained in QSO surveys by Kilkenny et al. (1997) and Boyle et al. (2000) which give respectively 10 per square degree to 20m, and 25 per square degree for 18.25 < bj < 20.85 from the 2dF survey with the Anglo Australian Telescope. While there are small uncertainties associated with the magnitude calibrations, and the total numbers are small, they do mean that there is a significant over-density of QSOs in the magnitude range down to 20m - 20m.5 near to
M82 compared with those in the general field."

The authors also make a good point when they discuss the sampling; sure, the limited sample sizes aren't ideal; but when do you cap sample sizes, and begin making inferences?
 
Hi Wangler. Welcome.

The issue I have is with the use of the statistics and I do plan to get back to each of the points.

I don't think Arp is terrible, I think he is very smart, I just disagree with his use of statistics. The basic premise runs like this:

- to say that a QSO has x probability of being in y area of the sky and that there is z chance that 4 objects would appear in y has intuitive appeal. It is not however a good use of population statistics. That would be making an argument that the members of the population would be evenly distributed through out the region, that they were spaced out on a regular basis.
Now assuming that this would be true, the stars that we could see would be allocated to the sky in a uniform sphere, if there was a ratio of x/y then for each section of y there would be x stars. So they would be distributed at the same regular intervals across the sky. So we would not have structure like constellation and other visible structures, so while you might determine a sample density per unit, it would be an error to construe the likelihood of a particular probability of a certain arrangement from the average population density.
Another analogy (and these are always full with peril) would be to look at the average populations density of human being on the landscape, these statistics are very common. But if you then found a city you could argue the low probability of the city occurring, if you assumed that all the human being were evenly spaced out, say that our city is a kilometer square and that the density if (.1 person/sq. km.) so if we had 10 people in our city you could falsely say that there was a (.1)^10 chance of this happening.
 
SDSS's quasar survey won't be 100% complete but a quick dig around suggests 95% completeness to the limiting magnitude. Of course translating that to a space density of qsos is a nontrivial task, but you might not want to do that. I'm not sure what Dancing David is planning to do with the numbers.

Wrangler - the 2dFGRS focuses on galaxies, so you'd not expect to see many quasars. 2QZ - the 2dF quasar survey - obviously gets lots of quasars and few galaxies! Check them out at 2dfquasar.org.
 
I am saying that argument that you can determine the probability of 5 QSOs in a certain area from the density ratio is not sound reasoning.(I am arguing against the population density as an indicator of significance)
That is not how population sampling is normally done, if you have a population you consider to be off the scale in significance then you compare it to other samples that are random (hopefully) and see what the difference is.
So in this case and others there is a claim made that the association of QSOs is abnormaly high. It is not accurate to do this from density ratio, which is what most of the studies are essentially doing.
The best method would be to choose one hundred 'normal galaxies', one hundred AGN galaxies and one hundred random points on the sky, then you compare the sample of QSOs in those three hundred sample areas and compare it to the one Arp and others are saying are abnormaly high.
That is sound population statistics, except that you also have to control for the lensing by massive bodies and how that will magnify distant objects.
 
Last edited:
Statistics?

Dancing David,

thanks for the welcome!

My statistics knowledge is less than I would like but your comments:


Dancing David said:
- to say that a QSO has x probability of being in y area of the sky and that there is z chance that 4 objects would appear in y has intuitive appeal. It is not however a good use of population statistics. That would be making an argument that the members of the population would be evenly distributed through out the region, that they were spaced out on a regular basis.
Now assuming that this would be true, the stars that we could see would be allocated to the sky in a uniform sphere, if there was a ratio of x/y then for each section of y there would be x stars. So they would be distributed at the same regular intervals across the sky. So we would not have structure like constellation and other visible structures, so while you might determine a sample density per unit, it would be an error to construe the likelihood of a particular probability of a certain arrangement from the average population density.
Another analogy (and these are always full with peril) would be to look at the average populations density of human being on the landscape, these statistics are very common. But if you then found a city you could argue the low probability of the city occurring, if you assumed that all the human being were evenly spaced out, say that our city is a kilometer square and that the density if (.1 person/sq. km.) so if we had 10 people in our city you could falsely say that there was a (.1)^10 chance of this happening.
Click to expand...

I have a hard time making sense of. It seems like you are stating that Arp has incorrectly assumed that QSOs and galaxies are evenly distributed; or he has incorrectly assumed that QSOs and galaxies are a "general" population.

If this is not the case, than how can we use statistical processes when discussing these populations at all? Does that mean we cannot make inferences about the galaxy population?

Or is it that Arp is making the wrong inferences. Aren't his claims similar to the following claim:

For stars of magnitudes 6-9, we find an approximately even distribution throughout the sky. However, we see a formation, Collinder 399, that looks curiously like a coathanger, a quite unnatural form. It would seem that the chances of this being a random occurrence is markedly low. Sure enough, there are few, if any other associations of this shape in the sky. And, as it turns out, this "random" arrangement is due to a close association of the brightest stars; i.e., they all formed at about the same time, in the same general region of space. Turns out they are a physically associated cluster.

To me, this line of reasoning is sound, but like I said my statistics are very rusty.
 
Statistics?

Dancing David,

thanks for the welcome!

My statistics knowledge is less than I would like but your comments:


Dancing David said:
- to say that a QSO has x probability of being in y area of the sky and that there is z chance that 4 objects would appear in y has intuitive appeal. It is not however a good use of population statistics. That would be making an argument that the members of the population would be evenly distributed through out the region, that they were spaced out on a regular basis.
Now assuming that this would be true, the stars that we could see would be allocated to the sky in a uniform sphere, if there was a ratio of x/y then for each section of y there would be x stars. So they would be distributed at the same regular intervals across the sky. So we would not have structure like constellation and other visible structures, so while you might determine a sample density per unit, it would be an error to construe the likelihood of a particular probability of a certain arrangement from the average population density.
Another analogy (and these are always full with peril) would be to look at the average populations density of human being on the landscape, these statistics are very common. But if you then found a city you could argue the low probability of the city occurring, if you assumed that all the human being were evenly spaced out, say that our city is a kilometer square and that the density if (.1 person/sq. km.) so if we had 10 people in our city you could falsely say that there was a (.1)^10 chance of this happening.
Click to expand...

I have a hard time making sense of. It seems like you are stating that Arp has incorrectly assumed that QSOs and galaxies are evenly distributed; or he has incorrectly assumed that QSOs and galaxies are a "general" population.

If this is not the case, than how can we use statistical processes when discussing these populations at all? Does that mean we cannot make inferences about the galaxy population?

Or is it that Arp is making the wrong inferences. Aren't his claims similar to the following claim:

For stars of magnitudes 6-9, we find an approximately even distribution throughout the sky. However, we see a formation, Collinder 399, that looks curiously like a coathanger, a quite unnatural form. It would seem that the chances of this being a random occurrence is markedly low. Sure enough, there are few, if any other associations of this shape in the sky. And, as it turns out, this "random" arrangement is due to a close association of the brightest stars; i.e., they all formed at about the same time, in the same general region of space. Turns out they are a physically associated cluster.

To me, this line of reasoning is sound, but like I said my statistics are very rusty.

Basically, I think that the problem with your criticism of Arp's statistics is evident in your example regarding human population density: the odds of a high people density in a particular location would be astronomical, if the density was a random occurrance, the high density is not random, the people are part of a city or association, and that is why the density is high.

That is exactly what Arp is trying to say: the probability of X number of QSO's being around a bright galaxy is astronomical, if this was truly a random occurance. Perhaps the occurance is not random; perhaps there is an association?

The odds of their being hundreds of kit-built planes at one time in a Oskosh Wisconsin are astronomical..unless their is an association! There is, believe it or not! No flawed statistics here!
 
The space aroung galaxies

Dancing David said:
That is sound population statistics, except that you also have to control for the lensing by massive bodies and how that will magnify distant objects.
Click to expand...

IMHO, Arp sees an apparently greater incidence of QSOs around bright galaxies because these galaxies have gravitationally scrubbed much of the intergalactic gas and dust from their environments, permitting a clearer view of objects further away (=with higher redshifts).
 
Wangler said:
Here is a link to the Wikipedia article on the 2dF survey:

http://en.wikipedia.org/wiki/2dF_Galaxy_Redshift_Survey

Apparently, it covered 1500 square degrees, and includes spectra of 232,155 galaxies, but only 125 QSOs!

Apparently finding any number of QSOs around a bright galaxy is a relative long shot!
Click to expand...
.

Perhaps that's because it is the 2dF galaxy survey?

There was, at the same time, using the same telescope etc, a 2QZ (2dF QSO Redshift Survey), which got spectra for > 25,000 QSOs.

However, SDSS is a much more extensive survey, of both galaxies and quasars.

Somewhere in the BAUT thread on Arp Dancing David cited in another, recent, JREF thread there's some discussion of a landmark paper - by Ryan Scranton (no doubt et al.) - on how the distribution of SDSS quasars around galaxies matches what is expected from LCDM models (a.k.a. 'big bang') with galaxies having dark matter halos as well as lots of ordinary mass and lensing distant quasars accordingly. This seems to be an SDSS page on that finding; from there it should be pretty easy to find the ApJ (or whatever) paper, or at least the arXiv preprint.

Sorry for the diversion ... :o

ETA: sorry edd, I'd not read down far enough in the thread to see that you'd already answered Wrangler :o :o
 
Last edited:
Dancing David said:
I believe the question was, what evidence is there for anamolies in the theory of cosmological redshift?
Click to expand...
.

Apologies in advance; I haven't yet gone through all the details of the introductory posts ...

Surely the question is much narrower than this? Aren't you asking about just Arp's papers, quasars (or QSOs) and galaxies, and how robust his statistical analyses are?
 
You must log in or register to reply here.

Back
Top Bottom