I have not read the book; from the publisher's blurb and comments such as Ian's (and your own), I expect that if I did I'd get extremely angry.

To remind readers of something I wrote:Here's the distinction jaydee: in writing a book, you can say anything you like; in publishing a book, the publisher has no obligation to check anything the author says (except where it might be break laws on copyright, libel, etc; I note that in the US the laws on libel allow an author to say almost anything without fear of being prosecuted, let alone convicted). In getting a paper published in a peer-reviewed journal, you need to meet a much more rigourous set of standards, on 'trivial' things (formatting, layout of tables, references, and so on), and on not so trivial things (e.g. consistency, correct attribution, logic, sufficient detail, and so on).

And before I forget, an important distinction which hasn't been all that clear in our discussion so far: it's not like Arp is an entirely lonely figure; if you check the literature, you'll see that he often (frequently?) is but one of several authors (and sometimes not the lead author?); you will also find many, many papers with conclusions apparently quite consistent with Arp's views, written by people other than Arp. It might actually be interesting to do an analysis - how clearly can one demarcate an 'Arp camp' and an 'anti-Arp camp'?

Now, for the 'sloppy science'. Let's start with a backgrounder; Bill Keel's is, IMHO, a good one. He's spent much of his professional career studying interacting galaxies, quasars, etc (i.e. much the same field as Arp), and I think you'll find his summary quite non-partisan.

My main 'beef' with Arp et al. is in how he turns the 2D sky that is all that we can 'see' into 3D pictures (AFAIK, parallax is the only 'sure' method of getting a 3D picture, and that is presently limited to a depth of no more than a thousand pc or so) - firstly, how do you distinguish between what's 'foreground' and what's 'background', and second, how do you estimate at least the relative distances between 'foreground' and 'background' (if you take the trouble to read the posts in this thread, you'll get an idea of what this is about).

The 'classic' paper in this regard, for me (for historical reasons), was not even written by Arp! It's López-Corredoira and Gutiérrez - "Two emission line objects with z > 0.2 in the optical filament apparently connecting the Seyfert galaxy NGC 7603 to its companion". I have no particular problem with the observations - they seem to have been competently made and the data reduced using standard methods. However, it's what they make of these observations that I found astonishing (caveat: this is a preprint; I do not know how different the actual published paper is). Leaving aside such hyperbole as "A knot (object 2) is perfectly centered in the line of the filament and positioned where the filament connects to NGC7603B" ('perfectly'? in a paper reporting an astronomical observation??), the content part (i.e. NOT in 'Discussion and conclusions') of the paper ends with these words: "That is, there should be one object like these per each square of 3-7 arcminute size (20 arcminute size for NGC 7603B); much larger than the area of the filament (~100 arcsec2).". In other words, they have taken the 'sky' average distribution of certain classes of objects (quasars/Seyfert1s; HII galaxies) and applied it to this one particular case! No mention of the 'sky' dispersion (more or less 'we know that these objects are not distributed uniformly on the sky; a measure of this non-uniformity is {x}; applied to this case, we would expect that there would be between {a} and {b} (1 sigma)'), no caveat about the pitfalls of this kind of statistical analysis, nothing.

I'd be happy to walk you through this 2D/3D thing jaydee (I think I saw - elsewhere - that you mentioned that you don't have formal training in this field), if you are interested.