Talk:Impact factor

From Wikipedia, the free encyclopedia

Contents 1 Copyright 2 Title caps 2.1 Result 3 Skewness 4 objectivity and coverage 5 Link to selected IFs online 6 edited first half for clarity, specificity and NPOV 7 reply to comments 7.1 what is important, and what can be changed 8 Inclusion of Link for Manually Tabulated Impact Factors 9 Illegal site? 10 apology

[edit] Copyright

ISI impact factor list is currently not publicly available. It has to be subscribed. (I don't know whether it was publicly available in the past.) -- Zeyar Aung

It was never free. It is copyrighted by ISI - Thomson Scientific. If you are interested in "alternative" sources, see this. Karol 19:52, 15 December 2005 (UTC)

I propose that links to both the Journal Impact Factor mentioned by Karol and the Hirsch number be added to this article in a section called "alternative impact metrics." Right now the article is something of an advertisement for ISI. I'm not going to just go ahead and make the change because I appreciate that impact factor is already a well-crafted article that someone has taken pains over. Alison Chaiken 01:23, 28 December 2005 (UTC)

This is no means an advertisement for ISI: it is a straightforward description of one part of an extremely import bibliographic tool, available at all major uuniversities (for a very high price to the university library) . It's a world standard (for better or worse). Of course there are alternatives, and they should be included.

The page on Se 9, 2006 has indirect links to sources of illegal content, and these links are being removed as copyright violations in accordance with Wiki Policy. DGG 03:38, 10 September 2006 (UTC)

[edit] Title caps

Should this be Impact factor ? - FrancisTyers 14:34, 15 December 2005 (UTC)

I think it should. Note that Impact factor redirects to here. On the other hand, it is notoriously used with both words capitalized, as a proper name. I think ISI - Thomson Scientific is responsible for this :) Karol 19:45, 15 December 2005 (UTC)

Of course it should. No question. --Stemonitis 09:30, 19 December 2005 (UTC)

Can't say I agree. It has a name, and its name should be used as it is by the creator. With both capitals. But as I seem to be in a minority I won't change it. DGG 03:38, 10 September 2006 (UTC)

[edit] Result

Page moved as requested. WhiteNight ^{T | @ | C} 17:09, 29 December 2005 (UTC)

[edit] Skewness

I would have thought that the number of citations followed a Poisson distribution. Jeremymiles 16:23, 21 February 2006 (UTC)

I think they might follow the Zipf bibliometric distribution -- 80% of the refs are generally to 20% of the articles

[edit] objectivity and coverage

Nevertheless I do agree on the advertisement like of the article. In particular, I would appreciate somebody has time and expertise to refresh my souvenirs, and complete information on ISI's “wide international coverage” and the protocole of selection of this coverage and the consequences of it (national researches tend to be disgarded). The assertion “An objective measure of the importance of different publications” has to be put, in my opinion, in relation with the absence of explanations on where to count the citation numbers in the phrase “The impact factor […]to be the average number of times published papers are cited up […]”. The word “objective” then should rather be “relative according to the journal coverage of the database”. I would like also to add this reference in external links ("du bon usage du facteur d'impact" -on the good usage of impact factor) :http://www.udfapse.lib.ulg.ac.be/aide_publication/facteur%20d'impact.htm

Will add.DGG 00:50, 15 September 2006 (UTC)

[edit] Link to selected IFs online

Re: Impact factors of selected scientific journals in 2001

I understood that IF listings were copyright, and that Thomson have patrolled this vigorously since they took them over; should we link to this? Espresso Addict 22:35, 28 August 2006 (UTC)

The link to this data must be removed. It links to an essentially complete copy of the 2001 data in JCR. I would guess that this year was chosen because it is the year the ISI makes available at no charge to library schools, in order to familiarize their students with the system. Conceivable the people posting it were under the mistaken impression that it was being made generally available.

I shall edit to remove the link, to anticipate my friends at Thomson/ISI. There is another link to copyrighted ISI data on this Wiki page (the external link to "List of impact factors from recent years--which is a linl to a site that says "The lists of Impact Factors have been removed from this site due to copy right violations. Below are the lists available elsewhere." [and then proceeds to give links to other illegal copies]"

i shall remove it as well.   (and check Wiki's related pages to see if there are any more)  

--I am following the instructions on http://en.wikipedia.org/wiki/Wikipedia:Copyrights

The Wiki "Get Cited" page links to an independent database apparently compiled by asking people to join and enter the reference lists from their papers. It appears to be legitimate, and I will edit its page to describe it in more detail.

DGG 03:24, 10 September 2006 (UTC)

[edit] edited first half for clarity, specificity and NPOV

To me, the article pre-edit exhibited a POV deprecating the use of impact factors. This is one of the possibly reasonable positions, but it is not the majority position. As shown by the wide availability and universal use of these IFs, the majority position is that they are valid, although other measures should also be considered. There are, naturally, more recent references proposing other measures than defending the existing ones, but what else could be expected? One cannot get a paper published saying that h values are unreasonable, use IFs instead. I could fish out comments from the various lists, but so could the reader. I'm not including the comments from various lists that those seeking other metrics are those whose journals don't show up well. I could correlate if I wanted to, but this isn't a product review.

The second half has most of the alternatives, and I promise not to delete any of them.

I've used IF for convenience, but I'll change them. to match ISI practice. I ask again about capitatilization for Impact Factor, since capitals are the consistent ISI use. I'll follow the apparent policy here.

As in the previous article, I didn't link to ISI every time mentioned, and I didn't link to subject fields used as examples. I could bold them, but whom does linking or bolding help?- I have not yet checked that all ISI links are reachable without subscription--I will, but i need to use a different computer.

If someone disagrees with me on the NPOV, I wish they'd comment here before reverting or changing drastically. DGG 01:14, 14 September 2006 (UTC)

My comments are mostly going to be about some things that I disagree with, but don't let that discourage you. I'm taking the time to write this because of your promising start. I didn't fix any of the things I mention because you may want to do that yourself.

• Use bold when you use the article title (and any redirects pointing here) for the first time (preferably in the first sentence or paragraph). Not everyone agrees on when to set links, but most terms should only be linked once per article (or per section, in some cases). Linking ISI once, as you said, is perfect; I'd link subject fields used as examples, too. Having a look at some featured articles and at the MoS may help.
• Try never to make things worse than they were before. Your edits removed the interwiki links (to the article in other languages) and added a category a second time. There's a half-finished sentence now ("Most of these factors are thoughly didiscussed on the"), new spelling mistakes (Citaation, posiible), and a sentence ending in two periods.
• The intro was better before because the first sentence explained what the article is about – now it says who invented and owns it.
• This sentence needs explanation for readers not versed in the subject at hand: "There can be no Impact Factor for an article, because the denominator is always 1"

• "The best explanation of the details and the proper use are at the [Thomson website]" is bad style. What we can or need to say belongs into the article body. If we want to suggest further reading, then that goes into "External links". On the other hand, if you want to point to an external source you used as a reference, use <ref> (there are examples in the article).
• Capitalization would probably best follow what 3rd party publications on the subject usually do — the creators/owners of such terms often have some very odd ideas of how to make their special term even more special.
• As for POV on IFs: IFs do measure something related to some quality of journals. The use of IFs in the evaluation of producers (institutes, teams, researchers), however, has been an unmitigated disaster – the serials crisis is the least of the problems it helped create. Even Garfield himself calls that an abuse. And there's little controversy about this because it's just too obvious:.
• It is certainly true that journals with a low impact factor have an incentive to question the validity of IFs. But the criticism quoted in the article was made by Nature which as you know has a huge IF.

Rl 08:04, 14 September 2006 (UTC)

[edit] reply to comments

In general I do concede that a writer must follow the house style. Therefore:

• I apologize for the typos and thank you for the corrections.
• The interwiki links were an accidental deletion. I will restore them from previous versions.
• I've rechecked, and 1/4 of the authoritative souces use both caps, 1/4 first only, and 1/2 both lc. In titles, all use both caps. (and, to my surprise, ISI uses l.c. in its help pages) I will adjust.
• I still disagree with links to subject fields on grounds of readability, and the examples show exactly what i dislike. I think it's valid only if the target article has a relevant part, and then it should be a link to that specific section or anchor.
  • But since other people will probably add them if I don't it might as well be me. At least in WP it's easy.
• For the opening sentence, I will revise to include the definition if possible. (you're right, at least in terms of WP style--but it is very hard to define IF accurately in one understandable sentence.)
• As for the way I did the external links, I modeled on what I saw on other articles. I'll look at the help again, for there seem to be a number of different options there.
• I'll say more about the equation, and if you think my explantation inadequate tell me. This is the sort of thing one cannot well judge by oneself. It's for the reader to judge whether one is clear, saying "it would have been clear if you had read more carefully," is never right.

For more substantative matters:

• If I did not make it clear that there is significant criticism on the use to evaluate institutions & individuals, I'll make it clearer. There is however a distinction between relatively appropriate ways of using it, and totally ludicrous ones, which I will expand on.
• As for the effect on the journal crisis, i refer to the fallacy of libraries always buying the highest IF journals. I don't think it is a major effect, for they generally judge otherwise. I see no other effect, so if you do, please tell me so I can think about it.

For more general WP policy

• As for the use of the Thomson website, I will try to expand what I do say instead of leaving it all to them But the Thomson help pages are the best explanation possible, and much too extensive to include here. This is not a public domain source, and too much would be necessary to be fair use. I'm sure they'd permit me to insert it here. but it wouldn't be GFDL. I do intend to discuss the general issue at some more appropriate place, because if the WP policy is what I think it is, I think it should be reconsidered. It might have been appropriate as a way to get started. And I can see it might be unavoidable for certain descriptive articles. If its a matter of stable links, there's other ways for that.

I do not expect to make all these improvements tonight. :) DGG 00:50, 15 September 2006 (UTC)

I have now made most of the corrections, expanded some of the later paragraphs,and added the links to other wikis. DGG 02:42, 15 September 2006 (UTC)

• New typos, literal quote mangled, one external link to pubmedcentral that used to work broken, repeated sentences in the intro

I fixed all that but I suggest to use "Show preview" and "Show changes" buttons liberally before saving a page, or to check the differences between revisions in the history tab

• "impact factor" casing not fully consistent yet
• I advise against adding more articles to external links; we should quote relevant facts in the text and then add the articles as references
• It is of course not possible to "define IF accurately in one understandable sentence". But that's not the point. The first sentence just offers a rough description aimed at people who may a) have no clue what the term means or b) are wondering if they arrived at the right place. It is an attempt at capturing the essence, accurate definitions come later.
• The article as is contains a lot of valuable information but in a barely organized way. What it needs most is a clear structure (e.g. Overview, History, Use/Impact/Criticism, Alternatives).
• regarding the "more substantative matters" – criticism and serials crisis: I don't want to add more to the article as it is now. It would just make the necessary refactoring of the content harder.
• Fair use is not an issue. We can publish the relavant facts in our own words.

Rl 07:55, 15 September 2006 (UTC)

[edit] what is important, and what can be changed

Dear RI, Though Im glad for the proofreading--you may not believe it, but I do transfer the content into MS word for a spell check first-- (I use MS word because Ive built up a good dictionary there of what I use) Besides yourself, there seem to be many proofreaders looking around, a practice I approve of. I too can even do this if I havent written the text. Some of what you advise are style changes I can make, but for some, you may be misunderstanding what I am doing.

1. I will indeed describe it in one sentence, unreadable though it may turn out to be. You can then give it a try. I notice a great many leads have extremely long and convoluted sentences. I know how to do that too. 2. Neither i nor anyone else can rephrase the help pages of JCR to be as good as the originals. As I understand it, that is one of the purposes for outside links-- detailed explanations that others have done better. If it is considered better style here, as you suggest, I will put in summary sentences of my own, and then use the outside material as references , as i can find subscription free material. (There's a good deal of other non-subscription writings by Garfield on the web). I know it when someone is clearer than I am, Garfield for example. BTW, by my standards, a detailed paraphrase is plagiarism. Plagiarism is borrowing expressions and arrangement --you cannot make a writing of your own from someone else's work. (it is much stricter than fair use in copyright). I do see signs that some do it here, and I'm keeping examples for when I'm experienced enough to be taken seriously here. I would be taken very seriously already if I put them on liblicense, but not even I would go outside first. I don;t think anyone's reported on this outside WP yet, but I'd rather influence than criticise. 2A When I started, the external links were all to subscription sources--I am still changing this.

3. Yes, its disorganized. It's disorganized because I have not yet done the later sections--I am very reluctant to alter all of this at once, because it has clearly been inserted by some whom I consider distinct small minorities. (BTW, for serious dscussion of the alternate measures, see the SIGMETRICS list, and JASIST.) If I have to prove they're minorities I will, but all I'm doing is reducing their share from 75% to 25%, which is probably 2X what their arguments deserve. If you are suggesting that it may be better to go point by point than divide, you msy be right. ZI did see the overlap as i ws writing. 3. There is no history. JCR was created, and has remained the same except for the better readability as an e-resource. 4. The substantative issues are my removing material, not adding it. The entire skewness section is wrong. the inflence on libraries is wrong. the h calculations i think arewrong, but thats the one part whee there is someserious owrk to be accounted for. Someone mentioned in the talk the angloamerican centrism, and this does need expansion. 5. I am aware that i can write or revise a complex page in a subpage of my space. But I always like to go by stages, and I want feedback as I go. I will try that, and i may even try writing in msword. I have my own good macros for the html, and I could probably write some for the wiki style. But the next version of MS Word has a more complicated scripting language which I have yet to learn. 5a Equally important, for some of this, if I make too many changes at once, I suspect that people whose ideas I discarded would revert. I am willing to be corrected, but not for my work to be futile. 6. If things i truly care about like 2. are not the custom here, I don't belong here. If articles like I found this are considered NPOV, ditto.

7. BTW, the real problem I have here is the necessary changes in interconnected articles. I may try to figure out a way to do this offline. But if it isnt fast enough it throws out other peoples edits. DGG 06:50, 16 September 2006 (UTC)

Most of your arguments and plans sound reasonable – it's good to have you here. Some of your assertions though are likely to be opposed by some. I advise you not to assume that the current article reflects the extent of our knowledge – it represents what has been written given several limitations like time, other responsibilites, and knowledge. (for instance, doing their own proofreading and fixing the mistakes for others makes some people slower)

The fact that there are many lousy leads doesn't mean that complicated intros are a good thing.

External links are for content that either doesn't belong into an encyclopedia or that we didn't get around to covering ourselves yet. But the article must be complete without the content of external sources.

A "detailed paraphrase" may be a copyvio. Plagiarism, however, is passing off someone else's work as your own, something you don't do if you give your source. If you present the facts of a text in your own words and name the source of your facts, then it's neither a copyvio nor plagiarism. I am aware that there are those who try to extend copyright way beyond that – but so far, we don't need a license just to be a secondary source on anything.

Please specify exactly what statements you disagree with.

The disorganized state of the article is of course not your fault, it's just something that needs to be fixed.

What I meant with history: the vision Garfield had when creating the impact factor, initial public reaction, increase in the number of journals covered, etc. But I don't insist on having a history section, we just need something better than what we have now.

The entire skewness section is most certainly not wrong. I remember reading the quote in Nature (I can verify the quote, but it certainly catches the spirit of what I read), so that's a seemingly relevant source accurately reported. The following paragraph does have a quirk or two, but looks mostly correct as well. Can you say specifically what statements you object to, or suggest a better wording?

I recommend either re-shuffling the content (without losing any) or fixing the statements you disagree with. Mixing both kinds of edits together can be quite confusing. You may want to take a look at {{inuse}} if you plan to do a major surgery, it may come in handy.

Rl 21:29, 17 September 2006 (UTC)

Thanks for info on the template; I was hoping there might be such a device.

Eugene Garfield did so many related things in bibliometrics that his article needs expansion--maybe we can best do some of the history there, but I do know what you mean.

The skewdness is not wrong, precisely, but rather just the manifestation of a more general phonomenon, Bradford's law. There are so many ways of presenting

this & so many difficulties that I havent even begun to work on this.-::enough talk--I'll try to work on the article instea

[edit] Inclusion of Link for Manually Tabulated Impact Factors

Hi, I am a Harvard researcher who manually tabulates new ISI rankings by journals of different scientific discipline. My tabulation of ISI impact ranking list is unique-- it breaks it down by scientific disciplines: general science, medicine, diabetes, CVD, cancer, obesity/nutrition, epidemiology and public health, etc.

This information took a lot of hard work and time to tabulate, and thus, a unique contribution for interested readers and scientific community, which was I included it on a weblog. Also because I tabulated these numbers from each scientific journal myself (from their general info "about" pages), it is not violating ISI copyright.

After a previous discussion with another Wiki user (R1), he agrees that given the nature of the unique tabulation of data (not available anywhere else), it is fair to provide 1 link on the impact factor list. Let me know if others have feedback- if not, perhaps R1 can add back the links, so as to maintain neutrality. —The preceding unsigned comment was added by Epiding (talk • contribs) 04:24, 19 September 2006.

Note: I removed the quotation from the talk page that broke formatting in several interesting ways. The comment referred to above is this one. Rl 08:06, 19 September 2006 (UTC)

I'll admit that said comment was not as clear as it could have been – my main point was that if the blog link belongs anywhere, it belongs into this article, and nowhere else. External links should be added sparingly (references are a different matter), links to blogs are often problematic and external sites that are ad-infested is where the line to "nothing is better" is usually crossed. – That said, the limits of copyright protection for ISI's impact factors are untested (to me, anyhow; anyone know precedents?), and ISI would have a vested interest in exaggerating the scope, so I don't object to the link (which has only a few impact numbers out of thousands) on copyright grounds. Rl 08:06, 19 September 2006 (UTC)

Look at how much is published in the literature--very little is--we want to set a good example, not fight with ISI. Let me see if I can reach an agreement with ISI. I think I know the right people. DGG 08:34, 19 September 2006 (UTC)

Excellent. Maybe you can get them to clarify a few things:

• Is it okay for us to note current impact factors in the articles on specific journals (see Nature, 3rd paragraph, for an example) ? How about previous impact factors?
• Is it okay for us to list the most highly rated journals sorted by area, as Epiding did? With or without their impact factors?

Please note that answers need to be verifiable in order to be meaningful (for instance, if you get an email, you may add – with their permission, of course – their reply to this talk page, along with all the email headers). Rl 09:05, 19 September 2006 (UTC)

i am planning to ask first for something which will serve everyone in and out of WP, which is for ISI to make one sample year available OA. which is a very basic qy, and will take personal contact. For the specific instances of current or historical data, if asked directly, any legal dept will usually ask anyone who wishes to post to obtain permission each time, which would not be GFDL, and not imho a useful technique for our purposes, for Epiding would need to ask himself about his own blog. I know WP moves quickly, but ISI does not. And see fair use, a very good article. DGG 03:48, 20 September 2006 (UTC)

OA is not a license and not necessarily compatible with the GFDL. Rl 06:34, 20 September 2006 (UTC)

I understand the difference perfectly well, Rl, and said what I meant to say. OA from ISI will be legal, and we can link to it. I know it is not sufficient to insert.

WP needs GFDL but someone's blog doesn't--it just has to be an authorized use. DGG 04:26, 21 September 2006 (UTC)

[edit] Illegal site?

The link below has been removed from the article because it is an "illegal site" – according to the edit summary:

• N. Pinhas & C Kordon "Du bon usage du facteur d'impact -on the good usage of impact factor"

I'm not sure how helpful a French article on impact factors is to readers, so I'm not advocating its inclusion. However, I challenge anyone to make a coherent argument that the linked site or article are illegal. Rl 08:18, 4 November 2006 (UTC)

Sure. The site had impact factors on its site. It removed them, and says:

"The lists of Impact Factors have been removed from this site due to copy right violations. Below are the lists available elsewhere." So the site admits that posting list of impact factors reproducing identical or to a considerable extent such factors is a copyright violation. To then link to some other places that have them is ingenious, and equally wrong. "I will not sell you heroin, but I know several people who will be glad to: here are their names and phone numbers."

Fortunately, when dealing with electonic resources, anyone can produce a citation indexand derive impact fcors from scratch, I encourage those who support this approach--which I support as well--to become active in one of the Open Access Citation Index workgroups--in fact I am, personally.

But I have respect for your position, and I do not immediately remove, pending further discussion--and because there would be no way to avoid a revert war if we continued that way. DGG 15:31, 9 November 2006 (UTC)

I must be missing something here. I was referring to a single web page titled Du bon usage du facteur d'impact, linked above. It contains exactly one link – to ISI. It doesn't contain any mention of copyright, either. You must be talking about a different page, then (link, please?). – And the comparison of what would constitute a relatively minor copyvio with selling heroin doesn't seem appropriate to me. Rl 18:39, 9 November 2006 (UTC)

[edit] apology

there does indeed seem to have been a mix-up, & I seem to have deleted the wrong page. I have restored it. The problem page is http://www.sciencegateway.org/impact/. which is a page providing a number of links to copyright violating lists of impact factors. I left it in, and labelled it for what it is. Yes, heroin is an exaggeration. But it is, quite literally, grand theft. These lists cost over $10,000 per year.DGG 05:01, 10 November 2006 (UTC) There is nothing wrong, with the rather good French page.

Oh, I see. I removed the sciencegateway page. It looks like a site thrown together to make advertising money with a high pagerank, and there's no need for us to support that. – However, what the site does is not "grand theft" at all: a) copyright infringement is not theft by law, and b) very few people who'd pay ISI $10,000 for such a list would rely on a random list on the internet for whatever important use case they have, so the damage to Thompson should be pretty much zero anyhow. Rl 07:01, 10 November 2006 (UTC)