Tuesday, June 29, 2010

Seeking Perfection

Is your citation record in Web of Science (or the moral equivalent) perfectly correct? Or are there errors?

If there are errors, are they insignificant (not worth correcting) or significant?

If there are significant errors, have you done anything about it? (Or will you?) It is possible to request a data correction using a form provided on the Web of Science website.

There is one particular paper of mine that is particularly prone to being cited in various and sundry ways. In fact, the citations for this paper are strewn about in so many different apparent titles in my citation report that, were the errors to be fixed and the citations combined, my h-index would increase (gasp). There are other errors as well, mostly because authors citing my work used an incorrect volume, page, or year, but most of these errors do not affect my h-index.

Should I try to fix the errors, or, at least, the one that would affect my h-index? Would you, if you were (or are) in this same situation?

29 comments:

Female Science Professor said...

That's the question.

John Vidale said...

If there were a way, people with big shortcomings might want to fix them. I don't know if the distinct author sets can be corrected, which can be the only way to track people with common names.

I see no moral problem with fixing errors, but it would be rare for any result to be worth the effort.

My Web o' Knowledge entry looks correct. Except 13 fewer papers show up under my distinct author set than just with searching for my name, with 55 fewer citations, and the same h-index. So no, I won't fix anything.

David Stern said...

I requested that ISI fix an error which attributed about 15 citations of mine to a paper in a completely different discipline with the author name Starrett instead of Stern. They fixed that one. The authors had all cited me correctly. As far as the other variation in ISI goes we just have to live with that (it's better than Google Scholar by an order of magnitude :)). If you care about the h-index, compute the h-index correctly yourself and post it on your website.

Anonymous said...

Yes. I was listed as several people and asked them to consolidate the several me's into one and they did it within 24 hrs. I know other people who have done the same.

bob said...

I asked for a correction once. It was very straightforward and easy to do but it took them about two months to fix (they're upfront on the request form about the timescale).

I would go ahead and request the corrections. The database might as well be correct and as Wikipedia has taught us, small improvements can really add up.

Anonymous said...

Wow...I thought I was pretty neurotic about such things, but I have never thought to check this. I'm curious to hear whether anyone actually has...

Anonymous said...

I definitely care about it. I'm an FS-associate prof, with a slightly unusual name (multiple initials). PubMed can deal with names that have the initials A. B. C. but mine is A. B. X. C. So sometimes my papers come out with C. as the surname and sometimes with X. But no one can ever find the ones with X. and they don't contribute to my citation record. I've repeatedly filled in the PubMed correction form, and they do get corrected eventually, but the corrections don't filter through. Of course, I could drop the X. part of my name, but it is my name and I like it!

Scopus has a better system where you can link all the variants of your name together so they all contribute to your citation record, and my citations are definitely higher on Scopus than on WoS. I haven't tried correcting the WoS version yet because I haven't had time.

The current system definitely disadvantages people with non-traditional names, and I think a better author-identification system would be very helpful. The more metric become important, the more it matters that they are accurate.

GMP said...

I have requested changes in the past when I've caught them. Most were about incorrect citations of my own papers. ISI are relatively prompt about corrections (takes a couple of weeks to implement them).

About growing up: people will go look up your h-index all the time (I know I do and my colleagues do) and they will make judgments based on it, so there is nothing wrong in ensuring that it is correct. It's not like you are taking steroids.

Post-doc said...

People who make decisions about your career may care. Seems to me like it would depend on the stage in your career and whether that h-index increase would help your advancement.

Karen said...

Fix them. Someone out there is collected stats to analyze about male vs female publication records.

a physicist said...

One of my articles was published in Phys Rev Lett and has article number 031416. If ISI counted mis-citations to article number 31416 as the same as 031416, I'd have 40% more citations for that paper.

But no, I don't lose any sleep over it. It's a useful reminder about how noisy the citation data are.

a physicist said...

I should add: the above-noted problem with my article in PRL doesn't affect my h-index. I do have another publication that is often mis-cited and it does affect my h-index. I don't lose any sleep over that one, either, and haven't tried at all to correct it.

Anonymous said...

"If there are significant errors, have you done anything about it? (Or will you?) It is possible to request a data correction using a form provided on the Web of Science website."

I have done this (mostly to sort my refs to me and the newly appeared other people with my combination of last name and initials to them, though I have in the past tried other things) and it has been totally ineffective. I think this form is simply a ploy by ISI to pretend they attempt to be accurate, when in fact they are wildly inaccurate. Mike Rossner wrote a nice piece about their disregard to accuracy in the Journal of Cell biology--well worth a read.

Mark P

Anonymous said...

Before I went up for tenure two years ago I had asked Thomson to fix one of my papers. It was a highly cited paper, and I wasn't even the first (of many) authors. But when it came to tenure, it seemed prudent to make sure that the citations were mostly correct. That said, it took about a year before Thomson even responded and sort-of fixed it.

So, one can report errors (I think I did it through a normal web site contact form) and they will probably consider reasonable requests, but it probably won't be quick.

Mr. Post-doc said...

Comrade Physio and FSP may have less reason to worry about this being more established in their career. However, FSP earlier this year (I believe), had a post about "uncited publications... and are they worthwhile?" (paraphrasing from memory here) indicating that h-factor aside, citations are one way we get measured. I'm guessing the shorter the CV the more these things matter.

If you've got a paper that's serially mis-cited could you elaborate/speculate as to why? And does this change how you'll publish in the future (i.e. avoiding a certian journal or not having an easy-to-mangle title)?

Trabor said...

I thought once you had tenure you didn't have to worry about this crap anymore. Am I wrong?

michiexile said...

I found the Web of Science just a few days ago, and swiftly came to the conclusion that not only are my own publications not visible there - neither are those of ANYONE I've worked with thus far in my career.

It seems to be rather irrelevant for my field.

mixlamalice said...

I would say that if a correction increases your h-index from, let's say, 20 to 50, it would be worth it.
If it is to change it from 20 to 22, I wouldn't bother.
Since I am a post-doc, my h-index is pretty low so I don't really care anyway (but my publication record and citations are pretty much correct in WoS, as far as I know).

Anonymous said...

I think it might be one of those things that I might notice if I were updating my CV in the hunt for a new job. As petty as it sounds, you never know what minor detail might differentiate you from another candidate. That said, I don't know if it came to it whether I'd actually bother!

Tara.

Anonymous said...

Web of Science has a few errors in my records, though not nearly as bad as Google Scholar, which is loaded with typos and errors in citations.

Does Web of Science even *have* a process for fixing errors, especially when the errors are in the original citation, rather than a transcription error on their part?

What about when the error is in the original publication? On one of my group papers with several authors, one student's name was mis-spelled and no one caught it, not even the student himself, though he read and approved the submission and the final proofs.

Anonymous said...

Having become more curious about this, I just searched Google Scholar for citations to a highly cited review I co-authored, using the most common mis-spelling of my last name. I found about 30 citations to my pubs but under that name on the first three pages of many, many pages...

This used to be also true in ISI but I can't get it to search citations rather than pubs anymore...

Perhaps there would be advantages to being named Smith
Mark P

Alex said...

If a database exists and is used, it might as well be accurate. And if lots of people are finding errors, well, that's something that the database's managers need to know about.

Besides, even if you are at a stage where this doesn't matter, you probably have some co-author who might also be affected if there's an error in how the paper is listed.

Will they find you a bit annoying if you keep sending corrections? Maybe. But as long as you are polite about it I don't see any significant negatives for your reputation.

Finally, a lot of people download citations from these databases to whatever software they're using to put citations into papers. Correcting this data will prevent errors in future papers. So there's another good reason. Yeah, yeah, a smart reader will know that "A. L. Jones et. al., Physical Review Letters 2010" is probably the same as "AL Jones et. al., Physics Review Letters 2010", but if the citation is there it might as well be right, you know?

Anonymous said...

Comrade PhysioProf - You may need to lower your steroid dosage - it seems you're raging. FSP asked a question. You can add a response but it's bullshit to attack the question.

As a postdoc about to be asst prof I'd say that for most of us at this career stage we ARE relative nobodies and even if they know some of our work they may not know all of it or its impact - this can send people to ISI or Publish or Perish in order to compare you to the others who've applied. There they get an impression of you based on what shows up there. I haven't corrected things in ISI but committees do look at the h-index. Should they? Different question. I am not in the position to lecture to them about it at this stage.

Comrade PhysioProf said...

Comrade PhysioProf - You may need to lower your steroid dosage - it seems you're raging.

AHAHAHAHAHAHAHAHAH!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

I haven't corrected things in ISI but committees do look at the h-index.

I have served on numerous faculty hiring committees over the years and not a single fucking time have I ever heard the term "h-index" referred to.

Anonymous said...

Who has the time? Might take as much time to do the reports, figuring out whats wrong and correcting as it would to put effort forth into revising or editing a paper already in the hopper and by doing so increasing the h-index.

Dave Backus said...

One of the reasons I prefer Google Scholar -- it sorts this kind of thing out better. Misses lots of things through typos than any modern IT system should find. My prediction is that WOS will decline in popularity over time unless it makes drastic changes.

Anonymous said...

I've been on lots of committees that discuss a candidate's h-index. Not so common for hiring committees, but more common for promotion and awards committees.

And has the last anon looked at the reporting form? If so, I'm either impressed with how fast you revise/edit papers or shocked at how slow you are at filling out a short and simple form.

Female Computer Scientist said...

It seems bibliometrics are rather malleable to begin with...

Materialist said...

As a grad student, one way I research a topic is to start from one paper and see who has cited it (from ISI). So fixing the citations doesn't just help your h-index, it also helps connect the literature.
One can, however, use incorrect citations to track how one paper leads others to cite without checking.