Ed Brayton is a journalist, commentator and speaker. He is the co-founder and president of 
The Discovery Institute has published this chart of yearly publications and citations in the scientific literature for all the members of the Iowa State astronomy department. But they're making the wrong comparison. Rob Crowther focuses on comparing Gonzalez' record to those who already have tenure, but he should be comparing that record to the other two members of the department who were up for tenure during the same time period.
There were three professors in that department awaiting tenure decisions: Gonzalez, Martin Pohl and Charles Kerton. Pohl and Gonzalez were evaluated around the same time and Pohl was given tenure while Gonzalez was not; Kerton has not yet had his tenure evaluation. So the relevant comparison is particularly that between Pohl and Gonzalez to see whether Gonzalez' record was close to Pohl's.
The answer: not even in the ballpark. On their total number of publications:
Gonzalez: 35
Pohl: 81
On their number of citations in the literature:
Gonzalez: 400
Pohl: 1503
This isn't even close, folks. He has less than half the publications that Pohl has and just over 1/4th the number of citations by other astronomers. Add to that the fact that he brought in no research money to the department, that he didn't graduate a single grad student and that he spent part of his time while trying to get tenure putting out Privileged Planet and no one could seriously be shocked that Pohl got tenure while Gonzalez did not.
Comments
I believe the point they're trying to make is that given the number of "normalized citations" where the value of the citation is inversely proportional to the number of scientists who worked on the paper, Gonzales comes out ahead of the others.
I'm not a scientist, so I don't know how significant normalization is to the measurement of a scientist's record of publication (anyone?) but since some astronomy papers do appear to have dozens of authors/contributers, I'm guessing there is some merit in differentiating between papers written by a few and those written by many.
But in any case, any benefit Gonzalez gets from normalization is wipe out by the fact that many of the citations from 2001, his best year, are almost certainly from papers he wrote or worked on before he got to ISU. From what I can gather, there is no attempt by the DI to filter out citations from his earlier work (which ISU professors are happy to point out was very good).
Only work performed while at ISU was considered in his application for tenure.
Posted by: tacitus | December 18, 2007 10:03 AM
How don't these guys see that doing a comparative analysis just makes it more obvious that Gonzalez didn't meet muster. Move on. This is embarassing.
DI takes off really fancy shoe, puts it in their mouths and pretends that it's foie gras.
Posted by: Peter | December 18, 2007 10:21 AM
Hey, I have one paper in the Astrophysical Journal with 56 citations. Can I have tenure now?
Posted by: Moopheus | December 18, 2007 10:40 AM
I don't think normalization means much at all. Some kinds of work simply require more hands on board than others because they are more complicated and involve more methodology and manipulation of the data. A person shouldn't be rewarded simply because he doesn't work in as large a group.
What matters is the total output while at that institution. The total citations to work done at that institution, which is a measure of how important it is to other scientists. The number of grad students successfully supervised and the amount of grant money brought in, both which attest to the current and continuing health of the researcher projected into the future. On these measures Gonzalez fails. He has only himself to blame. Of course that won't stop him from blaming others.
Posted by: Dave S. | December 18, 2007 10:44 AM
I agree that Gonzalez's overall record appears to have plenty of grounds for rejection. I was just curious about the normalization part of the DI's argument. The astronomy research database they used provides it as an option on the search parameters, so I suspect that some people use it as a measure of something, but it's not clear what that is.
Posted by: tacitus | December 18, 2007 12:20 PM
If they wanted some kind of normalization that reflects reality, those are the factors they would have considered. So unless someone demonstrates otherwise, I'm going to assume that this is merely a bit of DI numerology that they came up with in the hope of making things look better than they actually are.
Can't wait to see the comments when this story hits Panda's Thumb.
Posted by: Bobby | December 18, 2007 12:47 PM
Their attempt to make an issue out of normalizing for number of authors is quite absurd. It's an ad hoc and completely irrelevant argument. Does the number of authors on a paper determine its quality or worthiness for being cited by other scientists? Of course not. This is a bit like a baseball agent arguing that his client should be MVP because he had the highest batting percentage in the league against left-handed pitchers shorter than 6'1" on odd-numbered Tuesdays when there was a full moon.
Posted by: Ed Brayton | December 18, 2007 12:54 PM
I agree, but I thought it was worth discussing it since the DI is making a big deal out of it.
Posted by: tacitus | December 18, 2007 1:50 PM
It's a pretty desperate and I think transparent ploy to keep the DI (and the Gonzalez case) in the news with that apparently execreble movie Expelled coming out in a couple months. Problem is, nobody seems to care much. I think people are finally wisening up to theor tired old schtick. You can only travel so far on manufactured outrage when you have no results to show anyone.
Posted by: Dave S. | December 18, 2007 2:04 PM
In some fields it is common for a paper to have many authors and this isn't counted against the authors, though it does matter whether one is a lead author, if the paper distinguishes between the principal authors and the others who helped out in various ways. In fields in which multiple authorship is not the norm, such as my own, no explicit normalization of this type is done, but multiply-authored papers will be given less weight than singly authored papers unless it is clear that the person being evaluated is a major contributor to the paper. What one wants to discount are papers in which the person evaluated is being carried on the coattails of someone else, especially someone like his or her advisor. Committees don't like it when someone has nothing but co-authored papers because it makes it hard to tell what that person's contribution is.
There is a lot of variation among fields in how certain contributions to the work are credited. For example, there may be someone whose role was being the expert in the care and feeding of an instrument, but who had little or nothing to do with the conception of the project or the theoretical ideas involved. In some fields this person will be listed as one of many co-authors, while in others this person will just get an acknowledgment in a footnote.
Posted by: Bill Poser | December 18, 2007 2:23 PM
DAMMIT ED! Please don't give Satan, I mean Scot Borus any more bad ideas!
"This is a bit like a baseball agent arguing that his client should be MVP because he had the highest batting percentage in the league against left-handed pitchers shorter than 6'1" on odd-numbered Tuesdays when there was a full moon. "
You watch. It will be in some poor guys contract.
Posted by: J-Dog | December 18, 2007 2:53 PM
Comparison of the publications as between Gonzalez and Pohl is only part of the story. How much in the way of grants and contracts did Pohl bring in as compared with Gonzalez. This is a more important criterion (has one never heard of publish and perish).
Posted by: SLC | December 19, 2007 8:23 AM