Main Index: Trial Testimony June 10, 1997
160
1 UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF NEW YORK
2 ------------------------------x
3 GORDON & BREACH SCIENCE
PUBLISHERS S.A., STBS., LTD.
4 and HARWOOD ACADEMIC
PUBLISHERS GMBH,
5
Plaintiffs,
6
v. 93 CV 6656 LBS
7
AMERICAN INSTITUTE OF PHYSICS
8 and THE AMERICAN PHYSICAL
SOCIETY
9
Defendants.
10
------------------------------x
11
June 10, 1997
12 10:10 a.m.
Before:
13
HON. LEONARD B. SAND
14
District Judge
15
16
17 APPEARANCES
18 ORANS, ELSEN & LUPERT, LLP
Attorneys for Plaintiffs
19 BY: LESLIE A. LUPERT
ROBERT L. PLOTZ
20 PETER E. SEIDMAN
21 COVINGTON & BURLING
Attorneys for Defendants
22 BY: RICHARD A. MESERVE
JEFFREY G. HUVELLE
23 SUSAN L. BURKE
24
25
161
1 (Trial resumed)
2 THE COURT: The witness may resume the stand.
3 BRUCE KINGMA, Resumed.
4 THE COURT: I remind you, you are still under
5 oath.
6 CROSS-EXAMINATION (Continued)
7 BY MR. HUVELLE:
8 Q. Dr. Kingma, I would like to start off by just
9 talking briefly about the role of scientific journals.
10 Are you familiar with the history of scientific
11 journals?
12 A. Briefly familiar with the history.
13 Q. You are aware that they have been around for
14 several hundred years?
15 A. Yes.
16 Q. Is it your understanding that scientific journals
17 have been the principal means by which advances in science
18 have been communicated within the scientific community?
19 A. Principally, yes. There's obviously a number of
20 other means of communication too, I guess. I would say that
21 they are an important part.
22 Q. But isn't it true that it is widely considered
23 that, in science, the scientific journal has been the
24 principal medium for the transmission of scientific
25 knowledge?
162
1 A. Yes.
2 Q. Over the hundreds of years that scientific
3 journals have been in use, there have developed traditions
4 regarding the writing of articles for scientific journals;
5 is that correct?
6 A. Yes.
7 Q. One of those -- one of the academic traditions
8 relating to articles for scientific journals is that the
9 author cite in connection with his article as references the
10 prior works that have contributed in a significant way to
11 his analysis in his own article; is that correct?
12 A. Correct.
13 Q. It's part of the academic tradition to treat such
14 citations as an important part of a scientific article?
15 A. As an important part of it, yes.
16 Q. And it is true that, for most scientific
17 journals, there is a peer review process that looks over the
18 article and makes sure that it is of sufficient quality; is
19 that correct?
20 A. Correct.
21 Q. In that it conforms to the standards of academic
22 scholarship?
23 A. Correct.
24 Q. Including the citation of prior works that have
25 contributed to the author's own analysis?
163
1 A. Correct.
2 Q. These academic traditions apply to the field of
3 economics, too, your own field?
4 A. Correct.
5 Q. When you write an article, you exercise care and
6 responsibility in citing those prior articles that have
7 contributed in a significant way to your own analysis?
8 A. Correct.
9 Q. And other scientists and academicians with whom
10 you are familiar have likewise exercised the same care and
11 responsibility in citing the prior works, prior articles,
12 that have contributed to their own analysis, correct?
13 A. Correct.
14 Q. Is it true that numerous citations to a
15 scientific journal tell us that the scientists most engaged
16 in advancing scientific knowledge find that articles in that
17 journal contribute in a significant way to their own work?
18 A. Most citations to a journal show that that
19 particular journal -- is that what you said -- makes a
20 substantial contribution, that would be correct.
21 Q. Do you agree that the total number of citations
22 to a journal is useful information for a librarian to have
23 in connection with the assessment of the effectiveness of
24 journals; is that correct?
25 A. I would say there is better information than
164
1 citations, because you're now taking it from the world of
2 journals down to the world of my library, my patrons, and
3 that's where the analysis starts to crumble in that a
4 particular journal might have a lot of citations but not be
5 particularly useful to the patrons at my particular library.
6 Q. You said there are other better pieces of
7 information, but would you agree that the total number of
8 citations to a journal is useful information for a librarian
9 to have in connection with the assessment of the
10 effectiveness of that journal?
11 A. I would say that there is better information in
12 terms of making collection management decisions. Citations
13 might provide you with -- the total citations might provide
14 you with a piece of information about a set of journals, but
15 those decisions still have to be made in concert with what
16 the faculty at your library want.
17 Q. But do you agree that the piece of information
18 provided by the total number of citations to a journal is
19 useful to a librarian?
20 A. It may be useful to a librarian.
21 Q. You have written a textbook on information
22 science?
23 A. I would call it economics information. I
24 wouldn't call it a textbook on information science.
25 Q. OK. I will accept your characterization of your
165
1 book readily.
2 And this is a book for students?
3 A. Correct.
4 Q. And it goes over the basics of economics as
5 applied to information sciences?
6 A. Correct.
7 Q. And one of the issues it addresses is the subject
8 we mentioned yesterday, market failure?
9 A. Correct.
10 Q. The failure of the consumer to have appropriate
11 information about the price or quality of a product?
12 A. Correct.
13 Q. In your textbook, you identified a number of
14 examples of markets where there might be a problem with the
15 available information, or in terms of the lack of relevant
16 information?
17 A. Correct.
18 Q. One of the markets that you identified was the
19 market for journal subscriptions?
20 A. Correct.
21 Q. The problem you identified was that the
22 individual subscriber or library cannot determine the
23 quality of a journal until after paying for it and receiving
24 the subscription?
25 A. Correct.
166
1 Q. Do you recall that?
2 A. Correct.
3 Q. You suggested a market solution for that absence
4 of information?
5 A. I believe it is a table that you are reading
6 from --
7 Q. Right.
8 A. -- in which a number of market solutions are
9 suggested.
10 Q. Right. Four solutions?
11 A. I believe it's four. I would like to refresh my
12 memory.
13 MR. HUVELLE: I would like to show the witness
14 Defendant's Exhibit DDD.
15 Q. At the bottom of the page, you identify the
16 market for journal subscriptions?
17 A. Correct.
18 Q. One of the solutions that you propose for the
19 problem of lack of information regarding quality of journals
20 is the number of citations to published articles?
21 A. The third solution there, correct.
22 Q. Is it also true that, in your writings, you have
23 referred to the problems that have arisen because of the
24 increase in the number of scholarly journals that publish
25 many articles never cited in later research? Do you recall
167
1 that?
2 A. Yes, I -- it's from one of the books that I've
3 written, yes.
4 Q. You said that these articles seem to some people
5 to exist solely for the purpose of padding the author's
6 bibliography; do you recall that?
7 A. Correct, yes.
8 Q. Isn't it true that in writing this article and
9 making that observation, you used the fact that some
10 articles are never cited in later research as an indication
11 of the low quality of the articles; is that true?
12 A. As a signal of this, correct.
13 Q. When did you receive tenure?
14 A. September of 1996.
15 Q. And you had to apply and be reviewed for that?
16 A. Yes.
17 Q. In connection with that, you had the opportunity
18 to submit a short statement of your accomplishments; is that
19 correct?
20 A. Yes.
21 Q. You had to specifically outline your achievements
22 in the area of research, teaching, and university services?
23 A. Yes.
24 Q. Isn't it true as an indication of the quality of
25 your work you included in that submission the number of
168
1 times your journal articles had been cited in subsequent
2 articles?
3 A. Yes.
4 Q. You referred yesterday to the Bensman article,
5 Bensman being a professor at Louisiana State University, I
6 believe?
7 A. He's a librarian at Louisiana State, actually.
8 Q. Librarian. And you cited that article to support
9 the proposition that faculty perceptions at Louisiana State
10 University in the Chemistry Department do not correlate with
11 rankings based on the impact factor?
12 A. Correct.
13 Q. Isn't it true that one of the central conclusions
14 of Mr. Bensman was that total citations to a journal highly
15 correlated with those faculty rankings?
16 A. Yes.
17 Q. An impact factor normalizes the size of journals;
18 is that correct?
19 A. That's one of the things it does. Obviously it
20 differs from total citations in another way, too.
21 Q. If you had total citations, you might have a huge
22 journal with many articles and thousands of citations?
23 A. Right.
24 Q. Impact factor divides the total citations by the
25 number of articles?
169
1 A. It's -- it divides --
2 Q. In a two-year period.
3 A. In a two-year period, right. So it's a snapshot
4 and then divides by the number of articles, correct.
5 Q. Divides the total citations in a one-year period
6 by the number of articles in the prior two-year period?
7 A. Correct.
8 Q. And so you get an average number of citations per
9 article?
10 A. Correct.
11 Q. And so that's helpful, or it's fair to a smaller
12 journal because you're dealing with averages on a per-
13 article basis; is that correct?
14 A. I wouldn't call it fair or helpful. I mean, what
15 you're doing is normalizing in a way that I disagree with.
16 But, nonetheless, you can take a smaller -- your statement,
17 part of your statement is correct: If you take a smaller
18 journal and divide it by the number of articles, that might
19 improve the impact factor for that smaller journal.
20 Q. You could have two journals, in each one, each
21 and every article cited five times during a particular year?
22 A. Right.
23 Q. Every prior article is cited five times. So the
24 impact factor is going to be five for both journals?
25 A. Correct.
170
1 Q. But if one journal has a thousand articles,
2 they're going to have 5,000 citations, and if the other
3 journal has 20 articles, they'll have 100 citations?
4 A. Correct.
5 Q. So it's a different analysis?
6 A. It's a different number of citations.
7 Q. It's a different number?
8 A. Yes.
9 Q. It's a different number?
10 A. Right.
11 Q. Both are true numbers?
12 A. Right.
13 Q. They're both based on objective facts. It's just
14 different ways of looking at the data.
15 A. I don't know if I would agree that it's objective
16 facts. I mean, there are certain other issues going on with
17 citations. But it's -- you're right. It's two -- the
18 mathematics of that are that there are two different bases,
19 two different numbers you are producing out of the impact
20 factor.
21 Q. Well, the objective fact is the number of
22 articles, the number of citations.
23 A. There is a number of citations, number of
24 articles, correct.
25 Q. Isn't it true in Bensman's analysis, that he
171
1 suggested that the reason impact factors did not correlate
2 with faculty rankings is that the faculty gave -- ranked
3 higher the large journals with lots of citations?
4 A. Correct.
5 Q. And so the faculty rankings tended to disfavor
6 the small niche journals at the expense of the large society
7 journals?
8 A. Correct.
9 Q. Your criticism of the impact factor is that it is
10 too small of a window?
11 A. That's one of my criticisms of it, correct.
12 Q. Impact factor looks at citations in a particular
13 year to articles in the two prior years?
14 A. Correct.
15 Q. You think it would be better to look at the total
16 number of citations, the total number of citations in that
17 year to all prior editions of the journal?
18 A. Citation -- total citations, right, is looking at
19 the total number of citations, and that's certainly better
20 than using a two-year/one-year window.
21 Q. You would look at all the citations in 1986 to
22 articles published prior to 1986, back to 1975, 1974,
23 where --
24 A. Correct.
25 Q. And of course that would disadvantage a new
172
1 journal?
2 A. Correct.
3 Q. Now, if you look just at the two-year articles
4 appearing in the two-year, prior year period -- the prior
5 two-year period, you get information as to what's happening
6 currently, don't you?
7 A. Looking at only a two-year snapshot gives you
8 exactly that, a two-year snapshot.
9 Q. And the reason you're looking at citations or
10 you're talking to the faculty, you're measuring use, is to
11 make a judgment about the future; is that correct?
12 A. If you were -- if we're talking about librarians
13 making collection development decisions, yes. It's a matter
14 of what's going to happen in the future, what's the expected
15 value.
16 Q. Is the past a reliable indication of the future?
17 A. One would hope. It is the only thing you can
18 use -- the past by definition is the only thing you can use
19 to guess at what's going to happen in the future.
20 Q. And certainly there are times when if you look at
21 the 10-year past history or the 20-year past history, that's
22 less reliable than if you look at the prior two years?
23 A. No, you're going to make fewer mistakes looking
24 at a 20 or 20-year past history than only the prior two
25 years.
173
1 Q. You don't know whether that's true in the case of
2 scientific journals, though, do you?
3 A. I don't know if that's true in the case of
4 scientific journals, but the -- when you think about how
5 you're sampling those journals, it's clear that if you're
6 looking at a 20-year history you're going to be better off
7 than only looking at the past two years. You may be
8 ignoring pertinent information by discarding 18 years.
9 Q. You may be misled, as well, by a journal that had
10 a glorious past and a troubled present?
11 A. Hopefully you would be able to pick out that
12 troubled present by their -- the editors of that journal,
13 any kind of knowledge you might have about that journal in
14 its recent history.
15 Q. But isn't it true whether you picked two years or
16 10 years or 20 years, in some cases that past history may
17 not be the best slice to look at?
18 A. In --
19 Q. For some journals.
20 A. For some journals, depending on the slice, yes,
21 correct.
22 Q. Have you read the deposition of Ms. Hunter in
23 this case regarding impact factors?
24 A. No, I have not read the deposition of Ms. Hunter.
25 Q. Do you know whether other publishers use impact
174
1 factors or total citations when they want to know how well
2 their journals are doing?
3 A. Yes.
4 Let me finish what I was about to say, is, I have
5 read the report from Ms. Hunter but not the deposition of
6 Ms. Hunter. And in the report I believe she says that
7 Elsevier uses impact factors.
8 Elsevier is the name of the publisher.
9 Q. Let me turn to the price side of the equation.
10 Do you agree that the librarian must make
11 decisions regarding competing journals based in part on
12 price information?
13 A. Yes.
14 Q. And librarians must make comparisons between
15 journals in terms of price, among other factors?
16 A. Yes.
17 Q. Do you agree that the information on price should
18 be in a format that allows the librarian to make wise
19 management decisions about journal acquisitions?
20 A. Yes.
21 Q. You agree that price information can be presented
22 in a variety of ways?
23 A. Yes.
24 Q. Do you recall that Professor Barschall, in both
25 of his articles, displayed the annual subscription price for
175
1 each journal as well as the cost per thousand characters?
2 A. Yes.
3 Q. He displayed that information for two years?
4 A. Yes.
5 Q. But in addition to the annual price or
6 subscription price, Barschall normalizes the size of the
7 journals to provide a common measure of cost per amount of
8 article information; is that correct?
9 A. Yes, that is what he did.
10 Q. Do you agree that it is a sound approach to
11 normalize the cost in that manner?
12 A. No, I do not.
13 Q. I believe you discussed the reasons for that view
14 yesterday; is that correct?
15 A. Yes.
16 Q. Were you here during the testimony of Dr. Taylor?
17 A. No, I was not.
18 Q. Did you have an opportunity to read the testimony
19 of Dr. Taylor?
20 A. No, I did not.
21 Q. Let me --
22 A. Let me --
23 Q. -- pose this as a hypothetical to you.
24 A. All right.
25 Q. The journal editor whose journal price was
176
1 $9,000, compared to $507 for another journal normalized
2 cost, in order to calculate the cost per article, would you
3 say that's a wrong approach?
4 A. Are you saying the publisher did that?
5 Q. Yes.
6 A. For the publisher to do it is actually a bit
7 different from the librarian doing it on the purchase
8 decision.
9 Q. Why is that?
10 A. Well, because a publisher is also in some sense
11 selling those articles through document delivery or
12 interlibrary loan and has to set copyright fees on a per-
13 article basis. So for the publisher to divide it up on a
14 per-article basis might give that publisher information
15 about how to set a copyright fee as articles are delivered
16 through this alternative means.
17 Q. But if the editor who did this, the editor of the
18 $9,000 journal, did it in response to complaints that his
19 journal's price was very high, and in order to demonstrate
20 the cost effectiveness of his journal, would you say that's
21 the wrong approach?
22 A. I think it's the wrong approach for a librarian
23 making a decision to tell a publisher what they should or
24 shouldn't do. But as far as librarians go, their
25 responsibility is to purchase the entire subscription or to
177
1 find an alternative access to that information through a
2 system like interlibrary loan.
3 Q. I believe you conceded yesterday that normalizing
4 price on the basis of a common unit of measurement is
5 appropriate in a grocery store?
6 A. Correct.
7 Q. Is it --
8 THE COURT: We will take a five-minute recess.
9 MR. HUVELLE: Fine.
10 (Recess)
11 THE COURT: Sorry for the interruption. You may
12 be seated.
13 You know, as long as we have interrupted, you
14 made some reference yesterday to a book you had written or a
15 study you had done on comparing ownership to lease access;
16 is that --
17 THE WITNESS: Well, to access. It's called the
18 access-to-ownership comparison, where ownership means buying
19 the journal subscription and access is acquiring articles on
20 an as-needed basis through interlibrary loan. So the
21 faculty member comes in, wants a certain --
22 THE COURT: Acquiring from, what? From other
23 libraries?
24 THE WITNESS: From other libraries, from
25 commercial document delivery sources, other places that have
178
1 the journals you're looking for, essentially. They're
2 photocopying an article on that journal and sending it on to
3 you at your library.
4 THE COURT: Which creates copyright fees?
5 THE WITNESS: Well, there is a copyright fee
6 associated with that, so --
7 THE COURT: I just want sure what the terminology
8 meant. Thank you.
9 BY MR. HUVELLE:
10 Q. If I may just pursue that subject for just a
11 moment.
12 In connection with your study, you worked out a
13 formula by which a library should determine if it is cheaper
14 to borrow an article than to buy the journal?
15 A. Correct.
16 Q. And you noted that one consequence of using your
17 formula, if libraries did, would be that the smaller
18 journals that were used less, that were more expensive,
19 would be canceled by some libraries?
20 A. I don't know if I said "smaller," but journals
21 that were used less, more expensive, clearly both these
22 things factor into the formula, the price of the journal and
23 the amount to which it's used. So, low-use, high-cost means
24 you're more likely as a library to cancel that particular
25 subscription.
179
1 Q. And one of the macro effects that you did not
2 address in your cost/benefit analysis was the consequence
3 that, as a result of canceling journals that are high-cost,
4 low-use, the cost of those journals would actually go up, or
5 the price of the journals would go up even further?
6 A. Right.
7 Q. Which would then lead to more libraries
8 canceling?
9 A. Correct.
10 Q. And your position was, well, an individual
11 library can't worry about that?
12 A. Correct.
13 Q. Back to the grocery store.
14 A. OK.
15 Q. The appropriateness of unit costs in a grocery
16 store, in your judgment does that apply to all products in
17 the grocery store? We focused a little bit on pasta, but I
18 want to know if it has more general application.
19 A. In each product, there is going to be a container
20 of a certain size, but also quality differences. Now,
21 clearly unit pricing on pasta doesn't allow me to compare it
22 to -- and there's unit pricing down a few aisles on beans,
23 but I'm not really making a comparison between pasta and
24 beans. There are quality differences. There are, you know,
25 significant differences between those products; that the
180
1 unit pricing is for a different reason than comparing the
2 cost per ounce for beans to the cost per ounce of pasta.
3 Q. But do you agree with the appropriateness of
4 displaying unit cost data for beans and fruit juice and ice
5 cream, different products?
6 A. Many of the different products in the grocery
7 store, yes. I mean, there are clearly examples in the
8 grocery store where you don't have unit pricing, you don't
9 have a price per ounce of a broom, because we're not
10 interested in the price per ounce of a broom. You just want
11 to know how much you're going to pay for the broom.
12 Q. As I understood it, the reason you gave why a
13 common unit price is not appropriate for scientific journals
14 is that scientific journals are more like a car than they
15 are like pasta?
16 A. Yes, it's tough to draw that comparison, but,
17 yes, I would agree with that. They're more like a car in
18 that there's information, there's something else embedded
19 within the product than they are like pasta. Yes, I would
20 agree with that, although it's -- it's tough to think about
21 it.
22 Q. You don't agree with my statement. I'm simply
23 reciting yours.
24 A. OK.
25 Q. What is the difference between a car and pasta
181
1 that is pertinent to the distinction you're making?
2 A. A car you're purchasing as a single unit, and
3 within that there are many quality characteristics of a car
4 that are influencing your decision to buy, depending on how
5 many people you want to fit in the car, how fast you want
6 the car to go, what color it is, etc. There are many
7 characteristics to the car, and you purchase it as a car.
8 Whereas with pasta you're looking at it and
9 within the grocery store setting making the comparison
10 perhaps between two different sized boxes of pasta. Within
11 each box is relatively the same quality good -- relatively
12 the same quality good for the decision you're making, but
13 you want to know the price per ounce of those two boxes,
14 which might be of different size.
15 Q. So are you telling us that the operative
16 distinction is whether there is variation in quality among
17 the products?
18 A. It's variation in quality, yes, along with the
19 amount you're purchasing.
20 Q. Do you say it's inappropriate as to cars because
21 when you buy a car you consider quality?
22 A. When you buy a car you consider quality, and
23 you're also, you know, purchasing a car as one solid good.
24 You could, even in theory, if you had two cars of the exact
25 same quality, I still don't think you would want to
182
1 normalize it on a price per pound basis.
2 Q. The cost basis -- a cost comparison is a cost
3 efficiency analysis?
4 A. Right, correct.
5 Q. Which compares cost, correct?
6 A. Correct.
7 Q. It has nothing to do with quality?
8 A. Correct. Doing a cost efficiency analysis makes
9 the assumption that the quality is very similar, not
10 identical.
11 Q. Is it your testimony that there are no
12 differences in the quality of pasta or other goods you see
13 in a grocery store?
14 A. There are certainly differences in the quality of
15 pasta.
16 Q. Isn't the whole point of unit pricing in a
17 grocery store that you see the brand name, the Proctor &
18 Gamble Tide, and then you see the Safeway detergent X, and
19 you go in with some assumption as to the differences in
20 quality, and then the price data on a unit basis allows you
21 to see whether for one you are paying $1.50 per ounce and
22 the other you're paying 5 cents an ounce. It just gives you
23 the cost comparison, and then you use that in making a
24 decision; is that correct?
25 A. Correct. And that's actually a very good
183
1 example, because as you're using detergent, you're obviously
2 using it by the ounce, not by the unit you're purchasing.
3 Q. So that's a factor in whether or not the unit
4 cost is appropriate?
5 A. In the sense that you're using it by the scoop,
6 by the ounce.
7 Q. If I have these journals and I know the annual
8 subscription price of these journals, I know the annual
9 subscription price of these journals, and I know the annual
10 subscription price of the green journals, and those are all
11 one year of that journal, and then the orange and white is
12 another journal, one year. The librarian knew the price of
13 all four journals. Is there any cost price information
14 about these journals that in your judgment would be helpful
15 in making acquisition decisions?
16 A. Well, just the subscription price, simply that.
17 Q. That's the only information?
18 A. Well, the only information about price, cost.
19 You want to know --
20 Q. What about --
21 MR. PLOTZ: Please let him answer.
22 MR. HUVELLE: I'm sorry.
23 A. You want to know to what degree these journals
24 are going to be used by your patrons. That's the other side
25 of the coin.
184
1 Q. Let's focus on price.
2 A. OK.
3 Q. What about telling the librarian what the cost
4 price per page is of the journals?
5 A. I don't think that's useful information.
6 Q. Isn't it true that if a librarian knew the price
7 of this journal and he knew the price of this journal and he
8 knew the price of those journals, that he wouldn't think
9 about that, or that what he would do is, he would take into
10 account the differences in the amount of information he's
11 getting; isn't that right? He would normalize in his own
12 mind?
13 A. One would hope that he or she would not do that,
14 that they would look at the usefulness of what they're
15 purchasing rather than simply the size of it. The example I
16 gave in my deposition earlier is this, you know, these
17 places that sell books by the foot, for aesthetic reasons.
18 The cheapest way to fill your library shelves is to purchase
19 books by the foot, but that doesn't mean that your patrons
20 are going to be using those materials, and it would be a
21 poor collection management decision to simply buy books by
22 the foot which are sold for aesthetic reasons of filling up
23 people's shelves in houses.
24 Q. Suppose we had four journals, they're all good
25 enough to get in the ISI database. In each case there's one
185
1 faculty member who says, I'm on the board of that particular
2 journal so I'd like to have it. It's embarrassing me for me
3 to go to board meetings and not have it.
4 So we know something about the quality. But the
5 librarian needs to justify to his boss, we're using the
6 money wisely, we're getting a good bang for the buck.
7 Right?
8 Isn't that what unit pricing is about? Bang for
9 the buck? What you're getting for your money?
10 A. I don't know if I'd call it "bang for the buck."
11 I just strongly disagree with normalizing pricing for
12 journals on a per-page or per-character basis.
13 Q. You're saying that if this journal is a little
14 bit less expensive than most journals, that the librarian
15 ought to, in making the decision, just consider this as less
16 expensive without regard to the fact that that journal
17 provides, or those journals provide perhaps ten times as
18 many articles?
19 A. I think your question is putting the librarian in
20 a vacuum without any information about use or possible use
21 of those journals. If you only have price and you have
22 nothing else, boy, that -- I mean, that would be an
23 impossible decision to make.
24 Q. I'm not --
25 A. OK.
186
1 Q. The librarian has lots of other information. Is
2 it appropriate for the librarian, in addition, in addition
3 to knowing the annual subscription price, and in addition to
4 knowing things about quality, to know what the cost per
5 character is, or the cost per page, or the cost per article,
6 or some other common measurement that allows a comparison on
7 a common basis?
8 A. No. The normalization that should occur is the
9 cost per subscription, how much I'm paying for a
10 subscription. That's really the only normalization that
11 should occur under these circumstances.
12 Q. Isn't it true that in the case of the green
13 journals over there, it's not like a car in either you get
14 the whole thing or not, that contains several thousand
15 articles; is that correct?
16 A. Correct.
17 Q. No one is going to read all the articles?
18 A. I don't know if there's anybody that would read
19 all the articles. The editor would read all the articles.
20 Q. Well, we can even wonder about that.
21 In the pasta example, it's divisible into meals,
22 correct?
23 A. Correct.
24 Q. The university cafeteria can buy pasta on a
25 per-unit basis and then students come in and they take a
187
1 large portion or a small portion, but they do that, right?
2 A. Correct.
3 Q. And in the case of the journal, the 2,000, 3,000
4 articles, the librarian buys it, and different students come
5 in, read one article, two articles. They treat it not as an
6 entity but as a divisible product; is that correct?
7 A. The students treat it as a divisible product.
8 Q. True.
9 A. Divisible in the sense that it's their use,
10 right? It's a student coming into the library and using
11 that journal. You could have 50 students all using the same
12 article from that journal.
13 Q. So in your view, the only -- the only -- price
14 information is the annual subscription price?
15 A. The annual subscription price, yes. There is
16 the -- these fixed costs associated with buying a journal,
17 too, that involve shelving and cataloging, etc., so --
18 Q. You mentioned yesterday that Dr. Barschall did
19 not take that into account?
20 A. Correct.
21 Q. Should he have?
22 A. It's relatively the same for all these journals.
23 Q. So there's no reason to?
24 A. There's really no reason to, under these
25 circumstances. Those costs are about $60 to $70 per journal
188
1 subscription, so when you're talking about prices and in the
2 early '80's when you did this, they were probably less than
3 $60 or $70.
4 Q. But it washes out?
5 A. It probably washes out as a small percentage of
6 the whole.
7 Q. If you were a librarian and you looked at this
8 set of journals, this set of journals and those two sets,
9 would you make any assumptions about what the subscription
10 price would be?
11 A. I am not a librarian, but, no, I wouldn't make
12 any assumptions.
13 Q. You have no way of telling which one would be
14 more expensive?
15 A. No. Hopefully you would be able to find that
16 price information.
17 Q. So it would not be surprising to you if this
18 journal cost more than both of those journals combined?
19 A. It wouldn't be particularly surprising, no.
20 Q. And it would not be surprising to you if the
21 price per page of this journal was 67 cents and the price
22 per page of those journals was 5 cents?
23 A. I wouldn't be --
24 Q. The same with this? It wouldn't surprise you?
25 A. No.
189
1 Q. But that's not something a librarian ought to
2 think about in deciding whether or not to subscribe to this
3 journal or that journal?
4 A. They should not think about the price per page,
5 no.
6 Q. But it certainly allows us to say this journal is
7 much more expensive?
8 A. The subscription price might allow us to say that
9 that journal is much more expensive, yes.
10 Q. And this one, which has the same per-page price
11 as this one, can we say that this journal is more expensive
12 than those two when the total price is less than those?
13 A. I'm sorry. I'm getting a little confused. The
14 per-page price --
15 Q. The per-page price is the same?
16 A. Same.
17 Q. Total price is less of this than this and less
18 than those?
19 A. How can the per page price be the same and the
20 total price when that's a smaller stack? I don't
21 understand.
22 Q. The per-page price is the same.
23 A. Right.
24 Q. The total price of the smaller stack is therefore
25 less.
190
1 A. Oh, is less, OK. Yes.
2 Q. Would you agree that this journal also is more
3 expensive than those journals?
4 A. I don't know that -- you're asking me to make a
5 mathematical calculation of multiplying the number of pages,
6 so --
7 Q. If the total subscription price of this journal
8 is less than those journals --
9 THE COURT: You know, I think if somebody reads
10 this record, it might be helpful to know that what you are
11 referring to are stacks of blue-covered --
12 MR. HUVELLE: Your Honor, that's an excellent
13 point, and I can try to clarify it, although it would have
14 been helpful, I guess, if I would tried earlier, but I will
15 do my best now.
16 THE COURT: Sometimes you read a record and --
17 MR. HUVELLE: I can imagine.
18 THE COURT: It hurts me here, Doctor.
19 Q. The journals with the 67-cent-per-page price that
20 I have been referring to, one is an orange journal; is that
21 correct?
22 A. Correct.
23 Q. And that bears the title Ferroelectrics?
24 A. Correct.
25 Q. And then the other journal that's small in number
191
1 but high in price is a white-and-black journal called
2 Physics and Chemistry of Liquids?
3 A. Correct.
4 Q. And let me represent that the two larger stacks
5 are a green stack, which is Physical Review B, and a black-
6 and-orange stack, which is Journal of Applied Physics.
7 MR. HUVELLE: I will represent that for the
8 record. And thank you, your Honor, for --
9 THE COURT: And the relative heights of the
10 stacks I think are part of what you can do to make your
11 question meaningful.
12 Q. Would you estimate for us the size of the orange
13 Ferroelectrics stack?
14 A. It looks to be about ten inches high.
15 Q. Or a little less?
16 A. That's --
17 Q. And the Physics and Chemistry of Liquids?
18 A. It looks to be two inches high.
19 Q. And the green Physical Review B?
20 A. That's those two stacks, right?
21 Q. The two stacks, each about a foot and a half.
22 A. Yes, so about three feet high together.
23 Q. And the Journal of Applied Physics is --
24 A. Is just that one stack?
25 Q. Yes.
192
1 A. So about a foot and a half.
2 Q. Do you have any of the exhibits up there from
3 yesterday?
4 A. I have Physics Today, Bulletin of the Physics
5 Society, a table from my book, and the impact numbers from
6 the Science Citation Index for '86.
7 Q. I would like you to look at Plaintiff's Exhibit
8 2.
9 MR. HUVELLE: Your Honor, I have an extra copy if
10 you need one.
11 THE COURT: I think I have it up here. Yes.
12 Q. Do you have Plaintiff's Exhibit 2?
13 A. Yes.
14 Q. Can you turn to Table 3. Is this a ranking of
15 journals by cost per character?
16 A. Yes.
17 Q. And it's a three-page list?
18 A. Yes.
19 Q. Now, in the first page, the journals run from
20 39 -- .39 to 3.4 cents?
21 A. Yes.
22 Q. Now, on page 2, they run from 3.4 to 7.4 cents?
23 A. Yes.
24 Q. Is that correct?
25 And these are ranked in order of increasing cost?
193
1 A. Yes.
2 Q. And the total on the three pages is about 200
3 journals that Barschall studied?
4 A. Yes, about.
5 Q. And then on page 3, you go halfway down the page
6 and you get to the first journal that has a cost of 10 cents
7 per thousand characters. Do you see that?
8 A. Yes.
9 Q. And then you go down a little further and there's
10 a journal with a cost of 11.9 cents per thousand characters?
11 A. Yes.
12 Q. Then you jump up to 14 cents per thousand
13 characters; is that correct?
14 A. Yes.
15 Q. One journal.
16 Then you jump again to 16 cents per thousand
17 characters?
18 A. Yes.
19 Q. And then we have the bottom 14 journals, the most
20 expensive -- 14 most expensive journals on a cost per
21 thousand character?
22 A. Yes.
23 Q. And of those 14 most expensive journals on a
24 cost-per-thousand-character basis, 11 are Gordon & Breach
25 journals; is that correct?
194
1 A. Yes.
2 Q. And so the least expensive of the Gordon & Breach
3 journals ranks 187 out of the 200?
4 A. If there's 200 here, yes.
5 Q. Can you turn to Plaintiff's Exhibit 3.
6 A. Yes.
7 Q. Do you have that? Can you look at Table 1.
8 A. Yes.
9 THE COURT: Bear with me just a moment.
10 MR. HUVELLE: I have an extra copy.
11 THE COURT: I will take your extra copy. Thank
12 you.
13 Exhibit 3, what are you looking at?
14 MR. HUVELLE: We are looking at Plaintiff's
15 Exhibit 3, Table 1.
16 Q. You looked at this yesterday, Dr. Kingma; is that
17 correct?
18 A. Yes.
19 Q. And you looked at the categories listed on the
20 left on Table 1?
21 A. Yes.
22 Q. Dr. Barschall has grouped the journals included
23 on this page under eight categories?
24 A. Yes.
25 Q. And you have no reason to dispute that those are
195
1 reasonable and appropriate categories?
2 A. I have no reason to dispute these categories.
3 I'm not a physicist, so --
4 Q. Do you agree that the top-ranked journal in the
5 Letters Journals category is an AIP journal?
6 A. Ranked by ratio of cost to impact? Yes.
7 Q. The top two are AIP and APS; is that correct?
8 Under Letters Journals?
9 A. Correct.
10 Q. And under Review Journals, the top-ranked journal
11 is an APS journal?
12 A. These are actually ranked by cost per thousand
13 characters. Yes, correct.
14 Q. The top-ranked -- but it's true as to cost, ratio
15 of cost to impact factor as well, is it not?
16 A. Correct.
17 Q. And in the Atomics Physics category, the
18 top-ranked journal by ratio of cost to impact is an APS
19 journal?
20 A. Correct.
21 Q. The same is true in the Condensed Matter Physics
22 category?
23 A. Correct.
24 Q. The same is true in the Nuclear Physics category?
25 A. Correct.
196
1 Q. The same is true in the Particle Physics
2 category?
3 A. Correct.
4 Q. The same is -- well, in the case of Applied
5 Physics, it's an AIP journal that is the top ranked?
6 A. Correct.
7 Q. And in the Instrumentation category, it's an
8 AIP-APS journal that is top-ranked?
9 A. Correct.
10 Q. And the number two and number three journals in
11 that category are also AIP journals; is that correct?
12 A. In instrumentation?
13 Q. Yes.
14 A. It says IOP.
15 Q. Well, if you look at the ratio of cost to
16 impact --
17 A. Oh.
18 Q. Is it --
19 A. Yes, correct.
20 Q. The top three are all AIP journals in that
21 category?
22 A. By cost to impact ranking.
23 Q. Right. Dr. Kingma, if you will indulge me in a
24 hypothetical or an assumption, the assumption is that the
25 ratio of cost to impact is an appropriate measure of cost
197
1 effectiveness. Can we proceed with that assumption?
2 Let me ask you, if we accept that assumption, is
3 it true that the data on Table 1 would therefore support the
4 conclusion that the journals of APS and AIP are the best
5 bargain?
6 A. Obviously I don't agree with the assumption.
7 Q. I appreciate that.
8 A. If you assume that that's the ranking, then
9 that's what the numbers say.
10 Q. Thank you.
11 Let me go back now to where we were when we ended
12 at the conclusion of yesterday. You had identified five
13 possible criteria that one might use in measuring the
14 effectiveness of journals. Do you recall that?
15 A. Yes. I think we talked about them as proxies,
16 numbers that could serve as proxies for use value of a
17 journal.
18 Q. I would like you to focus on effectiveness.
19 A. OK.
20 Q. Isn't it true that those are five possible
21 measures for the effectiveness of journals?
22 A. Some good measures, some bad measures, yes; by
23 definition they are measures.
24 Q. But they are all ones that you have identified,
25 whether good or bad?
198
1 A. Correct.
2 Q. And we discussed readership, one of the possible
3 measures, yesterday. I would like to go on to the second
4 one, use. OK?
5 A. OK.
6 Q. And can you tell us what you mean by "use"?
7 A. "Use" is the number of times somebody is going to
8 pull it off the shelf, pull a journal, an issue of that
9 journal off the shelf and use it. Now, they may use it and
10 photocopy an article from it and ultimately read that
11 article several times, but use is simply the pulls off the
12 shelf.
13 Q. And as you used the term, all uses are treated
14 the same?
15 A. You can treat all uses the same. You can weight
16 them by who uses it, if a faculty member pulls it off the
17 shelf versus a student pulling it off the shelf.
18 Q. One difficulty of use as a measure is that it is
19 difficult to tell, or difficult to differentiate between
20 uses; isn't that true?
21 A. It would be possible to construct a use survey
22 where you are differentiating between uses. That would
23 clearly be a more expensive use survey than one in which you
24 were not differentiating between uses.
25 Q. Typically, if a library tries to study use, it
199
1 tries to determine the number of times the journal is pulled
2 off the shelf, correct?
3 A. Typically -- can you repeat the question?
4 Q. Typically when a library does a use study, it
5 determines how many times the journal is pulled off the
6 shelf?
7 A. Correct.
8 Q. When you do that kind of study, you don't know
9 whether the person glanced at the article, read two
10 sentences, and then put it back?
11 A. Correct.
12 Q. You don't know whether the user reshelved it
13 before it could be counted?
14 A. Correct. Hopefully in designing your use study
15 you have taken care to try and prevent users from doing
16 that.
17 Q. And you don't know if, after the journal has been
18 pulled off the shelf, it has been used by multiple people?
19 A. Correct. Again, hopefully you have designed your
20 study such that you are reshelving it as quickly as possible
21 so that there is limited multiple use possible.
22 Q. And you don't know whether it is use by a
23 first-year physics student, who reads the article, becomes
24 discouraged by the complexity, and switches majors, or
25 whether it's use by a Nobel laureate who finds that the
200
1 article contributes importantly to his own research, do you?
2 A. Again, it would depend on your design of the
3 survey.
4 Q. But it would be legitimate for a library to
5 differentiate between different uses?
6 A. If they could, yes.
7 Q. It would be legitimate for a library to focus
8 entirely on uses by the faculty?
9 A. I don't think it would be legitimate to focus
10 entirely on uses by the faculty. Clearly students use that
11 library, and their concerns as patrons should be addressed.
12 Q. But might it be the case that the library would
13 decide that, by determining the uses of the faculty, that
14 that was a reliable basis for making decisions about what
15 journals to acquire?
16 A. The library might have a use study -- again, if
17 the library is having a use study and it's determining the
18 use of faculty and students, separating it, differentiating
19 it by that, then one would want to use both pieces of
20 information, and I believe any librarian would say that you
21 would not want to throw out the use of the students in
22 making a determination.
23 Q. But you could devise a study that only gathered
24 information as to faculty uses?
25 A. Possible.
201
1 Q. And if you did so, that would provide helpful
2 information to the librarian?
3 A. Correct.
4 Q. And a third measure that you have suggested is
5 faculty rankings?
6 A. Yes.
7 Q. And that, by nature, focuses entirely on the
8 faculty?
9 A. Yes.
10 Q. And it focuses on their views as to the journals
11 that contribute most significantly to their work?
12 A. The journals they find most useful, correct.
13 Q. And do you think that's an appropriate measure?
14 A. Yes. Falling short of having use statistics, I
15 think use statistics are actually better than that
16 particular measure.
17 Q. But it's appropriate in the sense that it
18 provides a reliable basis for management decisions by
19 librarians, faculty rankings?
20 A. Correct. It's not -- again, it's not the first,
21 best choice, but it's a useful number, falling short of
22 having use numbers.
23 Q. Citation data is a comprehensive faculty ranking,
24 is it not?
25 A. It's a list of how many times each journal is
202
1 getting cited by the entire universe of faculty, clearly not
2 by just the faculty within your library or that are served
3 by your library.
4 Q. It reflects the considered judgment of the most
5 productive members of the faculty as to the journals and
6 articles that contribute most significantly to their own
7 research?
8 A. Well, its citations include everyone's citations;
9 not only the most productive we are talking, but the
10 faculty.
11 Q. I was using it for the faculty members who
12 actually write articles.
13 A. Oh, OK.
14 Q. With that clarification, can you respond to the
15 question?
16 A. That for all the -- the faculty that are writing
17 articles and citing all the entire body of faculty citations
18 show how many times they are citing things.
19 Q. With respect to the five possible measures of
20 effectiveness, you have stated that different people have
21 different views as to which of the five measures is the
22 best, correct?
23 A. Correct.
24 Q. And indeed, you noted that Professor Barschall
25 believed that the impact factor was the best measure of
203
1 effectiveness; is that correct?
2 A. Correct.
3 Q. And are you familiar with Professor Barschall's
4 career?
5 A. As a physicist? No.
6 Q. As a physicist.
7 A. I am not.
8 Q. You know that he had a long and distinguished
9 career as a physicist?
10 A. My understanding is that he did.
11 Q. Received many honors for his contributions to the
12 field of physics?
13 A. That's my understanding.
14 Q. Would you be inclined to give some deference to
15 his views as a physicist to what is a reasonable measure of
16 effectiveness of journals?
17 A. No. If he received those same honors in
18 economics and information science, I would give deference;
19 and to the same degree that I don't think he would respect
20 my opinion on physics too much, or should he.
21 Q. You have never cited a physics article, have you?
22 A. Not that I can recall.
23 Q. Dr. Barschall did that many times, did he not?
24 A. Yes.
25 Q. He knew what was involved in that process?
204
1 A. In citing a physics article, yes.
2 Q. And do you know that he had examined the
3 relationship -- he had examined citation data, had he not?
4 A. Yes.
5 Q. And he concluded that it was a reliable measure;
6 is that true?
7 A. He concluded that, yes.
8 Q. Is it true that you do not know very much about
9 impact factors?
10 A. No.
11 Q. Not true?
12 A. No.
13 Q. In what connections have you worked with impact
14 factors?
15 A. In the connection, obviously, of this case, prior
16 to this, with my work on the economics of information, that
17 it would be simply part of the entire universe of
18 information science that one is aware of.
19 Q. Have you ever used impact factors in one of your
20 own analyses?
21 A. No.
22 Q. Have you ever had occasion to study impact
23 factors and how the data is presented and examine it for
24 different journals?
25 A. Have I looked at the impact factor numbers?
205
1 Q. Apart from in connection with this case, have you
2 had occasion to work with impact factor data?
3 A. I've seen the impact factor numbers in the
4 Science Citation Index prior to this case, but I have not
5 worked with those numbers.
6 Q. What about citation data? Have you worked with
7 citation data?
8 A. Again, I have seen the citation numbers, but I
9 have not worked with them in research. I have also -- am
10 familiar with research that used citation numbers and impact
11 factors, but I have not used citation numbers myself or
12 impact factor numbers.
13 Q. You mean the articles relating to citation data?
14 Is that what you're referring to, what you are familiar
15 with?
16 A. Articles relating to citation data.
17 Q. And you looked at those articles in connection
18 with this case?
19 A. Well, I looked at some of those articles prior to
20 this case, too.
21 Q. It's true that even when you collected the
22 citation data to present to the Tenure Review Committee, you
23 didn't gather that data yourself, did you?
24 A. Correct.
25 Q. You sent someone else to do it?
206
1 A. I had a graduate student do it.
2 Q. Could you look at Plaintiff's Exhibit 3 again,
3 the Physics Today article.
4 In the first paragraph, do you see the statement
5 at the end of the first paragraph, after Professor Barschall
6 refers to the cost per thousand characters and then to the
7 impact factor? He then states as follows: "The ratio of
8 these two measures is perhaps the best indicator of a
9 journal's cost effectiveness." Do you see that statement?
10 A. Yes.
11 Q. Do you disagree with that statement?
12 A. Yes.
13 Q. Apart from that statement -- strike that.
14 Does a similar statement appear in the Bulletin
15 of the American Physical Society?
16 A. Yes. On the top of page 1438, under "Cost per
17 Impact," the second sentence.
18 Q. And there it's just slightly different. It
19 reads, "This ratio is perhaps the most significant measure
20 of the cost effectiveness of the journal"?
21 A. Correct.
22 Q. Referring again to the same ratio?
23 A. Yes.
24 Q. And you disagree with that statement?
25 A. Yes.
207
1 Q. Am I correct that, apart from those two
2 statements, there are no statements in either PX 2 or PX 3,
3 whether express or implied, that in your judgment are false?
4 A. I would say that the implication of these
5 articles is that this information can be used for collection
6 development, which I think is a false implication.
7 Q. Apart from that, there is no statement in either
8 article that in your judgment is false; is that correct?
9 A. (Pause) I would say that that implication,
10 actually, was borne out in the last paragraph of the Physics
11 Today article.
12 Q. I'm not sure if you're addressing my question.
13 A. Oh, the implication -- I said that the
14 implication of this having influence on collection
15 management, it's false that one should use this kind of
16 procedure for collection development in a library, and
17 you're saying is there any other statement --
18 Q. Any other statement, express or implied, in
19 either article that in your judgment is false.
20 MR. PLOTZ: With respect, I think Dr. Kingma was
21 answering the question.
22 A. I guess what I'm -- in trying to narrow down
23 that, in turning to the last paragraph here, there are a
24 couple of statements that I would disagree with Professor
25 Barschall on.
208
1 Q. What document are we on?
2 A. We are in Physics Today, the last paragraph.
3 It's really just --
4 Q. Can we -- you'll have an opportunity --
5 MR. PLOTZ: He's in the middle of an answer,
6 Judge.
7 MR. HUVELLE: He's not answering my question,
8 though.
9 THE COURT: Let him complete his answer. The
10 question was that, apart from those two statements and the
11 implications derived from those two statements, is there
12 anything else in these two articles that you regard as being
13 false?
14 THE WITNESS: I think the implications of Table
15 2, that one that should compare publishers on this measure,
16 again, is a false comparison more making collection
17 development decisions.
18 Q. Anything else?
19 A. No. I believe that's it.
20 Q. Do you see -- there are no other statements,
21 express or implied, in either article that in your judgment
22 are false? Correct?
23 A. Correct.
24 Q. You were deposed in this case on April 18?
25 A. Correct.
209
1 Q. As of that date, you had spent between 40 and 80
2 hours working on this matter?
3 A. Since April 18?
4 Q. No. Before that.
5 A. Oh, before that date. Correct.
6 Q. How many hours have you spent on this matter
7 since April 18?
8 A. Somewhere between 10 and 20.
9 Q. When did you resume work on this matter?
10 A. Worked briefly on it over the weekend -- it was
11 probably a couple of days before that -- looking at some of
12 these issues, and then obviously within trial right here.
13 Q. Have you done any analysis in connection with
14 this matter since April 18?
15 A. Analysis of the numbers?
16 Q. Any further analysis that goes beyond what you
17 had done as of April 18?
18 A. No.
19 Q. Did you review any additional articles or other
20 written materials that you had not seen prior to April 18?
21 A. No.
22 Q. With whom have you spoken about this matter since
23 April 18?
24 A. About this matter, the attorneys, Mr. Plotz,
25 Mr. Lupert, the other attorneys. I believe that's it.
210
1 Q. Am I correct, then, that you have not learned
2 anything new regarding this matter except -- since April
3 18 -- except what was told to you by the attorneys?
4 A. Correct.
5 Q. I would like to show you a copy of Defendants'
6 Exhibit CCC. Could you identify this document?
7 A. This is a report I prepared for this case.
8 Q. Did you testify on April 18 that it was a full
9 and accurate statement of the opinions to which you expected
10 to testify?
11 A. Yes. I did make a couple of corrections on it.
12 Q. There were two minor corrections that you made?
13 A. Correct.
14 Q. Those were on page 7?
15 A. Correct, along with a clarification of the
16 benefit effectiveness efficiency issue that we spoke of at
17 the deposition.
18 Q. Apart from those corrections, this report, you
19 testified, was a full and accurate statement of the opinions
20 that you held and about which you expected to testify?
21 A. Correct.
22 Q. In your testimony yesterday, you asserted that
23 Professor Barschall erred in comparing journals of different
24 types; namely, review journals, letters journals, and
25 archival journals, and also journals in different physics
211
1 disciplines. Do you recall that testimony?
2 A. Correct.
3 Q. That was one of the major criticisms that you
4 articulated yesterday?
5 A. Correct.
6 Q. Can you refer me to the place in your report
7 where you discuss that major criticism?
8 A. On page 5, No. 6, where it talks about variations
9 in citation rate related to type of publication,
10 nationality, time period, and size and type of specialty.
11 Q. Is there anything about review journals?
12 A. No. It just says "variations in type of
13 publication." It doesn't specific -- you're right, it
14 doesn't specifically mention review journals.
15 Q. And that was subparagraph 6?
16 A. 6 on page 5, right.
17 Q. Now, on page 1 of your report, you refer -- and
18 I'm going back to the correction that you made during your
19 deposition --
20 A. OK.
21 Q. You referred to a cost/benefit ratio. Do you see
22 that?
23 A. Yes.
24 Q. You agree now that Dr. Barschall did not purport
25 to prepare a cost/benefit ratio?
212
1 A. Correct.
2 Q. And that that is a different kind of analysis?
3 A. Related but different kind, correct.
4 Q. Different from cost effectiveness, which is what
5 he did?
6 A. Correct.
7 Q. And so you -- all of your criticisms with regard
8 to the adequacy of the cost/benefit ratio are misplaced; is
9 that correct?
10 A. No. It's, as I said at the deposition, even if
11 you substitute the word "effectiveness" for "benefit" here,
12 all the criticisms remain. It's just a matter of using the
13 word "benefit" instead of "effectiveness."
14 Q. Why did you use the wrong term there?
15 MR. PLOTZ: Objection.
16 A. Well --
17 MR. PLOTZ: Objection to the characterization,
18 your Honor.
19 THE COURT: Overruled.
20 A. Cost/benefit analysis is a term in the common
21 vernacular of economists and those who work with economists,
22 and it's actually sort of the encompassing of all the types
23 of analysis. Cost-effectiveness is a piece. If you think
24 about a Vien diagram with the largest circle on the outside
25 being cost/benefit analysis, cost-effectiveness analysis
213
1 falls within that, and cost-efficiency analysis falls within
2 that circle.
3 Cost/benefit, using that term is really more of a
4 sort of global term, which included many things that it
5 shouldn't have. But it's -- I think it's one that's
6 throughout the sort of economics vernacular. You know, this
7 is "echo-speak" in some sense that people understand, using
8 the term cost/benefit.
9 Q. Did you analyze, in your report, did you analyze
10 his work with reference to the standards for cost-
11 effectiveness studies or cost/benefit studies?
12 A. As I said, since "effectiveness" falls within the
13 rubric of "benefit," it in some sense would be both,
14 although I didn't analyze it with respect to him estimating
15 the dollar value of a journal subscription.
16 Q. But in some senses it would be different?
17 A. Right. The difference is inserting the term
18 "dollar."
19 Q. The difference is also in terms of the
20 comprehensiveness of the analysis of cost effectiveness?
21 A. Correct.
22 Q. My question is, when you judged the article for
23 purposes of the report, did you use the wrong term or the
24 wrong analysis?
25 A. It would have been more correct for me to use the
214
1 effectiveness term, the cost-effectiveness term.
2 Q. But do you agree that you also used the wrong
3 analysis in judging his report?
4 A. No.
5 Q. Can you look at the second paragraph on page 2?
6 A. Yes.
7 Q. Do the statements in that paragraph apply if
8 you're judging his work in terms of cost-effectiveness?
9 A. No. That statement refers to cost/benefit,
10 right.
11 Q. So that's an indication that in fact you were
12 judging his work by -- in terms of the wrong analysis. You
13 were applying the wrong economic construct to his work.
14 A. Just for that -- those single two senses, you can
15 say that, but for the rest of the report, no.
16 Q. Are you saying that, for most of the report, you
17 analyzed it in terms of the cost effectiveness study, but
18 for those, that one paragraph, you deviated and analyzed it
19 in terms of cost/benefit?
20 A. Well, as I said before, cost-effectiveness --
21 cost/benefit is sort of the big circle and within that is
22 cost effectiveness. So analyzing it by the standards of
23 cost-effectiveness by definition fall within a cost/benefit
24 analysis.
25 Q. But a perfectly proper cost effectiveness
215
1 analysis may -- indeed will -- fall short of a proper
2 cost/benefit analysis --
3 A. Right.
4 Q. -- necessarily?
5 A. Yes.
6 Q. And there is no doubt in your mind that Professor
7 Barschall had done a cost-effectiveness study?
8 A. Correct.
9 Q. And also that major components of it were a cost
10 efficiency study?
11 A. By definition, you can't do a cost effectiveness
12 without cost efficiency.
13 Q. So paragraph 2 is a mistake on your part?
14 A. No, it's -- I mean, the -- both sentences in
15 paragraph 2 are correct.
16 Q. If you're talking about a cost/benefit study.
17 A. Correct.
18 Q. Which we are not.
19 A. Correct.
20 Q. Paragraph 3 on that page, am I correct in
21 understanding that the principal concern expressed there
22 with respect to Barschall's cost analysis is that a journal
23 might add pages and pages of meaningless numbers and
24 therefore lower its cost per impact; is that what you're
25 saying?
216
1 A. Correct.
2 Q. Do you think that Professor Barschall assumed
3 that his colleagues in the field of physics would not add
4 pages and pages of meaningless data in order to affect the
5 cost-per-impact ratio?
6 A. I don't pretend to know what he assumed.
7 Q. It would have been reasonable of him to make that
8 assumption?
9 A. It's not always his colleagues in the field,
10 though. It's also the publishers that are making that
11 decision about what is included with an article, whether or
12 not data might be included or additional pages might be
13 included.
14 Q. Do you believe that the points you make in this
15 paragraph are a serious criticism of his work?
16 A. Yes.
17 Q. Can you look at page 3 of your report, the first
18 full paragraph consisting of two sentences.
19 A. Yes.
20 Q. Do the statements in that paragraph have any
21 relation to a cost effectiveness study?
22 A. It relates to benefit, so in this case you can
23 substitute the word "effectiveness" for "benefit" and the
24 statements are true.
25 Q. But do the statements have any relationship to
217
1 your analysis of Dr. Barschall's cost-effectiveness study?
2 A. Yes, because I believe effectiveness is more
3 appropriately measured by demand or use than by impact or
4 citation.
5 Q. Paragraph 4 -- I'm sorry, page 4. You criticize,
6 starting on page 4, citation data; is that correct?
7 A. Correct. I paraphrased from an article by
8 MacRoberts and MacRoberts.
9 Q. You copied the seven points that MacRoberts and
10 MacRoberts made?
11 A. Right.
12 Q. You didn't add anything new?
13 A. Correct.
14 Q. Is it true that many of the criticisms that
15 MacRoberts and MacRoberts had suggested pertain only if
16 you're using citation data to evaluate individual authors,
17 not if you're using them to evaluate journals; is that
18 correct?
19 A. I don't know if that's correct, because you can't
20 make statements about the degree to which these problems
21 influence different journals, or at least I'm not familiar
22 with any research that has been done that have looked at the
23 differences in citation or the error rates in citations and
24 the possible differences between different journal titles.
25 Q. Well, one of the identified flaws is that an
218
1 author may not include his middle initial, and therefore in
2 terms of citation data the article may not be attributed to
3 the right person. That doesn't have anything to do with the
4 assessment of journals, does it?
5 A. Can you point to that on this report?
6 Q. Page 6, Subpart B.
7 A. Correct.
8 Q. The same is true of A and C?
9 A. Correct.
10 THE COURT: How much longer will you be on
11 cross-examination?
12 MR. HUVELLE: It will be a while.
13 THE COURT: An hour?
14 MR. HUVELLE: Maybe.
15 THE COURT: Half an hour?
16 MR. HUVELLE: Between that.
17 THE COURT: We will take a recess.
18 MR. HUVELLE: Thank you.
19 THE COURT: Off the record.
20 (Discussion off the record)
21 (Recess)
22 BY MR. HUVELLE:
23 Q. Dr. Kingma, just a couple more questions on your
24 report. Can you turn to page 7.
25 At the end of the first paragraph, with regard to
219
1 impact factors, you state, "This truncation of the year of
2 publication and the years of citations biases the measure of
3 impact." Do you see that?
4 A. Yes.
5 Q. The use of an impact factor precisely records the
6 data for the period that it purports to cover; is that true?
7 A. Correct.
8 Q. And in the next paragraph, you make a couple of
9 errors in your description of the impact factor, in the
10 same -- in the following paragraph. Do you see that?
11 A. Yes. That's with my corrections, when my
12 corrections came up.
13 Q. Did you misunderstand how the impact factor was
14 calculated when you wrote this?
15 A. No.
16 Q. The changes would be in the first sentence, the
17 year 1986 should read 1984 and 1985?
18 A. Right. It's correctly described in the paragraph
19 above that.
20 Q. And then the next sentence, again, the year 1986
21 should read 1984 and 1985?
22 A. Correct.
23 Q. And in the next sentence, "a single year" should
24 be "two years?"
25 I'm sorry. "Two years" should be "one year."
220
1 A. Right.
2 Q. And "one year" should be "two years."
3 A. Correct.
4 Q. Page 9, you talk about the problems in counting
5 pages.
6 A. Correct.
7 Q. Let me just ask you to look at a copy of Physical
8 Review B.
9 This has been marked as Exhibit XXXXX, is that
10 correct?
11 A. Correct.
12 Q. And this is a journal published by the American
13 Physical Society? I will represent that to you. And it
14 says so on the front page. And it's one of the journals
15 included in Dr. Barschall's study.
16 The concern you had about blank pages does not
17 apply to this journal, does it?
18 A. It would appear with the issue in my hand that
19 there aren't any completely blank pages that carry page
20 numbers. There are obviously half-blank pages, where one
21 article ends and the next article begins on the next page.
22 For example, on page 4699, half the page is blank, whereas
23 half the page carries the end of the article.
24 Q. But it's not the case that there are full blank
25 pages?
221
1 A. That are numbered. There is a blank page but
2 it's not numbered, near the end.
3 Q. Let me show you, for example, Ferroelectrics,
4 Exhibit VVV, just one example there. There's a blank page
5 that's numbered.
6 A. Correct.
7 Q. So the journals that would be advantaged by this
8 methodology would be one such as Ferroelectrics that number
9 blank pages, correct?
10 A. The journal -- it's correct that the journals
11 that would be advantaged by this methodology are ones that
12 carry blank pages. I don't know if these two issues are
13 representative of Ferroelectrics and Physical Review B.
14 Q. Well, I'm not going to ask you to -- we'll let it
15 stand there.
16 But in any event, no matter whose advantage, it's
17 not a very important point in terms of the scope of
18 Dr. Barschall's study, is it?
19 A. It depends on how many pages were blank or how
20 many pages were recognized. It's also how many pages carry
21 Roman numerals.
22 Q. Can you turn to page 10 of your report, Character
23 Counting Bias.
24 You refer to Dr. Barschall's methodology of
25 counting a page which contains figures or tables as if it
222
1 were completely filled with text. Do you see that?
2 A. Yes.
3 Q. And you state that this methodology inherently
4 discounts the value of figures, tables, and display
5 mathematics; is that correct?
6 A. Correct.
7 Q. But your statement is not correct, is it?
8 A. Pardon me?
9 Q. Your statement is not correct?
10 A. That this bias is the cost per impact measure
11 against journals?
12 Q. That it inherently discounts the value of
13 figures, tables and displays. What it does is treat a page
14 containing figures, tables and displays as equivalent to a
15 page of text?
16 A. It treats them as equivalent in size, even though
17 they might have different production costs.
18 Q. But production costs should have nothing to do
19 with Dr. Barschall's analysis?
20 A. Well, the price is part of his analysis.
21 Q. And the price is the price and it doesn't matter
22 what the production cost is, does it?
23 A. Oh, it -- part of his analysis, he separated out
24 translation journals for those production costs, so --
25 Q. Isn't it true that translation journals are
223
1 journals that rarely have impact factors?
2 A. To my knowledge that's true.
3 Q. So he could not have done a cost-to-impact factor
4 ratio for translation journals, or he would have to do it
5 for such a limited sample that it wouldn't do much good?
6 A. I'm not sure that's the reason he gave for
7 separating out translation journals.
8 Q. But it is a fact that he -- he did not have the
9 data to do the full analysis with respect to translation
10 journals?
11 A. I don't know if he had the data or not for
12 translation journals or how many translation journals he had
13 data on.
14 Q. But in any event, for purposes of Dr. Barschall's
15 analysis, what's important to the librarian is the price,
16 not how much it costs the publisher to compile certain
17 pages?
18 A. The price is an important component of the
19 analysis for a librarian.
20 Q. It's price and effectiveness?
21 MR. PLOTZ: He hadn't finished, your Honor.
22 A. The tables and display mathematics, one would
23 assume, have bearing on the effectiveness, have some
24 influence on the quality of information within there. Now,
25 if you remove it -- as we said before, this is any cost
224
1 effectiveness, cost/benefit, cost efficiency analysis -- not
2 cost efficiency. Cost effectiveness, cost/benefit analysis
3 has two parts to it, the cost part and the effectiveness
4 part, and while you can take table and display mathematics
5 and pull it out of the cost part, it winds up in the
6 effectiveness part.
7 Q. This is not pulling it out of the cost side, is
8 it?
9 A. His methodology is, by making an equivalence
10 between characters and a table of the same size.
11 Q. It's assigning a value to mathematical tables?
12 A. Correct.
13 Q. Now, you may think mathematical tables have a
14 value that's higher than text or lower than text, but I
15 don't know -- do you have a way of assigning a value to
16 mathematical tables in comparison to text?
17 A. He used a formula of one to one, basically.
18 Q. Right.
19 A. And he could have used a formula of 100 to 1
20 or --
21 Q. 1.2 to one.
22 A. -- 1.2 to one. It's just an arbitrary pick.
23 Q. But you don't know if he discounts the value of
24 tables. All you know is he treats it as equivalent to text.
25 A. He treats it as equivalent to text, correct.
225
1 Q. And can you look at Phys. Rev. B. Isn't it true
2 that that journal, for example, has many tables, display
3 mathematics, and figures?
4 A. Correct.
5 Q. You haven't looked at the journals in the study
6 so you don't know whether or not it's true that all of them
7 have such figures?
8 A. Correct.
9 Q. With regard to the including of journals, physics
10 journals in different disciplines, review journals and
11 letters journals, did you say that that is like comparing
12 apples to oranges?
13 A. Correct.
14 Q. What did you mean by that?
15 A. Well, there are different -- they are different
16 types of journals and carry with them different citations
17 based on the type of journal it is.
18 Q. And you know that because you have compared them?
19 A. Because I have seen the numbers on them in
20 reading the literature, correct.
21 Q. Not all the --
22 A. I think -- I think -- I'm sorry. But I think
23 that even Professor Barschall points to this in one of his
24 articles.
25 Q. In fact, that's where you learned it?
226
1 A. No, I wouldn't say that I learned it from there,
2 but it's one source of --
3 Q. Have you ever read a review journal in physics?
4 A. No, I have never read physics journals.
5 Q. Do you have any personal knowledge as to the
6 citation rates for these different kinds of journals?
7 A. From economics, not from physics.
8 Q. In your readings of economics, they discussed the
9 citation rates of review journals in physics?
10 A. No. My readings of the citation rates of letters
11 journals and economics and original research journals and
12 economics.
13 Q. Which you have done in connection with this case?
14 A. No, which was prior to this case.
15 Q. Is it true that not all review journals listed by
16 Dr. Barschall have higher impact factors than archival
17 journals?
18 A. I don't know. I would have to look at the table.
19 Do you want me to look at the tables?
20 Q. Yes. Table 1 of PX 3.
21 A. Correct. There are some review journals with
22 lower impact numbers than some original research journals.
23 I do not know -- I have to admit I do not know, under the
24 review journals, what specialty or subspecialty these
25 particular review journals might be reviewing.
227
1 Q. I believe you noted yesterday that Review of
2 Modern Physics has an impact ratio of -- impact factor of
3 27; is that correct?
4 A. For 1986, correct.
5 Q. Right. And that's more than three times higher
6 than any other review journal?
7 A. I believe so, correct.
8 Q. Do you think that's a fact that a librarian ought
9 to know in connection with acquisition decisions, that
10 Review of Modern Physics has a very high impact factor?
11 A. No.
12 Q. Doesn't that number tell the librarian that the
13 articles in the Review of Modern Physics are relied upon
14 with extremely high frequency by scientists?
15 A. It tells the librarian something about how
16 physicists as a whole cite articles in the past two years
17 from Review of Modern Physics, but that librarian has to be
18 more concerned about how her or his particular set of
19 patrons might use the Review of Modern Physics, and it might
20 be that they use it a lot, but it also might be that they
21 have better use of another journal.
22 Q. It suggests that if they -- even if they are
23 reading other articles, that an article -- a journal that
24 will be frequently cited in other journals is Review of
25 Modern Physics; it tells you that, doesn't it?
228
1 A. Correct.
2 Q. It tells you that Review of Modern Physics is the
3 Michael Jordan of physical journals, doesn't it?
4 A. I don't know if I would make that comparison.
5 Q. Well, 27 impact factor, the next one 7 in the
6 same category; isn't that like scoring 63 points in a
7 playoff game?
8 A. It certainly is more than three times higher than
9 the next journal.
10 Q. You can say you can't compare Michael Jordan to
11 other basketball players, can't you? People say that, but
12 you can?
13 A. They are using it in a different context,
14 obviously.
15 Q. You can make a comparison. The comparison is,
16 he's better. Right?
17 A. Correct.
18 Q. And that's what this data tells you, as to
19 physical -- Review of Modern Physics, that, as compared to
20 other physical -- physics journals, it is cited much more
21 frequently, and a librarian might want to know that; is that
22 correct?
23 A. I guess that's where the analysis sort of
24 crumbles, and that we can say Michael Jordan is better in
25 terms of all basketball players and we can say the Review of
229
1 Modern Physics is more cited in terms of all these physics
2 journals. But can we say that the Review of Modern Physics
3 is the best journal for our patrons in our library? That's
4 a different question. Can we say that Michael Jordan is the
5 best basketball player for our basketball team? I don't
6 know. The length of magic might be might be very broad
7 compared to any other player.
8 THE COURT: You said that the impact factor of 27
9 was not relevant to the librarian whose concern should be
10 the needs of the patrons of the --
11 THE WITNESS: Their library.
12 THE COURT: -- particular library.
13 THE WITNESS: Right.
14 THE COURT: Well, if we are dealing, for example,
15 with a large university, isn't it reasonable to assume that
16 the interests of the users of that library will bear some
17 relationship to the world of physicists as a whole?
18 THE WITNESS: It depends on what specialties that
19 university might support.
20 THE COURT: If you are dealing with a library for
21 a commercial entity which is engaged in some very refined
22 form of physical research, that might not be relevant, but
23 for let's just say university libraries, would it not be
24 relevant for a librarian in a university library to know
25 that the Review of Modern Physics is more cited than any
230
1 other such review?
2 THE WITNESS: Let me go back to the article by
3 Bensman which showed that there wasn't a correlation or a
4 very low correlation between impact factor and faculty
5 ranking of journals. And clearly within any discipline,
6 there might be some large general journals that are
7 important, but even universities are becoming very
8 specialized, each and every department, and with those
9 specialties there is typically a core of niche journals that
10 support each specialty.
11 THE COURT: Without meaning to disparage
12 academia, for which I have great respect, isn't one of the
13 needs which a researcher has to have sources that can be
14 cited for the particular propositions which are discussed in
15 the article? So isn't the citation itself a form of use of
16 the publication?
17 THE WITNESS: Right. You would have to use it in
18 some sense to cite it, although sometimes people might cite
19 things just because they read them in other things.
20 THE COURT: Sometimes you pad citations, like --
21 THE WITNESS: Right.
22 THE COURT: -- law clerks do when they do draft
23 opinions.
24 THE WITNESS: The trick -- the trick is, though,
25 that --
231
1 THE COURT: But you seem to be making a
2 distinction between citation and use.
3 THE WITNESS: They are two different things.
4 THE COURT: Well, use is the greater, but isn't
5 citation a form of use?
6 THE WITNESS: Yes. What citations recognize,
7 too, is a global measure of how much everybody cites it.
8 Use within your library is a local measure of how much
9 people -- the patrons you support are going to use a
10 particular journal.
11 THE COURT: Let me just -- with respect to
12 general libraries such as in universities, is it reasonable
13 to assume that the interests or needs of a person who is to
14 utilize the library will bear a relationship to the
15 interests and needs of physicists at large?
16 THE WITNESS: Actually, universities over the
17 last two decades have become more and more specialized, so,
18 no, it's not -- you can't make that comparison anymore.
19 Universities' libraries just simply can't buy everything
20 that they want, and universities can't support all
21 specialties that exist. So many of them tend to limit that
22 and decide to collect a set of faculty that have a certain
23 niche themselves.
24 At the university at Albany, for example, the
25 Economics Department has a handful of faculty that
232
1 specialize in public finance and a handful of faculty that
2 specialize in econometrics, but we don't have -- we don't
3 have a lot of economists that do agricultural economics or
4 consumer economics. We have one that does labor economics
5 but not more than one.
6 So we -- university departments are becoming more
7 and more specialized.
8 THE COURT: But how about at the undergraduate
9 level? Is the nature of these publications such that the
10 undergraduate level is not so relevant?
11 THE WITNESS: The undergraduate level is not as
12 relevant because of -- you know, the journal publications
13 themselves report research that is digestible to faculty and
14 graduate students. Some undergraduates certainly use those
15 journals. But it's, you know, it's -- I would say
16 undergraduates don't go off and do research and do research
17 to the degree necessary to have academic journals constantly
18 at their disposal.
19 THE COURT: But they write papers?
20 THE WITNESS: They do write papers. They do
21 write papers, yes.
22 THE COURT: Which gets them in the library, one
23 hopes?
24 THE WITNESS: Yes, one hopes.
25 MR. HUVELLE: I just have a couple of minutes.
233
1 BY MR. HUVELLE:
2 Q. The point at which you took issue with me, I
3 believe once again, is you said that the librarian is going
4 to look at it differently.
5 A. Look at the impact number differently, correct.
6 Q. And am I correct in understanding from your
7 testimony that you have emphasized the need of the librarian
8 to look at the local community?
9 A. Correct.
10 Q. And I'm going to use a word that's sometimes used
11 pejoratively, but I think it's the precise word. Your
12 perspective, or the perspective you advocate for the
13 librarian, is a parochial one?
14 A. Meaning --
15 Q. Local.
16 A. -- a local one, yes.
17 Q. And of course libraries, even public libraries,
18 often buy books based upon The New York Times best seller
19 lists or a more global perspective of what's developing in
20 the world of literature or whatnot, correct?
21 A. Correct.
22 Q. And Dr. Barschall was a respected physicist at
23 the University of Wisconsin, which is a major center for
24 studies in physics? He was a member of the National Academy
25 of Sciences, correct?
234
1 A. Correct.
2 Q. And he thought it was important that librarians
3 know the global perspective of what journals were important
4 to the scientists who are writing articles and developing
5 knowledge in the sciences; is that correct?
6 A. He thought that was important, correct.
7 Q. And journals historically have played the key
8 role in transmitting new knowledge from generation to
9 generation in terms of science?
10 A. Correct.
11 MR. HUVELLE: No further questions.
12 THE COURT: Redirect?
13 MR. PLOTZ: Yes, your Honor.
14 REDIRECT EXAMINATION
15 BY MR. PLOTZ:
16 Q. Dr. Kingma, you have been asked a lot of
17 questions about impact factor, a lot of questions about
18 cost.
19 With respect to the market for library journals,
20 you were asked whether this was an example of an imperfect
21 market or market with imperfect information?
22 A. Correct. I was asked.
23 Q. What does it mean for a market to have imperfect
24 information?
25 A. Imperfect information means that either the
235
1 seller or buyer does not have complete information about the
2 transaction or good. Usually we talk about it as the
3 consumer not having complete information about the product
4 that a seller is offering.
5 Q. And just speaking generically, what is a solution
6 for a market which has imperfect information?
7 A. Generically, to provide accurate information
8 about the quality, typically the quality of the product.
9 Q. Is it your view that the ratio of cost per
10 character over impact factor provides additional information
11 that helps solve the problem of an imperfect market?
12 A. No. The appropriate term here is having correct
13 information. Adding correct information to a market
14 improves the efficiency of that market. Likewise, adding
15 misinformation to that market may actually push that market
16 away from efficiency, in economic terms. I don't consider
17 this cost per character per impact measure to be correct
18 information for managing -- making collection development
19 decisions within libraries.
20 THE COURT: I take it that applies as well to the
21 decision by the author, at least in Exhibit 3, the Physics
22 Today, the pitch, if I may use that colloquial term, is to
23 authors rather than librarians. In the last -- I don't have
24 it. Isn't the last paragraph an exhortation to authors
25 rather than librarians?
236
1 THE WITNESS: Yes.
2 Q. Do you know, by the way, whether the journals
3 which have lower ratios of cost per impact have the capacity
4 to publish additional articles that authors might submit?
5 A. I don't know.
6 Q. Do you know whether anyone before Dr. Barschall
7 performed a cost comparison survey based on a ratio of cost
8 per character over impact factor?
9 MR. HUVELLE: Your Honor, I think we are outside
10 the scope of the cross.
11 THE COURT: No, I will allow it.
12 A. I don't know.
13 Q. You are not aware of --
14 A. I am not aware of other articles, prior to this.
15 Q. You testified on cross, with respect to the cost
16 side of the ratio, that a normalization of cost did not
17 provide useful information, correct?
18 A. A normalization based on cost per page per
19 character per article analysis, correct.
20 Q. And why is that?
21 A. It's simply not the purchasing decision that the
22 librarian is making. A librarian has to be concerned about
23 how much they are paying, and that is paying for the
24 subscription as a whole, and how much use their patrons are
25 going to receive from that. And clearly that's use for the
237
1 subscription as a whole.
2 Q. Now, with respect, for example, to normalizing
3 price per page -- you were asked some questions about
4 that -- why doesn't that provide useful information to a
5 librarian making an acquisition decision?
6 A. I don't think normalizing by price per page
7 provides useful information to a librarian. It's not --
8 it's not the relevant decision for a librarian who is
9 purchasing a subscription as a whole.
10 Q. In terms of the subscription to a journal -- a
11 journal contains information, doesn't it?
12 A. Correct.
13 Q. And in determining the value of a journal, is one
14 of the things that one must look at the quality of the
15 information that's in the journal?
16 A. Correct.
17 THE COURT: Well, are you assuming that all of
18 the editorial material in all of these journals are subject
19 to a process of peer review?
20 THE WITNESS: No, certainly not -- I mean, the
21 articles typically are subject to peer review but not
22 always. You know, I -- again, I don't know within every
23 physics journal that appears in this which ones were subject
24 to peer review and which were not. And clearly there's some
25 information that was not subject to peer review, like
238
1 editorials and such.
2 THE COURT: Whether a particular journal has peer
3 review or not, is that something which a librarian would
4 consider in making a subscription?
5 THE WITNESS: Personally, I think that those
6 journals with peer review are going to be of higher quality.
7 But, again, a librarian has to be concerned about what their
8 patrons use. And if that means their patrons would rather
9 have a journal that's not peer-reviewed but it's in some
10 specialty that they really need, then they shouldn't make a
11 decision based on the fact that there is another journal
12 that is peer-reviewed.
13 THE COURT: Let's assume that you have two
14 journals, both of which are subject to peer review by
15 referees of comparable knowledge or standing in the field
16 and who apply the same criteria. If you indulge in that
17 assumption, is the number of pages per subscription dollar
18 of greater relevance?
19 THE WITNESS: I don't see in which case they are
20 ever of relevance. I mean, it's how much you're going to
21 pay for the subscription and how much use you're going to
22 get out of it.
23 THE COURT: Isn't --
24 THE WITNESS: There are some --
25 THE COURT: Isn't the use to be made -- doesn't
239
1 it have some relationship to the quality of the editorial
2 matter?
3 THE WITNESS: One could assume that more quality
4 peer review would result in more quality articles and that
5 might result in more use. But again, within, you know,
6 within the local decision making, does that result in more
7 use for my patrons?
8 THE COURT: Similar to the criticism that is made
9 of some political figures who are said not to exercise any
10 leadership but merely to follow popular polls, you are
11 assuming that the librarian's role is simply to cater to an
12 existing demand and that the librarian has no input as to
13 the quality?
14 THE WITNESS: I truly believe that the
15 librarian's role is to cater to, as you say it, existing
16 demand and expected future demand, because they are
17 entrusted with university dollars in order to purchase
18 journal subscriptions that are going to be used by the
19 clientele at that university. And I think that is -- it's
20 important for them in terms of the management decisions.
21 THE COURT: Let me make sure I understand what
22 you're saying. You're saying, if we have two journals
23 subject to the same peer review by equally conscientious and
24 knowledgeable referees, and for the same subscription price
25 one has 50 pages and the other has 500 pages, that that
240
1 price differential is not something which the librarian
2 should consider?
3 THE WITNESS: The price per page, is that what
4 you're saying?
5 THE COURT: Yes, the price per page or
6 subscription price. Yes, price per page.
7 THE WITNESS: Well, the subscription price is
8 important. The subscription price is important. That's
9 key. I just disagree with normalizing it on a per-page,
10 per-character basis. It's the value of the information, and
11 the information can't be divided up.
12 THE COURT: Well, but my hypothetical assumes
13 that there is the same peer review by the same standards in
14 both publications, both journals. Doesn't that give some
15 comfort that the quality level will be comparable?
16 THE WITNESS: If I -- if you can stretch your
17 hypothetical one more assumption and say these are both on
18 the same topic, OK, two journals on the same topic --
19 THE COURT: The same topic.
20 THE WITNESS: Same topic, same everything, same
21 quality of referees. Then I would say, actually, that the
22 economics of that situation means that the one journal won't
23 be sold. I mean, if you're saying that it's the exact same
24 topic, exact same referees, etc., ultimately the market will
25 decide as to which of these journals people will subscribe
241
1 to. And it's a decision in some sense of, you know, how
2 faculty will use -- hopefully how faculty will use those
3 journals.
4 THE COURT: It's also a function of how
5 knowledgeable the market is.
6 THE WITNESS: True. The market needs accurate
7 information to make those decisions.
8 BY MR. PLOTZ:
9 Q. Dr. Kingma, on the cross-examination, there was
10 an analogy of -- the grocery store analogy of pasta and
11 beans. Is considering journals in different topics
12 something like considering a choice between pasta and beans
13 in a grocery store?
14 A. I would say that's analogous. It might even be a
15 stronger difference among the journals, though.
16 Q. Why is that?
17 A. I mean -- well, pasta and beans are both sort of
18 consumable items that, if I'm looking for a side dish that
19 evening, I might eat either pasta or beans. But if I'm a
20 faculty member in one specialty, journals from the other
21 specialty might be of no use or value to me at all, only the
22 journals within my specialty.
23 Q. Would it make sense for a library to buy an
24 inexpensive journal that's seldom used?
25 A. No.
242
1 Q. Would it make sense to buy an expensive journal
2 that's seldom used?
3 A. No.
4 Q. Would it make sense to buy an expensive journal
5 that is frequently used?
6 A. Yes, if there's enough use to justify that
7 subscription price, yes.
8 Q. Just to complete the box, would it make sense to
9 buy an inexpensive journal that's frequently used?
10 A. Yes.
11 Q. Now, in terms of, you said, in response to a
12 question from Judge Sand a few minutes ago, that your view
13 is that the librarian's function is to respond to the
14 current needs of its patrons and the expected future needs,
15 correct?
16 A. Correct.
17 Q. Who will predict what those expected future needs
18 or trends are?
19 A. Typically the librarian is going to be making the
20 decision about --
21 Q. Based on what information?
22 A. Based on the information they have available.
23 Hopefully they have information on use. Hopefully they have
24 information on requests for different journal titles like
25 interlibrary loan, or hopefully they have talked to the
243
1 faculty to determine what journals would be the best buys
2 for their library.
3 THE COURT: Wouldn't the key factor be to see
4 what the curriculum will be for the near future?
5 THE WITNESS: I think that would be -- yes, I
6 mean, that's extremely important. Hopefully that will come
7 out in the expressed wishes of the faculty.
8 THE COURT: My limited contact with academia in
9 law school is that there is an inquiry made by the librarian
10 with respect to what material the faculty members would like
11 to see in the library in the light of what it is that they
12 will be teaching.
13 THE WITNESS: Right.
14 THE COURT: Isn't that standard?
15 THE WITNESS: I don't know if that's standard at
16 every library. That's an important thing that should be
17 done, correct.
18 BY MR. PLOTZ:
19 Q. I want to turn for the moment to the denominator
20 of the ratio impact factor.
21 Citation analysis -- rather, citation count is
22 not the same thing as impact factor, right?
23 A. Correct.
24 Q. In assessing whether or not either citation count
25 or impact factor reliably measures the effectiveness or the
244
1 benefit of a journal, one must consider what citations
2 measure, right?
3 A. Right.
4 Q. And if there are flaws in terms of what it is --
5 or the value of what a citation measures, would that have an
6 adverse impact on the validity of citation count?
7 MR. HUVELLE: I object to the leading nature of
8 this question.
9 THE COURT: Overruled.
10 A. Errors in citation or impact would affect using
11 it, correct.
12 Q. Is that true in a cost/benefit analysis?
13 A. Correct. I mean, given the errors that are known
14 to exist in citation analysis, those errors will produce
15 errors within subsequent analysis that relies on them,
16 pretty much by definition.
17 Q. So they would produce errors in a cost
18 effectiveness analysis?
19 A. Correct.
20 Q. Now, one of the -- I want to direct your
21 attention to one of the articles that you cited in your
22 report, an article by Moed and Vreins. What did Moed and
23 Vreins --
24 THE COURT: Now, aren't you really going entirely
25 beyond the scope?
245
1 MR. PLOTZ: No. Actually, this is going to be
2 directly relevant to the cross with respect to whether or
3 not the flaws in citation analysis are directed to
4 individual authors or to journals as a whole.
5 THE COURT: You may continue.
6 BY MR. PLOTZ:
7 Q. Just to put it in context, though, what issue
8 were Moed and Vreins looking at?
9 A. Moed and Vreins were looking at clerical errors
10 principally, by authors in citing materials. They took a
11 set of journal titles and then looked for whatever errors
12 might have been produced by the authors in those citations.
13 Q. Were there any differences across the journals
14 they looked at in the types of errors that they found?
15 A. They weren't going to find any errors in terms of
16 journal titles, because of the way they collected the data,
17 by establishing a set of journal titles.
18 Q. What errors were they looking at? What types of
19 errors?
20 A. They were looking at errors such as misspelling
21 of names, misspelling of articles, etc. And they found
22 among those that there was roughly a 9-percent error rate
23 among citations.
24 Q. Were there any differences among the journals
25 they looked at in terms of the types of errors that were
246
1 predominant in each journal?
2 THE COURT: When you say "a 9-percent error
3 rate," do you mean an error of sufficient magnitude so that
4 the correct reference was not credited with the citation?
5 THE WITNESS: In errors -- an error, period,
6 meaning misspelling of a word, misspelling of a name.
7 THE COURT: With the consequence that the
8 citation was not credited in the analysis?
9 THE WITNESS: They weren't looking at the
10 ultimate citation database. They were just looking at a
11 collection of articles. So you can't say it goes to the
12 next step, with the consequence.
13 THE COURT: And they were saying?
14 THE WITNESS: That they were errors in the way
15 authors put their citations together; they make mistakes.
16 THE COURT: For our purposes, the significant
17 question is whether the nature of the errors was such as to
18 skew the analysis?
19 THE WITNESS: Correct.
20 THE COURT: Were these errors of such a nature?
21 If someone misspelled your name, puts an E at the end
22 instead of an A, but in the analysis it's clear that it's
23 your book which is being cited so that they, I guess, would
24 call it a hit in today's language, then the error is of no
25 significance?
247
1 THE WITNESS: Right. I don't know the answer to
2 that, your previous question. That is what it comes down
3 to.
4 BY MR. PLOTZ:
5 Q. Leaving clerical errors aside, is there a
6 subjective component to an author's decision to use a
7 citation?
8 A. Clearly, there is subjectivity based on -- as was
9 mentioned before, positive and negative cites carry the same
10 weight. Even though one is -- might not be favorable and
11 the other is favorable, in the final analysis they both
12 count as a hit. There may be, you know, minor cites of some
13 related research which is counted equally to a major piece
14 of work that had major influence on your research. But
15 those, again, are both counted as equal hits.
16 Q. In connection with the question Judge Sand asked
17 you a few moments ago, the relationship of citations and
18 use, are there uses of journals beyond citation?
19 A. Yes, there are.
20 Q. What kind of uses would there be?
21 A. In reading the literature, knowing what is out
22 there, looking for things that might be related to your
23 research, there are, in the physical sciences, there are
24 clearly a lot of uses within the corporate sector that don't
25 result in citations. This is well known.
248
1 There are, you know, obviously, a body of uses
2 beyond simply those that you are going to cite.
3 THE COURT: How much longer do you think redirect
4 will be?
5 MR. PLOTZ: Longer than five minutes.
6 THE COURT: We are adjourned until 2 o'clock.
7 (Luncheon recess)
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
249
1 A F T E R N O O N S E S S I O N
2 2:05 p.m.
3 BRUCE KINGMA,
4 Resumed, and testified further as follows:
5 THE COURT: Good afternoon. You may be seated.
6 REDIRECT EXAMINATION (Resumed)
7 BY MR. PLOTZ:
8 Q. Dr. Kingma, on cross you were asked a whole lot
9 of questions concerning five theoretical measures of
10 effectiveness of journals, do you recall that?
11 A. Correct.
12 Q. And just so we have them in front of us, what are
13 those five theoretical measures?
14 A. The five theoretical measures are readership, or
15 readings; use; survey of faculty opinions; citations; and
16 impact.
17 Q. And is that list you just gave us in any
18 particular order of importance?
19 A. Yes. I regard it to be listed as best to worst,
20 best being readings or use of a journal in terms of
21 measuring effectiveness.
22 Q. Are any of those five measures in your view
23 invalid measures of effectiveness of journals?
24 A. Yes. I consider the impact factor to be an
25 invalid measure of effectiveness. It has serious problems
250
1 inherent in it that don't make it a worthwhile measure.
2 Q. And what, briefly, are those problems?
3 A. The problems that we discussed before, of its
4 dramatic variability from year to year, of it makes it
5 difficult to compare cross subspecialties, the different
6 kinds of types of journals, the inherent problems in
7 citation analysis, and one more, the fact that it's a global
8 measure rather than a local measure.
9 Q. In terms of its year-to-year fluctuations, you
10 were asked some questions on cross relating to the review of
11 modern physics impact factor as reported in the Barschall
12 study, which was 27, correct?
13 A. Correct.
14 Q. Do you know what happened to that impact factor
15 in the next year, 1987?
16 A. In 1987, actually there was a 40-percent decline
17 in the impact factor down to a little more than 16.
18 Q. Now, what is it about citation count that you
19 rank at above impact factor in your list of theoretical
20 measures?
21 A. Well, citation count gives a more complete
22 picture of a journal. In some sense, if you wanted eve |