Features - The Great Citator Debate

Victoria Szymczak is Electronic Information Specialist & Adjunct Professor of Law at Brooklyn Law School Library.   She is primarily responsible for instructional use and collection of electronic materials. Victoria co-teaches an Advanced Legal Research course with her colleague James Murphy, and a Foreign and International Law Research course with her colleague Jean Davis.   On July 21, 1999, she will be moderating an exciting session on Canadian and Mexican legal research at the AALL Annual Meeting in Washington, D.C. Victoria has co-authored two law journal articles and many columns in Law Lines, the newsletter published by the Law Library Association of Greater New York. She received her MS from Pratt Institute in 1995 and a JD/LLM from Duke University School of Law in 1992.


It was a beautiful day in Brooklyn, New York on May 12, 1999. An invitation brought approximately 50 librarians to Brooklyn Law School that day to attend the Great Citator Debate. Many others were turned away due to space limitations. The audience represented legal information professionals from private, government, academic and publishing house libraries. All were concerned with the state of electronic citations services from Westlaw (Keycite) and Lexis-Nexis (the new Shepard's online). It would be the first time that these two competitors would be in the same place at the same time to dispel any myths and misconceptions about their products.

Once our attendees enjoyed some chitchat at the coffee bar, the activities commenced. Westlaw retained seating at the rear of the classroom as requested. Lexis-Nexis folks stole some front row seating. At 9:40 am, Sara Robbins, Director of the Library and Professor of Law at Brooklyn Law School, welcomed our guests and reminded everyone that we were about to embark on an educational endeavor. The floor was then turned over to yours truly.

The Rules

First, I laid down the ground rules. Each side would have 15 to 20 minutes to make their pitch. After laying down this hard and fast rule, I waffled and admitted that I would not govern with an iron fist. If we were involved in an intellectual discussion that aided our understanding of electronic citators, I would not interrupt the discourse for the sake of slavishly keeping to a schedule. Dan Dabny preferred to be peppered with questions throughout his presentation. Sharon Cort asked the audience to hold their questions until she had finished her presentation. Once both company representatives completed their presentation, attendees would be given a chance to ask further questions regarding KeyCite and Shepard's. If time permitted, the session would close with a random citation given to each speaker who would update and verify the chosen case with great spontaneity! In order to decide who went first, I flipped a coin. This is where Dan’s early arrival was instrumental. The person who arrived first was allowed to choose heads or tails on the coin flip. Dan chose tails but lost the flip to Sharon who led us into the intellectual part of the morning.

Representing Westlaw

Dan Dabny
Lee McCallister
(Account Representative)
Bill Benish
(District Manager)
Joni Marra
(District Manager)
Yadana Latt
(Marketing Program Manager)
Alyin Baumin
(Account Representative)

 

Representing Lexis-Nexis

Sharon Cort
Thomas Daniels
(Account Executive)
Christine Angland
(Regional Sales Manager)
Allison Nussbaum
(Account Executive)
Thomas Geodde
(Products Manager)
Donald Lodge
(Marketing Programs Manager)

 

Shepard's & Lexis-Nexis

The approach Lexis-Nexis chose to illustrate their side of the story allowed Sharon to present a methodical exploration of the new Shepard's online via the Lexis-Nexis web product. Many librarians may be interested in what exactly is so new about Shepard's online. Sharon chose 414 US 573 to demonstrate some of the new features of Shepard's.

Historical Treatment

First, the user is asked to choose between Shepard's for validation, or Shepard's for full research. The first provides subsequent appellate history and citing references that have editorial analysis. The latter provides prior history and subsequent appellate history plus all citing references from cases, statutes and a variety of secondary resources. Although Sharon did not make this point during the presentation, the inclusion of subsequent history in new Shepard's does not seem to be limited to one layer of appellate history either above or below your citing case. The existence of negative treatment of your cited case in the session software is still murky. For example, if you "Shepardize" a trial level case that was subsequently vacated by the United States Supreme Court, this negative history is not readily indicated in the session software. It is simply called "the same case." You would need to "Shepardize" the Court of Appeals citation in order to see the negative treatment of the highest court of the land. New Shepard's on the Internet makes the negative analysis by the Supreme Court much more evident, thus eliminating the need for double Shepardizing and double fees.

Customizing Your Displays and Restrictions

New Shepard's permits researchers to select the display, customize restrictions, and choose their default options based on their individual needs. Lexis-Nexis has our default display option set to "unrestricted." To change your Shepard's default display, you should click on the Options hot link at the top of the screen and select your desired default. Other Shepard's display default setting are

  1. show prior history,
  2. show subsequent history, and
  3. show legislative history.

You can also select your results to display in KWIC or FULL. The default is set to KWIC. Once you change the default, it will stick with your password until you decide to change it again.

An "unrestricted" display will provide citations to cases analyzing the issues in the present case, as well as simple string cites. Other possibilities for display include

  1. all positive analysis,
  2. all negative analysis, and
  3. any analysis.

The difference between any analysis and unrestricted is that the any analysis option will display cases that actually discuss the cited case and eliminate string citations. Unfortunately, after checking several citations on Shepard's, it seems that using any analysis will also eliminate citations to secondary resources. It does not seem logical to omit scholarly commentary in the any analysis view and include it in the unrestricted view with the string citations.

Further customization can be accomplished using the Custom Restrictions Form (CRF). This option can be found in the string of hypertext defaults that appear on your Shepard's results screen mentioned in the previous paragraph. The CRF allows you to limit the display by specific treatment (criticized, distinguished, limited, followed, concurring, dissenting, explained, questioned), jurisdictions, law reviews, ALRs, secondary resources, statutes, case reporter headnotes and date. The category called secondary sources is a bit vague. It would be an improvement if the available secondary resources on Lexis-Nexis were readily evident so we knew what was covered. The same comment can be made about the KeyCite product. An interesting point was raised during the morning concerning the limiting feature for headnotes. If you want to use this feature to its fullest potential, you need to know the number of the headnote from each reporter. For example, in 414 US 573 you can limit the headnotes to those from Lawyer’s Edition and from West. Since the headnotes on issues of law are numbered differently in each reporter, it presuppose you were in possession of both versions of the case. Still, the CRF is a strong tool that allows full customization just one click away from your Shepard's results.

I would like to offer a word to the wise about the display in Shepard's. Although the case used for demonstration purposes listed all the results on one web page, this is not always the situation. When I Shepardized five different cases recently, only one of them listed all the hits on one web page. The display I saw most often was limited to ten hits per web page (using Netscape Communicator). This forces the user to click on the next page link which does not always result in an instantaneous response. If your results do not load up all together, you are limited to printing to your stand-alone printer . Otherwise, you will have to print 10 citing references at a time to your attached printer. In addition, the categories of citing references should be clearer. At the top of the screen you are told the total number of citing references. Each category also lists the number of citing references in that category, but the delineation of those categories is lacking. This can be confusing to the novice Shepard's user.

Focus

Another new feature Sharon discussed with the audience was the ability to focus within your Shepard's result. The focus feature allows boolean searching within the results list. Although librarians welcomed this addition, there was some concern that the focus feature could not search the full text of the cases. We were assured by Sharon that the search engine does search through the full text of each element listed in the results. Other members of the Lexis-Nexis team pointed out that the focus feature is and will remain a "free" perk. Researchers will incur a lexsee/lexstat search fee in addition to the Shepard's fees only if they click on the hypertext link to the full version of the listed cases. Several librarians found the ability to do boolean searching provided a significant level of freedom to customize results.

Audience Commentary

When prompted for commentary regarding new Shepard's, we received a sizable showing of hands. Most of the questions that were submitted by E-mail were answered during the session. For example, some librarians were wondering how to change the default option of KWIC to FULL in their Shepard's search. Sharon illustrated the Options feature at the top of the screen to select their preferred viewing method. A couple of notable comments are described below.

Jody Armstrong from Columbia University indicated that she liked the new Shepard's very much but would like to see a feature that indicated the depth of treatment given a particular case. The representatives from Lexis-Nexis suggested that researchers use the "any analysis" feature to ensure substantive discussion of the case in point. Presumably, the editors at Shepard's filter this type of information through their predetermined editorial codes. Simple string citations would be omitted from the display unless "unrestricted" was chosen.

Bill Mills from New York Law School illustrated an ongoing problem with the Shepard's statutory citators. Apparently, Shepard's does not recognize the abbreviations for the statutory titles. As an example, Bill chose the New York Vehicle and Traffic Law title which is abbreviated as NY Ve & Tr followed by the section number. If you run this search on Lexis-Nexis an error message displays indicating that the citation cannot be found. Lexis-Nexis is aware of this problem. It exists in statutory citators throughout the country and they are working on a "fix" for the snafu. Of course, they hope the problem will be resolved very soon. It should be noted that the abbreviations do work on the soon to be departed version of Shepard's on Westlaw. We could not test it on Keycite since it does not yet support statutory citations.

Yale Lansky from the firm of Jackson, Lewis, Schnitzler & Krupman confirmed that the ability to customize restrictions was only available on the Internet version of Lexis-Nexis. Representatives from the company assured the audience that steps were being taken to include these features in the session software. This fact is important for librarians and patrons who may not be using the web product in their libraries.

KeyCite and Westlaw

For those who have never heard Dan Dabny speak, I can best describe his style as effervescent, particularly he describes KeyCite. Dan spent most of his allotted time discussing the philosophy behind the development of KeyCite, and, consequently, spending a small amount of time on the nuts and bolts of the product itself. Dan also went out his way to refute many negative rumors about KeyCite that are percolating throughout the legal information industry. It should be noted that some of the negative remarks mentioned by Dan are reprinted on the Lexis-Nexis Web page under the link, Learn More About the New Shepard's.

What is KeyCite

According to Dan, the term "Keycite-Shepard's cage match" is an unavoidable term in Eagan. KeyCite, however, was in development for many years prior to their impending loss of the Shepard's Citator Service on Westlaw. It was in their best interest to create a service that merged Shepard's, Shepard's Preview, QuickCite and InstaCite into one product that could be as up-to-date as their databases. In addition, their marketing research among information professionals allowed them to incorporate research elements that we, as librarians, thought were important.

The Process

Dan was very adamant about human intervention in the KeyCite process. Cases are retrieved electronically from the courts. They are then loaded onto Westlaw with field searching capability and instantly appear in the KeyCite database at that time. Within four hours, direct history is added by the legal editors at West and includes red flag treatment. These individuals read every case that comes through West contrary to the belief by some that a computer algorithm is responsible for editorial analysis on KeyCite. Finally, these editors add indirect history within 24 to 48 hours from West’s receipt of the case. The author would venture to guess that long and complicated United States Supreme Court cases might be worked over for several days to ensure accuracy. In fact, Dan later pointed out that the 24 to 48 hour time period is what the editors aspire to reach. When the courts are busy or a case is considered to be a landmark decision, the editors may take as much as three to four days to post the analysis of the case.

Development of Editorial Analysis

One of the advancements promoted by West in connection with its citator service is the use of more descriptive language in the editorial treatment of a case. West contends that the Shepard's editorial analysis codes can be a little mystifying at times. In fact, some of the questions submitted via my e-mail made mention of this fact. For example, the use of the letter D for distinguished postings is ambiguous in Shepard's. We are not sure if "distinguished" means different facts, law, results, etc. . . . Further, Shepard's traditionally includes D analysis only for cases that would appear in the same regional reporter. Still, West marketing research indicated that for all the confusion about the editorial codes such as "distinguished," librarians still felt comforted by their existence. KeyCite finally did adopt the use of "distinguished" in their editorial analysis. Under KeyCite, distinguished will mean that the results of similar cases in fact and law, ended in different sides winning their legal argument. In addition, on KeyCite distinguished will cover all jurisdictions.

A large part of the editorial analysis, however, uses the language of the court itself. Legal editors at West attempt to track the language of the court. Some courts use phrases that are particular to that jurisdiction and are recognizable by those that conduct legal research in that region. Examples provided by Dan were the phrases "disapproved of by" and "overruling recognized by" which are used specifically in California courts. This "is red flag language" for West editors. Frankly, it is red flag language for all of us. The only drawback to this method that I can see is that someone from New York might not realize that "disapproved of by" in California is functionally equivalent to being overruled. I believe this is accounted for by West because they list it under negative treatment. Also, common sense would tell a researcher that a phrase such as "disapproved of by" indicates negative treatment that deserves a better look.

Post-Great Citator Debate information leads me to believe that the origin behind the idea that West uses computers to evaluate their case load began with the process they use to identify possible negative treatment. Apparently, West uses a computer program to sort out the hundreds of cases that arrive in Eagan each day. The computer puts likely candidates for negative treatment on the top of the editors’ lists based on textual indicators. Phrases such as "disapproved of by" would probably be a significant textual indicator -- although I am guessing at this point. This program does not eliminate cases. It sorts them. According to West, the sorting program is correct about half of the time. The editors will still read all of the cases and make their determination concerning analysis. My theory, however, does not explain what benefit a sorting program with a mere 50% accuracy rate is to the West editors. Perhaps they can explain this better through their very efficient direct mail program, or let me know if I misunderstood what I was told.

Use of Graphics and Depth of Treatment

After leading us through a discussion of how and why KeyCite was developed, Dan was ready to go online. He used the citation of 530 P2D 589 to illustrate the strengths of KeyCite.

West likes to use of visual cues in KeyCite. The use of flags and stars are designed to quickly clue a researcher into the status of a case. A red flag warns you that the case is no longer good law for at least one of the issues for which it stands. A yellow flag warns that the case has some negative history, but hasn't been reversed or overruled. A blue "H" indicates that the case has subsequent history. West does not have the market on graphical indicators. It seems to me that Shepard's has made use of the red stop sign and the blue "I", among other visual aids, to help researchers orient themselves before the advent of KeyCite.

Arguably, the most helpful graphics that West incorporated into their citator were the green stars to indicate depth of treatment. Any of us who travel or dine in reviewed restaurants adapt readily to the green star system. The more stars, the better! In the case of KeyCite, the more stars a citing case has, the greater the discussion of your cited case. It was developed to help researchers understand the editorial analysis that was created by Shepard's. The greatest number of stars a case can have is four. The least is one. According to West, four stars indicates more than a page discussion of your cited case. Three stars indicates discussion length of more than a paragraph and less than a full page. A two star treatment usually has a paragraph discussion of your case. Finally, a single green star indicates a string citation or other brief mention of your cited case. Judging from the response in the audience, this idea was very popular.

Customizing Your Options

Although there was barely any time left for a review of the customizing options on KeyCite, they should be mentioned since there was commentary on them during the morning. I looked at them in more detail once the session was completed. The initial screen you see on KeyCite provides the direct history and negative indirect history of the case. You then have the option of reviewing negative history only or to omit minor history. Presumably, minor history indicates one star treatment. A third option is to list the citing cases. In Dan’s example, the list was 242 cases. If you click on this third option, further limiting options are listed.

The most obvious limitation offered is to restrict your hit list by West digest topic and key number. Some audience members pointed out that this is much more restrictive than using a boolean query such as that which is allowed in the Shepard's focus feature. Lexis representatives chimed in that this type of limiting feature restricted you to parsing out cases according to issues of law rather than differences of fact. Hopefully, researchers are trying to match up issues of law rather than fact, but it would be wonderful if you could match up cases according to digest topics, key numbers AND fact patterns. If someone is unfamiliar with the digest topic or key numbers for the case, the ability to create their own query could prove very helpful. Other limits available on KeyCite are jurisdiction, type or name of a reporter in which cases are published, date, depth of treatment, and document type (secondary resources, highest court, administrative material). These features were not reviewed or discussed during the session. It would be helpful to know what secondary resources are covered in KeyCite.

KeyCite of the Future

West plans on adding more extensive secondary resources to its citator coverage. It is already quite deep since it includes references to more than 600 law journals compared to about 400 law journals on Shepard's via the web. West also points out that it provides citations in KeyCite to many treatises and legal encyclopedias such as Am Jur, ALR, Wright & Miller, and Practicing Law Institute materials. Inclusion of pre-National Reporter System case law, administrative law decisions, and patent and labor citators are imminent. Of course, Shepard's already provides access to this type of material. Since these are very important tools on Shepard's, West should probably hustle to provide these services if they do not want to risk market-share impact.

West will unveil their statutory and regulatory citator at the American Association of Law Libraries (AALL) Annual Conference in Washington, D.C., in mid July. They promise a citator for the federal government plus the fifty states. The statutory citator product will also include a reference to any pending legislation that could impact on that statute for the federal government. They expect to roll out a similar product for the larger states such as New York and California. Adding this type of information to the citator is an extremely interesting and useful feature, especially since the codification of bills does not take place in Congress prior to its enrollment. Assuming West can handle the labor intensive chore of codification, this could save the legal researcher a few hours of reading time.

Commentary from the Audience

Since Dan preferred questions to be asked while he was speaking, there were very few hands raised for additional comments on KeyCite. Jody Armstrong felt that it was very important to include boolean searching in order to limit results even further than what was already provided for by Westlaw. The remainder of the questions were write-in questions directed at both companies.

Mary Matuszak of the New York County District Attorney’s Office was interested in the editorial policy behind "related cases." Dan responded that West will consider a case to be related if it is the same case, shares the same docket number or involves most of the same parties. The Lexis-Nexis camp did not have their policy about that type of classification with them but promised to get back to Mary and took down her name. Perhaps Mary can divulge this information to the public at large if she has received a response.

Several people wrote in to ask how quickly will we see negative analysis on either system. Dan already pointed out that the West editors aspire to a 24 to 48 hours response time. Lexis said that they will definitely post negative history within that same time slot.

Some Final Comments

Although we were already running 20 minutes late, I asked Sharon to Shepardize the case that Dan used for illustration, and asked Dan to KeyCite the case that Sharon used for illustration. Since we did not have time to discuss the differences during the session, I ran this exercise myself to see if there were any earth shattering differences in my results.

I used Dan’s citation to 530 P2d 589. On KeyCite, Dan found 224 citing references. On Shepard's I found 189 citing references. These include secondary sources and primary sources of law. When using Sharon’s citation of 414 US 573, Shepard's found 635 citing references. When I used KeyCite, I retrieved 658 references. Since both cases had significant negative history it is interesting to note how that information was displayed in each system. On Shepard's the researcher will see the familiar "Superceded by" editorial comment. On KeyCite the display read "opinion vacated by." The numbers between the two are not too far apart. I venture to guess that the greater numbers in KeyCite may best be explained by the use of extensive secondary sources in the KeyCite product.

It was not until I sat down and actually compared my results on these two cases that I could articulate what I preferred about each system. On Shepard's I liked to see my results listed by order of jurisdiction. On KeyCite, the order of positive and negative history is listed by star depth. My personal preference is to see the treatment afforded by case law in my local jurisdiction or from the Supreme Court. I realize I can customize my KeyCite display by jurisdiction but there are two problems with this method. First, it requires an extra step on the part of the researcher. Second, even if you select certain jurisdictions they are still grouped together under the appropriate star level treatment. This forces the researcher to read through the citation to figure out which court is involved. In contrast, the CRF on Shepard's permits greater latitude for the researcher who wishes to customize his or her results. Also, all of the restrictions on Shepard's appear on one web page for easier selection.

The most outstanding feature on KeyCite for me is the list of digest topics and key numbers. The ability to see what the headnotes are in the cited case and limit your results to cases discussing those particular legal issues is very useful. Shepard's does not provide this information. In fact, I find it even more confusing when I have two sets of headnotes to choose from on the CRF. I also like the maneuver away from the traditional editorial codes found in the Shepard's citators. Phrases such as "superceded by statute as stated in," and "opinion vacated by" provide a better idea of the type of impact made by a citing case.

At the end of the day, we have the choice of two good products to update and verify our research. Neither will please everyone all the time. Some of us will continue to use the system with which he or she is most comfortable. One thing is certain though, we will not be updating or verifying in hard copy unless we absolutely must!