Peter A. Hook
Doctoral Student, School of Library and Information Science (SLIS)
Indiana University--Bloomington
1202 North Lincoln St. Apt. #1
Bloomington IN, 47408-1823
pahook@indiana.edu
http://ella.slis.indiana.edu/~pahook/index.html

        I am interested in applying cartographic techniques to rigorously created concept spaces (using information science techniques) to create maps for teaching.  There are a number of ways that domain maps might enhance classroom pedagogy.  At the beginning of the semester students might be given an overview map of their subject.  In different weeks, the instructor might zoom in on different sub-fields in the discipline.  Sometimes the maps might reveal bibliographic elements like the most cited papers in a particular field or the co-authoring relationships of a particular author.  Sometimes the maps might be based on semantic/topic space.  At other times, the maps might be based on literal geographic elements like circuit of origin for United States Supreme Court cases.

        Presently, domain maps are created using a very limited array of databases.  While good article databases exist, there are problems with cross platform searching.  Also, citation data for monographs (books) is hard to obtain.  Hopefully, increased public awareness of the utility of domain mapping might result in changes as to how the data necessary for domain visualizations is produced and stored.  Citation formats and practices might be standardized across many different disciplines and novel approaches for collecting citation data might be tried.
        For instance, with little added burden to book publishers, it might be reasonable to require as a condition of copyright registration, that a publisher submit to a non-profit entity such as the Library of Congress, an electronic list of all citations and references from the book for which the publisher wants to register a copyright.  Additionally, it is possible to envision a Web-based, public collaborative endeavor, in which people submit to a central repository, all of the references and citations from their favorite book (or at least a chapter of it).  This type of data harvesting would facilitate the mapping of non-science domains that rely as much on books as articles for their scholarly distribution (law, history, political science, etc).  Also, the widespread collection of older citation data would facilitate the portrayal of the evolution of particular domains.  More historical data would also allow researchers to pinpoint major shifts in and contributions to the various bodies of knowledge.  This would also provide researchers a better understanding of how knowledge is defused. 

        Presently, many members of the public expect maps to be literal.  It will take some work for the public to accept spatial metaphors to portray inherently non-spatial data.  While the layout left to right and top to bottom of domain maps might be arbitrary, the proximities and adjacencies are created with rigor and reflect the actual structural dynamics for a given knowledge domain.  It will take the public some time to understand this.  Some of the public will also need to understand how the domain maps are produced.

        There are strong pedagogical advantages to creating domain maps.  Domain maps provide a cognitive structure from which students can more effectively problem solve.  Having a good understanding of the “structural organization of a knowledge domain is a better predictor of being able to problem solve in an area than aptitude (as measured by standardized test scores) or past performance on a similar set of problems.”  (Jonassen).   Furthermore, when a domain map becomes part of the navigational environment for accessing information in a particular domain, the use of the domain map to navigate the information reinforces the user’s conceptual schema for that particular domain.  This in turn leads to better mastery of the domain.

        I have domain expertise in the law and have worked recently on projects that attempt to visualize the concept space of various legal topics.  I have used the co-occurrence of topics in Supreme Court cases to produce visualizations of the topic space of the Supreme Court.  Presently, these maps do not utilize cartographic techniques.  That is why I am interested in attending a geographer’s conference to better understand how cartographic techniques may be used in domain maps and to identify potential collaborators.  Recent work on this topic may be found at two places:

Poster for an Information Visualization Open House:
The United States Supreme Court: Visualizations and Metrics (60 years of data), http://ella.slis.indiana.edu/~pahook/product/iv_open_house.ppt

PowerPoint Presentation at a Doctoral Student Research Seminar:
The United States Supreme Court: Visualizations and Metrics (60 years of data), 
http://ella.slis.indiana.edu/~pahook/product/sct_2004-09.ppt

While I must acknowledge the generous assistance of Dr. Katy Börner, numerous graduate students, and my former colleagues at the Indiana University Law Library, the work is my own.

Teachers and students of legal information. 

The formation of conceptual schemas (mental models and concept maps), for use in studying a domain and becoming a better problem solver.     

This dataset consists of all top level West topics that have been assigned to United States Supreme Court Cases since 1944.  The data was downloaded from the Westlaw commercial database licensed to students and faculty of the Indiana University School of Law.  A partition was created based on the combined categories of Crimes and Remedies from the West Outline of the Law.

Network graphing techniques and visualization tools (Pajek).

(West Topics in the Supreme Court) - The graph is too densely pulled to the center to reveal any structural insights.  A fisheye view of the center reveals the clustering of the criminal law node with several remedies.  This is actually an area of the law known as criminal procedure.  By comparing the Pajek rendering of the graph with a criminal procedure textbook, it is possible to make a domain map of criminal procedure.

The pros are suggested by the education literature on cognition which I have alluded to above.   These will have to be born out by rigorous user testing.  Certainly, cons will also be discovered during user testing. 

        I want to use the information science techniques described in the work above, combined with cartographic techniques, to make a domain map for a specific law school course—Contracts.  More specifically, the domain map will portray a specific casebook--Randy Barnett’s, Contracts: Cases and Doctrine, Third Edition.  This map will be given to law students for preparation for their final exam.  Feedback will be collected as to its utility and usability.  Ideally, I would also examine whether test scores go up when students use the domain map.