The book, "Statistical Methods for Human Rights" [1], is modestly presented, fitting the style of its contributors. People who work to promote human rights tend not to boast about what they have done. In advocacy, quiet diplomacy often works best, as oppressive governments are less likely to grant concessions if they think bad publicity will result anyway. Also, publicizing analytical methods can help violators learn how to escape detection.

Nevertheless, the stories are spectacular. In Chapter 9 is the story of how one letter, drafted by a handful of people in less than a day, apparently persuaded the Soviet Union, at the height of the Cold War, to end a serious, widespread abusive practice. Another letter is credited with saving a former government minister from execution in Liberia following the 1979 coup there. A small group of analysts used text retrieval and search capabilities, primitive by today's standards, to bring many perpetrators of Argentina's "dirty war" to justice. In Chapter 2, Mary Gray and Sharon Marek mention how another small group of analysts was instrumental in the prosecution of Slobodan Milosevic for war crimes in Bosnia-Herzegovina (Chapter 10, a detailed review of preparing evidence for the International Criminal Court, describes this effort in more detail; also see [2]), and describe how yet another group was less successful in an attempt to bring the perpetrators of the Rwandan genocide to justice. Still another group, not cited in the book, did economic impact analysis later that proved to be essential to the International Campaign to Ban Land Mines, which won the 1997 Nobel Peace Prize [3].

The subject has come a long way since the pioneering work in the field 20 years ago that showed what was possible, including the observation by an O.R. analyst familiar with consumer credit scoring that even missing or distorted data could provide information about violators: the applicant who leaves "employer" blank is providing important information, and so is the government that can no longer tell you where the members of a threatened ethnic group live [4].

Analytical Contributions


While the stories are fascinating, the book also contains a remarkable collection of innovative methods, most likely applicable to many other topics. Chapter 3 is a careful review of how to estimate collateral and indirect casualties from war and why it is important, even from the strictly military perspective, to do so, using the first Gulf War ("Operation Desert Storm") as the primary example. The author, demographer Beth Osborne Daponte, who worked for the

U. S. government at the time of the conflict, focused especially on what can be considered excess mortalities relative to the military effect. She estimated a disturbingly high number — 111,000 civilian deaths due to indirect effects of military action — for this war. (As a result of her early work, Dr. Daponte became the only scientist the American Statistical Association's Committee on Scientific Freedom and Human Rights ever recognized and adopted as a case of government suppression of scientific freedom in the United States.) Despite this and other studies, she concluded, the issue of proportionality of casualties to military results still has received little scientific attention.

Chapter 4 is a careful, closely reasoned study, using a comparison of Cuba and other Latin American countries, of how to distinguish between government actions that were necessary because of economic hardship and actions that can properly be viewed as human rights violations. The method is a combination of comparative analyses, longitudinal studies and modeling the differentials in certain indicators over the same time period in different countries. This approach seems promising as a way to assess governments' claims that human rights violations were justified sacrifices for the pursuit of a "greater good." In this case, too, there is a story behind the statistical analysis: Jorge Luis Romeu, the author of this chapter, is well acquainted with the data and the details of this subject, as he came to the United States in 1980 as part of the Mariel boatlift.

In Chapter 5, forensic pathologist Clyde Collins Snow and several co-authors reexamine what data can be gathered, and what can be inferred, regarding the disposition of "disappeared" persons and unidentified decedents, using the events in Guatemala in the 1970s and 1980s as the primary source. In Chapter 6, Romesh Silva and Patrick Ball expand on previous approaches to estimating indirect casualties in war, using East Timor as the example. Their comparisons and contrasts of different data sources and methods yield more credible results than could be obtained from any one source. Multiple systems estimation (MSE) was first applied to a demographic problem nearly 60 years ago, by no less renowned an analyst than W. Edwards Deming [5], but it is not widely known and has been sparsely used in more recent human rights work. The technique is clearly applicable to other types of problems, as well.

Chapter 7 reports the results of a survey of people in Afghan refugee camps in Pakistan. The difficulties this team had to overcome to obtain solid results, and the creativity and persistence with which they addressed these difficulties, are informative and exemplary. Chapter 8 is a description of Metagora, a new, comprehensive set of reporting standards and metrics, to address the long-standing challenges of making reports and indicators consistent enough to be used in the kinds of comparisons that form the basis for assessments of human rights practices. A noteworthy point is that, although these problems had been widely known and discussed for some time, the first comprehensive professional conference on the subject took place only recently, in 2000. The field of human rights monitoring and assessment is still relatively new, with many worthwhile tasks still waiting to be done.

Broader Methodological Contributions


O.R. analysts who prefer conceptualization to exhaustive (and often exhausting) data analysis should especially appreciate the last three chapters of the book. In Chapter 11, David Banks and Yasmin Said reexamine the definitions of human rights and the relative emphasis placed on different aspects of human rights, with the objective of highlighting where new approaches seem most needed. As the emphasis shifts from civil and political rights to social and economic rights, analysts must reopen questions about defining what governments are required to do, determining what constitutes a violation of rights, and deciding what benefits are worth what costs. One culture's idea of basic human rights sometimes constitutes another's idea of an unworkable or undesirable ideology. Even with a consensus about objectives, many proposals to address these objectives have great potential for unintended and unpredictable consequences.

To help clarify definitions ad objectives and assess the implications of different ideas of social organization, Banks and Said propose agent-based simulation of alternative theories of justice. These models would be built around hierarchical probability models, to keep their complexity manageable. The authors' proposed approach includes provisions for taking grudges into account, along with different schemes for persuading people to discount past offenses. Readers familiar with recent work in agent-based simulation of societies and in models of economic cooperation will not be surprised to learn that the discussion next delves into game theory, in particular Axelrod's famous experiments about the evolution of cooperation in repeated plays of "Prisoners' Dilemma" [6].

They conclude, "Statistical thinking can address new problems in the evolving theory of social and economic rights. This chapter has explored some of the nonstandard things we can contribute, suggesting:

• ways in which human rights behavior can be simulated, in the context of a theory of justice based upon exchangeable agents;
  
  
• a procedure for determining how to fairly distribute the burdens of charity among the well-to-do and wealthy; and
  
  
• a methodology for studying the costs of compensation as a function of time, and this appropriately discounting past injuries."

Banks and Said go on to admit that they are not experts in the techniques they advocate, and readers of OR/MS Today may draw their own conclusions about whether O.R. analysts, or computational social and organizational scientists, might be better prepared than statisticians to tackle problems of this nature. The authors' point clearly is not to stir up squabbles, but to call attention to an exciting analytical opportunity and recommend that someone ought to be doing it. Indeed, given their comments about the desirability of cooperation in society in general, it seems highly likely that they would assume cooperation among disciplines would be beneficial here.

Chapter 12 goes further into the area of defining and assessing the more elusive human rights objectives in the social and economic area, focusing on the Millennium Development Goals adopted by the United Nations in 2000. They add that improved evaluation will be essential if the goals are to be met, and they note the major obstacles:

• lack of sound data collection;
  
  
• lack of a baseline;
  
  
• lack of good statistical analysts; and
  
  
• opposition to experimental evaluation.

This discussion proceeds, in turn, into a review of the most accepted quantitative methods for evaluation, such as clinical trials in medicine; suggestions for integrating quantitative and qualitative approaches; recommended ways to achieve greater "data transparency," enabling critics, people likely to be affected, and the public to understand where data come from and what they imply; and an explanation of why impact evaluations are relatively rare, and what can be done to rectify this. They cite the varying degrees to which organizations such as the World Bank, the U. S. Agency for International Development, the Inter-American Development Bank and the Millennium Challenge Corporation have provided incentives and requirements for evaluation as part of their programs.

They advocate the development of new statistical techniques to estimate program impacts; again, as organizational design issues are involved, readers of OR/MS Today may have ideas that go beyond traditional or even innovative statistical methods, encompassing the larger topics of evidence-based decision-making. Their bibliography and listing of Web resources is extensive and should prove extremely useful to anyone interested in learning more in this topic area.

The final chapter is deeply troubling, as William Seltzer and Margo Anderson present a number of examples of how well-intended, seemingly harmless population data systems can be and have been used to target vulnerable individuals and subgroups. In the United States, they identified five occasions of interest: the use of census data to facilitate forced relocations of Native Americans in the 19th century; the use of 1910 census data to support draft registration prosecutions; the use of federal statistical data to direct forced internment of Japanese Americans during World War II; possible violations of confidentiality of education statistics data in investigations of suspected terrorism since 2001; and possible use of 2000 population census data to track Arab Americans. They list many other instances from other countries, some — such as the Nazis' use of occupied countries' population registries to help round up Jews for deportation and extermination — resulting in much more serious harm than any of the U.S. cases. They recommend five improvements:

• improving safeguards and prevention strategies;
  
  
• more attention to risks associated with mid-level data;
  
  
• training;
  
  
• further research; and
  
  
• the establishment of an incident register, "both as a deterrent against efforts aimed at undermining statistical confidentiality and as a record of successes of federal statistical agencies in resisting efforts to compromise their responsibilities to data providers."

This chapter, too, has an extensive bibliography that should prove very useful to anyone who wants to learn more about this vital subject.

Concluding Note


At the end of the movie "Schindler's List," as Oskar Schindler is about to be presented a major award for what he did, he protests to one of the people who nominated him, "But I did so little." The nominator, who is one of the people Schindler saved from extermination, replies, "Oh, but you did so much." Human rights work continually involves this dual, seemingly contradictory truth, and the stories of the people who contributed to this book reflect their sense of accomplishment and their frustration at not having done much more.

The book is dedicated to the memory of Graciela Mellibovsky Saidler, an Argentine government economist who "disappeared" in 1976, shortly after she did a statistical study on conditions in the slums of Buenos Aires. The junta reportedly found this study embarrassing. Despite intensive efforts by the American Statistical Association and others, her fate and the disposition of her remains are still uncertain. She exemplifies, and the junta evidently knew, what many of her colleagues have yet to learn: a few analysts can make a big difference.



Douglas A. Samuelson is a principal decision scientist at Serco-North America, a general professional services company in Vienna, Va. He is also president of InfoLogix, Inc., a research and consulting company in Annandale, Va. He was one of the contributors to the book and a member from 1979 through 1996, and chair from 1985 through 1988, of the American Statistical Association's Committee on Scientific Freedom and Human Rights. He is a frequent contributor to OR/MS Today.

References

1. Jana Asher, David Banks, and Fritz Scheuren, eds., "Statistical Methods for Human Rights," Springer, 2008.
2. Patrick Ball and Jana Asher, 2002, "Statistics and Slobodan: Using Data Analysis and Statistics in the War Crimes Trial of Former President Milosevic," Chance, 2002.
3. "Statisticians Contribute to 1997 Nobel Peace Prize Work," AmStat News, April 1998.
4. Richard Claude and Thomas B. Jabine, eds., "Human Rights and Statistics: Getting the Record Straight," U. of Pennsylvania Press, 1992.
5. C. C. Sekar and W. E. Deming, "On a Method of Estimating Birth and Death Rates and the Extent of Registration," Journal of the American Statistical Association, 1949.
6. Robert Axelrod, "The Evolution of Cooperation," Basic Books, 1984.

Table of Contents

OR/MS Today Home Page


OR/MS Today copyright © 2008 by the Institute for Operations Research and the Management Sciences. All rights reserved.


Lionheart Publishing, Inc.
506 Roswell Rd., Suite 220, Marietta, GA 30060 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com


Web Site © Copyright 2008 by Lionheart Publishing, Inc. All rights reserved.