I, Dr. William Barclay Allen, hereby declare as follows:
1. I am currently Dean and Professor at James Madison College of Michigan State University in East Lansing, Michigan, where I teach classes in political philosophy, American government, and jurisprudence. I taught formerly at Harvey Mudd College in Claremont, California. I was also formerly an active member of the United States Commission on Civil Rights, having been appointed by President Reagan in 1987, and I served as Chairman of the Commission from August 8, 1988 to October 23, 1989. I also serve currently under a grant from the State of Michigan Department of Education to devise a “New Framework for Public Education in Michigan.”
2. I received a Bachelor of Arts degree from Pepperdine College, and I earned both a Masters degree and a Ph.D. in Government from Claremont Graduate School.
3. Attached hereto as Exhibit A is a true and correct copy of my curriculum vita, which accurately sets forth my educational and career background and various of my accomplishments.
4. Throughout my educational and post-graduate experiences, and because of my specific academic concentrations, I have become intimately familiar with the complexities, concerns and intricacies of voting rights in this country, and I have developed an expertise for assessing the effects of redistricting on the voting rights of minorities and the dynamics of electoral competition within defined districts.
5. In this regard I have acted as an expert on voting rights issues in the following civil cases: Carillo, et al. v. Whittier Union High School District; Romero, et al. v. City of Pomona; Badillo, et al. v. City of Stockton; and Arizonans for Fair Representation, Inc., et. al. v. J. Fife Symington, Governor of the State of Arizona, et al, the latter three of which are reported respectively at 665 F. Supp. 853 (C.D. Cal. 1987), 1992 WL 19400 (9th Cir.), and [CIV 92-256-PHX-SMM, U.S. Dist. Court for the District of Arizona]. I have testified on and analyzed the issue of voter polarization among racial and language minority groups, including in the Carillo case, and have also provided testimony and analysis concerning cohesion of voting patterns among different minority groups and electoral dynamics, including in the Romero, Badillo, and Arizonans for Fair Representation cases.
6. I am also responsible for the introduction of a methodology known as “surname analysis” to the field of voting rights litigation. This methodology is a means to determine the approximation of the number of Hispanic registered voters in a district or precinct, based on a review of the surnames of registered voters as compared to an official Hispanic surname list provided by the United States Department of Commerce (Census Bureau). In my experience, I have found that this methodology produces a reasonably accurate measure of Hispanic registered voters in a particular district or precinct, and surname analysis is now relied upon by experts in the field of voting rights analysis. I first employed this methodology in the Carillo case and more fully elaborated upon it in the Romero case.
7. Moreover, I applied techniques from developed methodologies in political science analysis of citizen participation to create a “Profile of Winning Candidates” in order to assess the electoral dynamics within a jurisdiction in relation to assessing the chances of minority candidates prevailing and minority voters succeeding in electing candidates of their choice in at-large elections. In my experience I have found that this methodology produces an accurate assessment of chances for election based on prescribed practices. I first employed this methodology in the Romero case, and relied upon its principles in evaluating the Arizonans for Fair Representation case. Its focus on networking has been reflected in the ruling of that Court, which held that “This principle is important because it recognizes the importance of shared local experience and the ability of groups and candidates to ‘network’ within their communities.” (p. 6, emphasis added.)
8. Finally, I have been a candidate in California for the Republican nomination to the U. S. Senate.
9. In conjunction with my knowledge and expertise in the area of voting rights, I was asked to review, and have in fact reviewed, numerous election and demographic data, charts, and maps relating to voting patterns, demographics, and community characteristics in the State of Louisiana. Specifically, I have reviewed the following: the geographic pattern of Congressional districts as provided by Louisiana’s Act 1 following the 1990 census; a breakdown of voting by precincts following the Act 1 district lines for several elections prior to and after the reapportionment was implemented; media directories and maps for the State of Louisiana; voter registration, voter turnout figures, and election results for 1986, 1987, 1990, 1991, 1992, and 1994 elections for several district and statewide offices; reports by Dr. David Bositis, Drs. Charles Tolbert and Forrest Deseran, Dr. George Castille, III, Dr. Ronald Weber, Dr. Engstrom, Dr. Reaves, and Dr. Susan Howell.
10. Based on my review of the information outlined in paragraph 9, and consistent with my knowledge and expertise in this area, I have formed the following conclusions about the Act 1 redistricting.
11. I conclude that the congressional redistricting plan enacted as Act 1 operates to diminish the role of black influence in Louisiana congressional balloting, in terms of concentrating a substantial minority vote (30% statewide) in only two of seven districts with the result of marginalizing the incremental black vote that remains outside of those two majority-minority districts.
12. As set forth in more detail hereinafter, the effect of the Act 1 redistricting on localized communities of interest among black voters in Louisiana is to deprive them of the opportunity to determine the outcome of balloting for representatives in Congress in as many as five out of seven districts, which when weighed against the two majority-minority districts, produces a net reduction of influence of three districts or 60%. A singular fact in this regard is the effect of isolating black communities in Shreveport and Alexandria from the opportunity to determine the general outcome of congressional influence within their own regions, by means of sweeping down from the Northwest of Louisiana all the way through central Louisiana in order to incorporate substantial black voting populations in District 4. This dynamic is illustrated by the fact that the Gale Directory of media outlets in Louisiana shows only two black community newspapers outside of New Orleans, one in Shreveport and the other in Alexandria. As a result of this districting plan, all the black community newspapers in the state are included in only two-sevenths of the state’s congressional representation, although black voters make up 30+% of the state’s registered voters. A related observation emerges from reviewing Tolbert and Deseran, who report in a discussion of “spurious correlations” that the socioeconomic standing of whites and blacks in District 4 is far more nearly similar than in the remaining districts. Though this similarity can be supported in this manner, it obviously also bears strong relationship both to the dissimilarites between blacks and whites in the remaining districts and also to the phenomenon of the isolation of blacks with interests similar to those of blacks in District 4 from the potential of competitive influence within the districts where they are left.
13. A useful image of the notion of local community suggested here and supported by research is conveyed in a recent description by Hugh B. Price, President of the National Urban League, who described the dynamic in an essay published in a black community newspaper. “Our society also consists of the communities that comprise it and of the people who live in those communities. I really worry that we’re losing sight of what community means, much less what it means to a civil society. Communities aren’t abstractions or statistics. They are neighborhoods in towns and cities where real people live and work. They are the building blocks of a civil society... As the social critic Christopher Lasch observed, informal community hang-outs like coffee houses, beauty shops and bars are crucial to democracy, because they’re the venues for conversation, the reinforcement of values, arguments about local issues... I am hard pressed for time these days. I decided a couple of months ago to try one of those ‘barbers by appointment.’ After only one sitting, I gravitated back to the neighborhood barbershop...that I’d patronized for 17 years. Why? Because I sorely missed the conversation among plumbers, mechanics and ministers that keeps me in close touch with my reality as an African American... The black barbershop is the ultimate democratic institution... A community is much more than an aggregation of individuals and households whose only contact with one another is via Internet... When you bundle communities together and surround them with a border, they become what we call cities. Cities matter to our society, therefore, because communities matter.” Black Voice News, September 21, 1995, p. A-2.
14. Two questions have held my attention. First, do the consequences of the districting plan under review impair the capacity of a community or communities to intervene in the electoral process, for agenda setting and candidate selection right through eventual election, with reasonable opportunity to influence the outcome along the lines of identifiable communities of interest? Second, whether the Act 1 or the Court-ordered districting plans respond to peculiarities of racial polarization and community isolation in Louisiana in such a way as to foster participation and effective choice that would not otherwise occur.
a. The question of whether one or more communities enjoy full and fair opportunity to intervene in the electoral process, from agenda setting and candidate selection right through eventual election, cannot be determined on the basis of the record I have seen. One might make an informed speculation that the most usual and known factors conditioning electoral participation will operate less efficaciously in any electoral system in which the participants, however united in general views, are nevertheless divided by dramatic geographic and social interaction frameworks. To provide a thorough analysis of this question one would require to analyze the network characteristics of candidate selection in Louisiana politics as a whole and by regions, as well as the actual social patterning of political and governmental decisions.
1) This carries especially heavy weight in a state which permits the election to be carried at the primary stage and which, therefore, does not limit primary participation to party adherents. Since the data reveal substantial party and race cross-over voting in Louisiana in particular circumstances, then it must follow that the particular electoral framework enforced will affect coalitional practices and opportunities. For example, it is clear that the black voters concentrated in District 4 enjoy appreciably reduced coalition opportunities in relation to their hypothetical opportunities under a congressional districting plan which permitted as full weight as their natural geographic dispersion would allow.
2) In fact, a fair reading of the data suggests, and I’m certain an actual analysis would confirm, that the voting results in at least three of the remaining six districts would be very different from their present configuration, and the difference would reflect the impact of the choices and influence of black voters. To the extent that this observation is correct, it means that the residual black voters in those districts have suffered the effect in the current districting plan of being unable to select representatives of their choice. The significance of this last observation runs not to the theory that democracy owes anyone in particular a right to elect in addition to a right to vote. It runs rather to the conclusion that folk whose circumstances geographically and socially would almost surely confer on them an opportunity to see their choices elected have lost that opportunity in a districting plan that abstracted from those geographic and social realities. This criticism does not apply to the court-ordered plan, which minimizes such effects.
b. The inquiry whether the existing districting plan responds to peculiarities of polarization and community isolation in such a way as to foster participation and effective choice that would not otherwise occur cannot be answered on the basis of the data thus far submitted. That data, to the extent I have seen it, relies almost exclusively on ecological analysis and that even includes a survey (Howell) based on phone listings rather than voter rolls and exit poll tabulations borrowed from national news network analysts that are not susceptible to cross-tabulation with the polling results, the census data, or any particular voters. It is too well known to dwell upon, the problem of the “ecological fallacy.” In this case it means assuming the distribution of votes based on general population distributions (which renders accurate measures of cross-over voting far less likely). That kind of analysis fails to identify the choices that voters actually make. In sum, it is impossible to discriminate constituent elements as independent variables in an ecological analysis, while, at the same time, the analysis of electoral dynamics for the purposes in view here absolutely requires discrimination among constituent elements in order to generate a reliable algorithm. For example, the question, “What do voters want?” requires that there be voters and not merely census demographics. Otherwise one can draw no nearer than an approximation. The Bositis study, moreover, focuses on outcomes rather than inputs, which is at best an indirect measure of voter polarization, inasmuch as it forces one to infer the intentions and wishes of voters rather than to weigh them directly. The established methodologies for studying polarization focus on voter choices through such devices as Homogeneous Precinct Analysis, Direct Polling, Deviation Analysis, etc. The “Responsiveness” analysis (the legislative performance of Representative Cleo fields) does not directly address the question of polarization and is not included in this observation.
1) The weight of this inquiry hinges almost entirely on the answer to the trailing inquiry, namely, whether there results participation and effective choice that would not otherwise occur. The reason for this is twofold. First, polarization between blacks and whites in voting choices is clearly discernible in some elections. It would be more helpful, of course, if analysis were done in cases that were not defined ex hypothesi as racial cases (which is the chief liability of the Engstrom study, in addition to its apparently exclusive use of ecological regression). In that manner one could judge how far the results and not the particular subject identified the polarization. The conclusion in Howell, at p. 10, for example (“These result [sic] suggest that the creation of District 4 has increased the political awareness of its black residents.”) is clearly a case of post hoc ergo propter hoc. What really happened is that the creation of District 4 left black voters no option. Thus, the conclusion regarding “black/white differences in opinion” cannot address directly the opinions of persons who actually participated in the election (p. 7).
2) Secondly, and more importantly, the question is less whether polarization exists but what is its particular configuration. It is possible, for example, that enhancing coalitional opportunities would be the indicated strategy under certain conditions of polarization. Since the election data show a substantial potential for white cross-over voting, configurations that have the effect of constraining their options for cross-over voting may not only reduce such cross-overs but also advance other candidacies than those now encountered (Sometimes it matters for black voters which white representative is elected as much as whether a representative is white or black). Therefore, it means everything to say whether increased black participation (if it has increased, which the data do not in fact show, since no attempt has been made to assess the fall-off impact of isolation of small and less influential black votes in Districts 1, 3, and 5-7; this observation also applies to the Howell conclusion at p. 10 with respect to the political awareness of black and white residents in the putative District 4) and effective choice represent a net gain of influence, a loss of influence, or no change. In the absence of an answer to that question no reasonable inference may be drawn from the fact of polarization in the State of Louisiana in respect to the issue under review.
15. The inquiry whether the Act 1 or the court-ordered districting plans will foster civic community between and among blacks and whites in the State of Louisiana draws a complete if not a comprehensive response in the testimony and data submitted. Surprisingly, there seems to be a rough or working consensus between the plaintiffs and defendants (or, at least, the U. S. Department of Justice). The leading premise of the contributions from the latter seems to entail the demonstration that no plan with such a goal would be acceptable because such a goal is unrealistic. On the other hand, the plaintiffs’ view (or, at least, the testimony of Dr. Weber) runs from the premise that the Act 1 plan precisely thwarts prospects of civic community between and among blacks and whites (at the same time as reducing the influence of black voters through the device of electoral ghettoes). Since they largely agree that the Act 1 plan is inconsistent with community between black and white voters, the relevant inquiry would seem to be whether the research offered for review shows that to be a necessary result or an unnecessary result. The Howell survey, for example, at p. 14, suggests that the effect of District 4 is to dilute the relative strength of black conservative voters in Districts 5 and 6. This is amplified by the testimony of Weber at p. 5, paragraphs 12 and 13.
16. One might say more about this than the research proffered allows, if one were in a position to examine the actual practices and dispositions of Louisiana voters (beyond what Weber has provided, which is helpful but not comprehensive). This further inquiry cannot and ought not be carried out in the model of ideological litmus surveys. It requires rather to surface community concerns and community resources and networks characteristically relied upon for problem-solving. For example, the manner in which Louisianans deal with public education, in the parishes and in the state, should reveal much about the potential of Louisiana voters to cooperate and act in coalitions in a well designed system. The fact that the state holds both state-wide, district school elections (analogous to congressional districts) and also parish elections provides a rich resource, without employing which I do not see how one otherwise could conclude definitively regarding the prospects for fostering civic community. It takes neither study nor genius, however, to observe that what requires cooperation will invariably serve better than what requires separation to build the possibility of civic community.
Pursuant to 28 USC Sec. 1746, I declare under penalty of perjury that the foregoing is true and correct and that this declaration is executed on this 27th day of October, 1995 at East Lansing, Michigan.
William B. Allen
 The primary nomination of Representative Cleo Fields on a state-wide basis in the current election, with a reported 40% of the black voters selecting a candidate other than Fields, would constitute an excellent basis for determining the extent of polarization and the potential for influence voting. Because this election takes place on a statewide basis, it illustrates well the kind of dynamic which should take place on any district level where the voter demographics would mirror those of the state at large.
 Since writing the text at this point, I have learned that the original Howell study in fact misidentified even the residents of CD 4, rendering the results still less applicable to the conclusion drawn. The correction applied at best restores that analysis to the level criticized here, to repeat which, it must be observed that a reading of the opinions of residents in a specified jurisdiction does not constitute a sufficient test of the choices of voters. The set of residents includes the set of all voters. However, one requires to implement a step-down transform in order to make the first order, general opinions of residents in any way reflective of the opinions and/or choices of voters. Such a step-down transform may be executed in several ways, among which, tests contained in the survey instrument of "propensity to vote," participation in voting, and reliability of responses regarding voting behavior. None of these measures were included in the survey data on which Howell relied.