RESOLUTION OF BIAS: TORT DIVERSITY CASES IN THE UNITED STATES COURTS OF APPEALS*

When the federal courts were created by the Federal Judiciary Act of 1789, many of the suits entertained by these forums involved disputes between aggrieved citizens of different states. section 34 of the act expressly provided for this grant of diversity jurisdiction: "And be it further enacted, That the laws of the several states except where the constitution, treaties or statutes of the United States shall otherwise require or provide, shall be regarded as rules of decision in trials at common law in the Courts of the United States in cases where they apply." Through this grant of diversity jurisdiction, the lower federal courts were given the power to resolve disputes when it could be proved that each party to a case legally resided in different (or "diverse") states. Mostly commercial in nature, these suits early on represented attempts by litigants to receive the most equitable treatment under law. By applying the "laws of the several states," federal judges, acting as unbiased adjudicators, would ensure that in-state and out-of-state litigants alike achieved equal justice in a neutral federal forum. The belief was that the parochial biases of state courts could, in effect, be sidestepped, while at the same time principles of federalism remained safeguarded through proper application of the relevant state law.

The grant of federal jurisdiction in diversity-of-citizenship cases remains today, as it did at the inception of the federal court system, an important part of the federal court docket. However, its history and its present role in American law have been riddled with controversy. There has been a continuing debate among legal scholars and practitioners over the propriety of retaining federal diversity jurisdiction. Members of the bench, numerous public-interest and legal-aid organizations, and academic scholars have typically become allies in urging the abolition of diversity jurisdiction. In addition to the claim that the grant of such power affects a drain on federal resources, these opponents of diversity jurisdiction argue that the administration of diversity jurisdiction is "inherently troublesome as it encroaches on state sovereignty" (Baker, 1995:766). Likewise, to these critics, the isolation of the federal courts renders them unsuitable for developing state law in a dynamic fashion. On the other side of the debate stands the private bar, which traditionally supported the status quo. To this group, diversity jurisdiction is a well-established practice in no need of alteration (Stone, 2001). The substantial arguments for retaining diversity jurisdiction are as follows: 1) Abolition of diversity jurisdiction would only shift a large caseload burden from the federal to the state courts. 2) Because the quality of justice is higher on the federal bench, diversity jurisdiction is necessary. 3) Diversity jurisdiction exists to serve its original functionto protect out-of-state litigants from state court bias (Kramer, 1990).

While these and similar arguments have been made for and against the abolition of federal diversity jurisdiction, little systematic empirical analysis has been undertaken to assess the validity of the claims made by both sides of this ongoing debate. This article seeks to address two major questions regarding the propriety of retaining diversity jurisdiction. The first is whether the invocation of federal diversity jurisdiction has the potential for eliminating bias toward in-state parties that would have arguably inhered in a decision rendered by a state court. The second question is whether the federal judges hearing diversity cases are applying the relevant and legally required state law necessary for resolution of the case at hand. Here it is argued that these federal courts are in fact doing their jobs and are not only rendering unbiased decisions, but also settling these disputes based on applicable state law and are thereby holding true to principles of federalism.

DIVERSITY JURISDICTION IN THE UNITED STATES COURTS OF APPEALS

Federal-court diversity jurisdiction is based on the perception that federal courts provide litigants with an arena free from state and local biases. When there is no question of federal law involved in a case, a litigant may achieve access to the federal courts only by means of the diversity statute, which provides that federal courts will have jurisdiction over any civil suit involving citizens of different states. For example, if a citizen of Pennsylvania is injured as a result of using a defective snow blower and the manufacturer of that defective product is a citizen of Delaware, the plaintiff can sue in either a Pennsylvania state court or the appropriate U.S. District Court asserting a manufacturing-defect theory under state negligence law. Even though tort law is typically based on state case law, when both parties are diverse, the suit can be entertained in the appropriate federal court based on this statutory grant of jurisdiction. A further requirement for the invocation of diversity jurisdiction is the meeting of an "amount-in-controversy" requirement that a certain damages figure be claimed. Congress first established this monetary requirement in the Judiciary Act of 1789, pursuant to its powers under Article III of the U.S. Constitution and set it at $500. This figure was increased over time-in 1887 to $2,000, in 1911 to $3,000, in 1958 to $10,000, in 1988 to $50,000, and, since 1996, to $75,000. If the statutory amount cannot be satisfied, the case can only be heard in the appropriate state court.

An understanding of the history of the creation of diversity jurisdiction in the federal courts is crucial. According to conventional wisdom, section 34 of the Federal Judiciary Act of 1789 was an all-encompassing mandate for federal courts to apply state law in diversity cases (Warren, 1923). This traditional view held that diversity jurisdiction was created to protect out-of-state citizens from the biases inherent in the numerous, already existing state courts (Chemerinsky, 1989). These early state courts, said to be rife with biased, nonprofessionally trained judges, often rendered decisions that overtly benefited the in-state party to the suit at bar. Federalists, concerned that creditors were not receiving a fair trial in these state tribunals, advocated a strong role for the national courts to protect the commercial interests of a budding young nation. The Anti-Federalists, on the other hand, desired a greater role for the state governments and the state judiciaries in these matters of local economic concern. Out of these debates, a compromise, albeit in the Federalists' favor, resulted. Federal judges sitting in federal courts could decide important cases involving issues of state law if the parties to the dispute were diverse and if the judges abided by state law and not federal law in rendering their decisions. Thus, Federalists' attempts at national harmony, at least in the commercial realm, were tentatively achieved in the early years of the republic as commerce among the states flourished (Kramer, 1990). Suits meeting the requirements of diversity jurisdiction could be brought in, or removed to federal court, where an impartial judge would apply the applicable state law in all fairness to both sides of the legal controversy.

However, this desired result often was not attained. In Swift v. Tyson (1842), the Supreme Court gave the lower federal courts the power to fashion law in those areas into which Congress could not reach. In effect, a body of "federal common law" emerged from those instances where no state statutes or constitutions were present for legal guidance. In the Swift opinion, Justice Story reasoned that a uniform body of law would result from such a practice. However, instead of promoting such a body of "general law" under which all litigants could obtain impartial renderings of a decision, a practice of "forum shopping" by plaintiffs' attorneys developed. In 1938, in the landmark case of Erie Railroad v. Tompkins, the Supreme Court overruled the holding in Swift. Justice Brandeis opined, "Experience in applying the doctrine of Swift v. Tyson, had revealed its defects, political and social; and the benefits expected to flow from the rule did not accrue" (Erie, at 8). Thus, since the Erie decision, no longer does there exist a federal common law, so federal courts are obligated to apply the laws of the states, including state common law such as tort law and property law, to avoid the inequities that occur when the outcome of a suit depends on the citizenship of the parties. Furthermore, if state law is clear and unambiguous, federal judges are charged with the duty of applying the applicable legal rule without regard for their own attitudes or intuitions.

Conversely, when state law is ambiguous or opaque, the members of the federal bench must "exhaustively dissect each piece of evidence thought to cast light on what the highest state court would ultimately decide" (Friendly, 1973). Finally, where there is no state law that coincides with the elements of the litigation in dispute, a federal judge presiding in a diversity case must predict how the state court of last resort would be most inclined to behave if such a law was present in the state's body of laws (King v. Order of United Commercial Travelers, 1948). In fact, the Supreme Court has even spoken on questions of conflicts of laws in diversity cases, holding that such predictions should also be made when conflict principles are at issue in suits based on diversity jurisdiction (Klaxon Company v. Stentor Electric Mfg. Co., 1941). That is, when a federal court is faced with the question of what law should be applied to the case at hand, that court in a diversity case must decide the conflicts-of-laws issue as if it were the highest court in the state in which it is sitting. In sum, as a result of the Supreme Court's holdings in Erie Railroad v. Tompkins and its progeny, the lower federal courts have been guided in upholding the principles of federalism through their obligation to apply the applicable state statutory and state common law, such as tort law, in all diversity cases.

However, just as controversy and debate plagued the Framers in the establishment of such a powerful grant of federal jurisdiction, disagreement continues to characterize the discussions regarding the modern propriety of retaining diversity jurisdiction. While the various arguments have been raised and counterattacks launched, little empirical research has been undertaken by either side to the debate. Most of these studies have specifically focused on attorneys' choice of forum in diversity cases and other federal questions, and the empirical evidence on the influence of the fear of prejudices has been mixed. An early study on a litigant's choice of forum found that when parties were of different citizenship, over 60 percent of the attorneys surveyed preferred to bring suit in federal court, out of a fear of bias that was perceived to exist in the state courts (Note, 1965). A similar study conducted several years later found that in Cook County, Illinois, 40 percent of lawyers filing diversity suits reported that local bias was a factor in their decision to file in federal, rather than state court (Goldman and Marks, 1980).

A series of more recent studies, involving extensive surveys of attorney attitudes, have shown that in addition to the existence of a continuing perceived fear of bias, typically by attorneys who represented out-of-state defendants, other factors that tend to guide lawyers' choice of federal or state courts are the quality of the judges and familiarity of court procedures (Flango, 1991, 1995; Miller 1992). Another study took a different approach to addressing the propriety of retaining diversity jurisdiction by using a matched sample of cases from five U.S. District Courts and five state trial courts to assess how state courts would fare if any of three popularly proposed changes in diversity jurisdiction would be established. This study found that if diversity jurisdiction was abolished, or even if a bar was instituted against in-state plaintiffs from filing diversity suits, state courts would be disproportionably burdened; whereas, if the amount-in-controversy requirement was raised, state courts would feel little to no impact (Flango and Boersema, 1990).

Finally, it has been stated, but not tested, that "since the decision in Erie, the federal courts have applied state laws consistently in diversity cases" (Grissom, 2001:385), while another study argues, without empirically testing, "Increasingly, the judicial power of the federal courts is being deployed to limit state tort law in violation of the spirit, if not the letter, of Erie Railroad Co. v. Tompldns" (Lind 2004:717). However, these studies are merely suggestive and provide no direct evidence about whether in-state bias occurs in state courts and, likewise, if such bias is absent in the federal courts, especially the courts of appeals, hearing diversity cases. Additionally, less is known as to whether federal judges are applying the applicable state statutory, constitutional, or common law to these suits as commanded by the Supreme Court in Erie and subsequent decisions.

The questions that remain unanswered in this long-standing jurisdictional battle are fodder for empirical research. An appropriate place to begin this inquiry is in the United States Courts of Appeals. While it is true that the U.S. District Courts are the primary actors in diversity cases, both because they entertain these cases in the first instance under their original jurisdiction and because the judges of these courts are undoubtedly more familiar with the relevant state law, study of judicial decision making in the courts of appeals in tort diversity cases remains important.

The mandatory jurisdiction of the courts of appeals requires the circuit courts to hear all appeals in tort diversity cases, but diversity cases are rarely subject to en bane review within those courts, for "an en banc session is considered an extraordinary procedure, generally reserved for important or novel legal questions" (Barrow and Walker, 1988:11-12). Moreover, the Supreme Court hears very few of these cases (Perry, 1991). Its Rule 10, on guidelines governing the grant of certiorari, does not mention diversity cases as among those to which the Court will give greater consideration in granting review; indeed, the Court did not grant review of any diversity cases included in this study. The result is that the courts-of-appeals panels are essentially rendering final decisions in this area of the law.

An appropriate category of civil cases through which to study the question posed is tort disputes. While images of the shady plaintiff's attorney come to mind when we first hear the term "personal injury," it is important to point out that this body of law encompasses much more than the million-dollar jury verdicts for injury associated with excessively hot coffee. This multifaceted body of law, comprising approximately 20 percent of the courts of appeals' civil docket, contains a variety of issues, including personal injury, medical malpractice, products liability, and defamation claims, which appear to be very similar to those financially charged types of cases which have typified our economic system since the time that diversity jurisdiction was created (Priest, 1990). These tort disputes, which obviously encompass many more case types than the all-too-common "slip-and-fall" injury and which involve the transfer of vast sums of money each year, represent the genre of commercial cases and controversies that leave a daily and lasting mark on our economy-past and present-and thereby provide an important basis upon which to assess whether the federal courts are both acting as unbiased forums and are upholding the principles of federalism.

A systematic analysis of these tort diversity cases in the United States Courts of Appeals may provide important information from which to better inform the debate over the legal suitability of retaining diversity jurisdiction in our federal court system and its implications for judicial federalism. In addition to being of interest to practicing attorneys, court administrators, state judges, and federal judges who daily meet head-on the reality of ever-increasing federal and state court dockets, this study should also be of import to those political scholars concerned about the role of federalism not only in American democracy generally, but also in the courts particularly. If federal courts operate in state court territory by entertaining and resolving suits in diversity-of-citizenship cases, then it should be appreciated that research on the role of federal courts in this area can have important implications for the understanding of federalism as a working concept.

RESEARCH DESIGN

The data for this project come from the Court of Appeals Data Base.1 The votes of the judges on the United States Courts of Appeals in tort diversity cases, for the period 1960-88, were examined in this study. A random sample of thirty published decisions per circuit per year for each circuit from 1961 to 1988 of this analysis is contained in this database; for 1960, only fifteen published cases were sampled. To explain the voting behavior of each individual courts-of-appeals judge in every tort diversity case in the 1960-88 time period, the votes cast by all appeals court judges are examined. Because only one diversity case in the time period was overturned by the full circuit, en bane rulings do not play a part in this study.

Time Period. The year 1960 was chosen as the starting date for analysis, because it was around this time that changes in modern civil-liability law began. Earlier, an appropriate body of tort law had not yet been established across the numerous state court systems. Instead, contract law determined recovery in most personal-injury cases, because many jurisdictions did not recognize negligence as a legitimate cause of action for personal injury. Predictability and uniformity also were not available to the litigants on either side of a civil dispute. However, beginning in 1960, important changes began to occur in American civil litigation. Older causes of action came to be replaced with more straightforward and concise doctrines of liability, thereby forming what has been typically regarded as the law of torts (Priest, 1990). Particularly, in the area of what has come to be known as "product liability," consumers' obligations began to diminish and the liability of manufacturers, distributors, and sellers began to grow. Thus, the rise of a more organized body of tort law in the early 1960s has led to the solidification of rules and procedures for handling the claims of those for whom great commercial and financial interests are at stake.

The end date of this study, 1988, coincides with the rise of advocacy for national tort reform. Because of the rise of tort claims and the accompanying expansion of liability insurance coverage, in the late 1980s, the insurance industry began to face what it claimed was a crisis. Because of the gravity of this situation, numerous state legislatures threatened or implemented modifications to tort laws to lower the costs paid by the insurers (Carroll and Pace, 1987). Furthermore, with the proliferation of asbestos lawsuits, the number of tort cases heard in federal court reached its high point in the mid-1980s. In fact, by 1985, tort-claims filings in federal court had begun to decline significantly. Between the period of 1985 and 2003, the number of tort cases decided in federal court fell by 79 percent (Cohen, 2005). Therefore, the period chosen represents the era in modern legal history when economic claims based on diversity jurisdiction reached a heightened state in the federal courts. If the need for such jurisdiction grew out of a historic desire to enhance the commercial capabilities of a young nation, then an examination of its relevance in the modern period should shed some light on the issues facing the legal community and political scholars alike. Finally, there is no reason to believe that conditions have changed since the period under study that would render any findings made or conclusions rendered currently inapplicable.

PROPOSITIONS AND MEASURES

Two broad propositions are examined in the present study. They are that 1 ) the invocation of federal diversity jurisdiction can potentially eliminate the bias toward in-state parties that would have arguably inhered in a decision rendered by the state court hearing a dispute between the in-state party and the out-of-state party, and 2) federal judges hearing diversity cases are applying the relevant state law necessary for resolution of the case at hand and are thereby holding true to principles of federalism by ensuring that causes of action in state law are being treated as state court cases.

Basically, it is posited in these propositions that courts-of-appeals judges in tort diversity cases apply the substantive laws of the applicable state to the case at hand in accord with the mandates of diversity jurisdiction, as set forth by statute and reinforced by the landmark decision in Erie Railroad v. Tompkins and its progeny. In a tort action, the applicable substantive state law is generally that of the jurisdiction where the event causing the injury to the plaintiff occurred, not where the defective product was made. The advocates of diversity jurisdiction are also postulated to have reason to believe that decisions rendered in federal court do, indeed, eliminate bias, for the in-state plaintiff in these diversity suits should enjoy no added advantage because of citizenship over the party from a different state had the litigation occurred in the plaintiff's home-state court.

To examine the first proposition, concerning whether the federal courts in tort diversity suits provide a neutral forum for those who would have been in-state (for example, an injured Pennsylvania citizen suing a Delaware defendant in the district court for the Middle District of Pennsylvania) and out-of-state citizens in state court litigation, two measures were created to determine whether there exists an in-state bias by the federal judges charged with applying the proper state law toward the nondiverse party (the in-state party) in the litigation. One, termed Same State, identifies whether the plaintiff was the non-diverse party (the in-state party) in the original filing of suit in the trial court. If the in-state plaintiff consistently found more favorable review in the federal court of because of his or her citizenship, then the grant of diversity jurisdiction did not provide the hoped-for neutral forum. However, if no relationship between residence status of the plaintiff and the judges' vote in the outcome of the case exists, then our federal judges may be acting as the unbiased adjudicators urged by the proponents of diversity jurisdiction.

The second measure, Same Judge, involves the state affiliations of the members of the three-judge courts-of-appeals panels who decided these tort diversity cases. This measure was created to determine whether a federal judge hearing diversity cases might be biased toward the party who possesses the same state citizenship as that judge. As it is typical that the judge appointed to the federal bench comes from a local community and had most likely practiced law in that very state, it is fair to assume that the courts-of-appeals judges will commonly maintain their citizenship within the jurisdiction of the relevant court and, likewise, be quite familiar with the statutory, constitutional, and common laws of that state. Likewise, courts-of-appeals judges typically maintain a home office in a federal building near a federal courthouse where district court judges and personnel are also employed and maintain offices. This variable was created by matching state affiliation of the courts-of-appeals judges found in the Auburn judge database2 to the state affiliation of the plaintiff in each tort diversity suit under study in this project. Although it is theoretically possible that a judge in a circuit court could be from the same state as a diverse party, instances of this were quite rare.

It is important to note that 50.6 percent of all appeals court panels in the data used had no in-state judges; 35.9 percent of all panels had one in-state judge; 10.4 percent of all the panels had two judges of the same state; and 3.1 percent of all appeals court panels had judges of the same state citizenship as the plaintiff. Once again, if a bias exists toward the in-state party to the litigation, there should be a relationship between the state of the judge and the non-diverse (or in-state) party. On the other hand, if the federal judge is acting in accord with the law of diversity jurisdiction, then there should not be a significant connection between either of these measures and the judges' votes.

To explore the second proposition, whether the federal courts are properly applying state law in tort diversity cases, two alternative measures of state law were used; these roughly reflect the degree of liberalism espoused by each state regarding important common-law tort policies. The first, labeled Tort Innovation, comes from a study used to test whether courts-of-appeals judges follow the law in addition to their own personal policy predilections (Songer, Ginn, and Sarver, 2003). This measure of state law is used to reflect each state's degree of liberalism regarding tort policies in the post-WWII period. The original tort innovation score for each state reported comes from a study analyzing the diffusion of twenty-three pro-plaintiff tort doctrines across the numerous state courts (Canon and Baum, 1981). Each state is ranked in relation to its degree of "innovation" in twenty-three pro-plaintiff common-law torts. The larger values for the Tort Innovation measure denote jurisdictions that were forerunners in adopting into their body of tort law important proplaintiff tort policies.

Tort Law, the other measure of state law used here, incorporates seven notable tort doctrines, which serve as a valid representation of the most highly litigated issues in state civil courts (Songer, Ginn, and Sarver, 2003). This variable was derived by summing the following doctrines (all weighted equally):

Tort Law = Limits of Recovery + Comparative Negligence + Collateral Source Rule + Frivolous Lawsuit Penalties + Right to Privacy (nonconstitutional) + Strict Liability + Negligence per se^sup 3^

For each of the elements of this linear combination, scores were ascribed depending on whether the policy was pro-defendant or pro-plaintiff. Higher values of the overall variable thus indicate that, on average, a state's overall approach toward tort policy is pro-defendant in nature, while lower values for the Tort Law variable represent those jurisdictions more likely to have laws designed to benefit plaintiffs. Because the year of adoption for each doctrine was coded for every state, the summary indicator of Tort Law is a dynamic measure of state law at any given time during the 1960-88 period used in my analysis. Essentially, the Tort Law variable and the Tort Innovation variable both roughly measure the degree to which each state is pro-plaintiff and pro-defendant in nature with regard to their overall tort policies. For example, Indiana is one of the most pro-plaintiff states as captured by these tort law measures. Thus, if a court-of-appeals judge is deciding a diversity case in which Indiana tort law is at issue, and since under Erie that judge is required to apply the relevant Indiana tort law, the decision reached should be relatively pro-plaintiff. Therefore, if statistically significant, either the Tort Innovation or Tort Law variable, or both, would indicate that federal judges are actually applying the substantive laws of the states to those tort cases that have been initiated in or removed to the federal court, so that principles of federalism would be upheld as the judges gave legally required deference to state law.

While it might be suggested that a reading of the actual cases contained in this analysis for evidence of citation to state law would be a better measure of whether courts-of-appeals judges are using state law, the method of coding described above is superior. Both the Tort Innovation variable and the Tort Law variable provide an objective measure of judicial adherence to state law, thus avoiding the difficulty of ascertaining the proper interpretation of state law, whether the appropriate state law cases have been cited, and whether those cases were cited properly in the opinion.

Because attitudes appear to play such a major role in judicial decision making on the Supreme Court (Segal and Spaeth, 1993, 2002), and also to an extent on the U.S. Courts of Appeals (Songer, Segal, and Cameron, 1994; Humphries and Songer, 1999; Songer, Sheehan, and Haire, 2000; Songer, Ginn, and Sarver, 2003; Hettinger, Lindquist, and Martinek, 2004), a measure of judicial ideology is added to test the assumption that the attitudes and personal policy predilections of lower federal court judges play a role in tort diversity cases in the courts of appeals. One can ascertain from the analysis of this measure whether appeals court judges are simply deciding cases in accord with their own values instead of faithfully applying state laws. This measure was obtained by first identifying all nonconsensual economic policy cases, except for tort cases, in the database. A case was defined as "nonconsensual" if the appeals court reversed the federal district court decision or if it affirmed the district court with a two-to-one vote on the court-of-appeals panel. A career score for each judge equal to the percent liberal of all their votes was then computed. This measure of judicial ideology appears to be an improvement over the more commonly used measure of party identification or appointing president, because it tends to capture state influences on the judicial selection process more directly.

Past studies have concluded that parties who possess superior material resources tend to experience greater success in trial court litigation (Galanter, 1974), on state supreme courts (Wheeler et al., 1987; Brace and Hall, 2001), and in the United States Courts of Appeals (Songer and Sheehan, 1992; Songer, Sheehan, and Haire 1999). Therefore, a control measure was created to test the effect of litigation resources in judicial decision making on the courts of appeals in tort diversity cases. For this Party Capability variable, numbers were assigned to each side based on a hierarchy in which national governments are presumed to have the most resources and litigation experience, followed by provincial governments, local governments, businesses, associations, and natural persons. The variable is actually a difference score: appellant party capability minus respondent party capability. The measure was constructed by assigning each party to a category based upon shared organizational characteristics that relate to resources and potential litigation experience. The categories are national government, provincial government, city/local government, business, organization or association, and natural person. While some of these categories are fairly general, this typology is consistent with the prior party-capability research (Brace and Hall, 2001; Parole, 1999; Sheehan, Mishler, and Songer, 1992; Songer and Sheehan, 1992; Songer, Sheehan, and Haire, 1999; Wheeler et al., 1987).

To ensure that any correlations between state law and judicial decisions were not simply a function of state liberalism, an independent control was also included for the dominant political values of each state in this analysis, Citizen Ideology. This was adopted from a study that created a measure of citizen ideology for all states for the period of 1960 to 1993 that in turn was derived from an analysis of the voting behavior of the members of Congress from each state in combination with an analysis of the partisan composition of state and national officeholders from the state (Berry et al., 1998). This approach is an improvement on the older approach in the scholarly literature of relying on region as a surrogate for political culture or values.

FINDINGS

As hypothesized, neither variable used to measure whether there exists a bias for the non-diverse (or in-state) party to a diversity suit, Same State and Same Judge, was related in a statistically significant fashion with how each appeals court judge voted in the tort diversity cases included in this analysis.4 Thus, it seems that the citizenship of neither the litigant nor the federal judge was found to exert any influence on the outcome of a tort diversity case in the courts of appeals. The expected lack of relationship between state affiliation of either party or judge lends support to the claim that the exercise of federal diversity jurisdiction tends to dissipate any bias that might exist against out-of-state litigants, at least in these important tort cases.

Both alternate measures of state law, Tort Innovation and Tort Law, were related to an appeals court judge's application of state law in diversity cases, and the relationship was statistically significant-the Tort Innovation variable in the hypothesized positive direction, and the other measure, Tort Law, in the hypothesized negative direction. These findings support the hypothesis that federal court judges are adhering to the mandates of diversity jurisdiction and are thereby applying the relevant statutory, constitutional, and common laws of the fifty states to the case at bar in federal court. Ideology was related to the courts-of-appeals judges' votes in tort diversity cases in the hypothesized positive direction, and the relationship was also statistically significant. Such a finding suggests, consistent with earlier work, that perhaps there is a relationship between judicial attitudes and decision making in tort diversity cases in the courts of appeals (Songer, Ginn, and Sarver, 2003).

Both the elements included as controls performed as predicted. Results for the Party Capability variable indicate that while the nature of litigants' resources affects judicial decision making to the extent that those with superior material resources tend to come out ahead, the effects of the main factors are not negated by its presence. Adding the Citizen Ideology measure does not measurably alter the other findings,5 as it was not related to courts-of-appeals judges' votes in tort diversity cases in the hypothesized direction in a statistically significant manner.

CONCLUSIONS

With the grant of diversity jurisdiction in the Federal Judiciary Act of 1789, the early proponents of states' rights became fearful that the ability to bring suits, especially those of a commercial nature, in the federal courts would threaten the workings of a new American federalism. However, these courts were given the direction to apply "the laws of the several states" in deciding suits of this nature. The notion that the newly created national courts could entertain important economic litigation, previously heard in the state courts, began the origins of a controversy that would surround the legal community for years to come. When the Court allowed for the creation of a "federal common law" in diversity cases, with its decision in Swift v. Tyson, old fears were again resurrected. Not until the overruling of Swift in Erie Railroad v. Tompkins were the lower federal courts once again warned of their duty to apply applicable state substantive laws. Through this command, it was hoped that federal judges would be relieved of any pressures that would tend to favor in-state litigants over their out-of-state opponents. However, the debate over the propriety of retaining diversity jurisdiction remained, with valid arguments both for and against its retention. This article has contributed to this debate by providing empirical evidence of judicial practice regarding the resolution of the major type of diversity cases, tort cases that appear before the federal courts.

The finding that federal court judges do not tend to favor non-diverse (in-state) litigants over the diverse party suggests that the prerogative of invoking diversity jurisdiction goes a considerable way to eliminate the potential for state court bias, real or perceived. Similarly, the finding that the citizenship of sitting federal judges also fails to influence judicial determinations lends support to the claim that federal courts appear to be the unbiased, neutral tribunals as envisioned in the original grant of diversity jurisdiction in the Judiciary Act of 1789. Thus, those litigants who choose to sue in federal court or to have their cases removed there based on diversity jurisdiction seem to have reason to believe that a real option exists in their choice of forum for claim adjudication.

Furthermore, the finding that state tort law does play a great role in the decision-making process of federal judges reinforces the claim that the mandates of diversity jurisdiction are being adhered to in these important tort diversity cases. That is, judges tend to apply the appropriate applicable state statutory, constitutional, and common law to the issue on appeal. These judges of the U.S. Courts of Appeals, arguably presented with an opportunity to exercise great discretion in deciding such cases as a result of their unique positions as the typical final arbiters of the law in tort diversity cases, instead seem to be adhering to the dictates of federal diversity law by applying the proper state laws, attitudes aside.

Those members of the private bar who support federal-court-diversity jurisdiction now have empirical evidence to back their arguments for its retention. For those scholars and critics alike who are suspicious that along with diversity jurisdiction comes abuse of judicial federalism, they too should rest assured. While the grant of diversity jurisdiction undoubtedly does give to the national government a vast amount of power, the federal judiciary appears to give more than mere lip service to the laws and policies of the individual states. A respect for the underlying ideals of federalism, in the faithful application of relevant state law, and a concern for equitable treatment of all litigants to a suit, seems to have motivated the judges of our federal appellate courts to adhere to a body of law, often criticized, in contrast to simply rendering their own personal judgments in a manner disregarding deeply held American legal principles and standards. If bias was meant to be resolved through the creation of diversity jurisdiction, then it appears that these federal judges (with state ties of their own) are concerned with this command and appear to take it very seriously.

Naturally, the research reported here has limitations, but it also provides the opportunity for further studies. As the United States Courts of Appeals Data Base only includes published opinions, and as diversity cases are less likely than other case types to be published (Wasby, 2005), only a small set of actual diversity decisions rendered are examined here. However, if courts-of-appeals judges appear to be constrained by law and precedent in tort diversity cases (Songer, Ginn, and Sarver, 2003), and because only published opinions have precedential effect and most likely represent the greatest ideological voting, it is safe to assume that the same judicial adherence to the applicable state law in question is still present in the unpublished opinions. A future project could assess whether these findings do in fact also apply to unpublished tort diversity opinions. Further research could also examine whether findings regarding elimination of bias and judicial adherence to diversity jurisdiction remain constant across different categories within the universe of tort diversity cases. Perhaps more important, an examination of tort diversity cases in the United States District Court could also be undertaken since it is in the district courts that decisions for or against litigants are initially made. Although it is true that courts of appeals, as the final arbiters of the law in tort diversity cases, are an appropriate place to study the questions raised here, an examination of these cases at the federal trial court level, where judges could arguably be susceptible to greater charges of parochial ties, is likely to prove fruitful for this important debate as well.