The Influence of Amicus Curiae Briefs on U.S. Supreme Court Opinion Content

Abstract

We address fundamental questions about the ability of interest groups to shape public policy by examining the influence of amicus curiae briefs on U.S. Supreme Court majority opinion content. We argue that the justices will incorporate language from amicus briefs into their opinions based on the extent to which the amicus briefs contribute to their ability to make effective law and policy. Using plagiarism detection software and other forms of computer assisted content analysis, we find that the justices adopt language from amicus briefs based primarily on the quality of the brief's argument, the level of repetition in the brief, the ideological position advocated in the brief, and the identity of the amicus. These results add fresh insight into how interest groups influence the development of federal law by the Supreme Court.

To understand constitutional law in the United States, judges and scholars point to a variety of sources. Some judges stress the need to focus on the preferences or words of the framers (Bork 1990; Scalia 1997), while others offer the view that the Constitution should be understood according to more contemporary values (Breyer 2005). Academics argue that constitutional law develops by other means, including through long term constitutional conversations (Friedman 1993), interbranch interactions (Meernik and Ignagni 1997), and social movements (Ackerman 1991). While each of these perspectives offers insight into American constitutional development, arguably the most important vehicle for comprehending constitutional law is found in the majority opinions of the Supreme Court (Shapiro 1968). Through its opinions, the Court establishes appropriate norms of behavior, provides guidelines regarding the constitutionality of particular programs, and affords direction to lower court judges and future Supreme Courts who are charged with adjudicating disputes touching on similar factual circumstances. Our purpose here is to contribute to our understanding of Supreme Court opinions by examining the extent to which amicus curiae (“friend of the court”) briefs influence their content.

Investigating this topic is significant for a number of reasons. First, it sheds new light on the ability of interest groups to shape the Court's policy outputs. Although there has been no shortage of research on friends of the court, scholars have overwhelmingly examined the ability of amici to influence case outcomes or the justices’ voting behavior in those cases (e.g., Black and Boyd 2013; Box-Steffensmeier, Christenson, and Hitt 2013; Caldeira and Wright 1988; Collins 2008a; Kearney and Merrill 2000). While these are important avenues for study, they do not address the content of the Court's opinions, which is the most significant means by which the Court contributes to legal and social policy. For example, the amicus brief filed by the American Civil Liberties Union in Craig v. Boren (1976) was instrumental in the Court's adoption of the intermediate scrutiny standard for evaluating gender discrimination claims (Campbell 2002). Such highly significant forms of influence would be missed by extant research that focuses on case outcomes, but can be captured using the content analysis techniques used in the current analysis.

More broadly, this research is valuable in that it contributes to our understanding of how organized interests use language to shape public policy. Traditionally, interest group scholarship has focused on the ability of groups to affect policy outputs by focusing primarily on the presence or amount of lobbying activity (for a review, see Hojnacki et al. 2012). Although significant, this research has generally ignored the ability of interest groups to use language to influence the content of public policy. These techniques are ubiquitous across legal and political venues, and include proposing rules to bureaucratic agencies, submitting model statutes to legislatures, and filing legal briefs with judicial bodies. Here, we join a growing group of scholars in examining how interest groups leverage language in their advocacy efforts (e.g., Chien 2011; Klüver 2009; Pedersen 2013).

Third, this topic is important because it provides insight into how justices craft the content of their opinions.¹ At the Supreme Court, the four primary sources of information presented to the justices are the opinions of the lower courts that initially handled the case, litigant briefs, amicus curiae briefs, and oral arguments (Stern et al. 2002). Although we have a fairly sophisticated understanding of why the justices integrate language from lower court opinions (Corley, Collins, and Calvin 2011), litigant briefs (Corley 2008), and oral arguments (Johnson 2004) into their opinions, research on the influence of amicus briefs on opinion content is more limited. For example, existing scholarship reveals that the justices incorporate the arguments from amicus briefs into their opinions (Epstein and Kobylka 1992; Samuels 2004) and cite amicus briefs in those opinions (Kearney and Merrill 2000; Owens and Epstein 2005), although we have yet to develop a thorough understanding as to why some amicus briefs are relied on and others are ignored. Consequently, while we can document examples of amicus influence on opinion content, the reason for that influence has generally eluded us. We remedy this by providing a systematic analysis into the extent to which the justices incorporate the language of more than 2,000 amicus briefs into their opinions.

Finally, this research is significant because it demonstrates a technique that has the potential to help bridge the divide between the two dominant methodological approaches to understanding the Supreme Court (and social science more generally). On the one hand, traditional approaches to studying the Court involve the in-depth analysis of the Court's opinions, often focusing on a single case or series of interrelated cases (e.g., Gillman 1993). On the other hand, other scholars study the Court using statistical techniques to analyze hundreds or thousands of decisions in a single investigation (e.g., Segal and Spaeth 1993). Each of these approaches has its own strengths and weaknesses. For example, while traditional research strategies provide in-depth insight into the Court's opinions, it is not clear how much one can generalize from findings related to a single case. Conversely, although large-scale quantitative studies offer a good deal of generalizability, these studies tend to oversimplify the Court's opinions by breaking down decisions into a single dichotomous variable (such as whether the opinion was conservative or liberal). The present study builds on recent work that applies plagiarism detection software to legal texts (Black and Owens 2012; Collins, Corley, and Hamner 2014; Corley 2008; Corley, Collins, and Calvin 2011) by investigating the influence of amicus briefs on Supreme Court opinion content. This approach combines the generalizability offered by quantitative studies with the depth provided by more traditional qualitative approaches that focus on the words used in judicial opinions. Although far from perfect, we are confident this approach can be applied to great effect in an effort to better understand the Supreme Court and the content of legal and political texts more generally.

Investigating Amicus Influence

There are a variety of ways that amicus curiae briefs can influence the Supreme Court. At the agenda setting stage, amicus briefs highlight the significance of appeals, increasing the likelihood that the justices will review a case (Black and Boyd 2013; Caldeira and Wright 1988). At the merits stage, amicus briefs are associated with litigation success (Collins 2004; Kearney and Merrill 2000), as well as with the ideological direction of the Court's decisions and the justices’ votes (Box-Steffensmeier, Christenson, and Hitt 2013; Collins 2007, 2008a). Moreover, amicus briefs enhance the likelihood that the justices will author concurring or dissenting opinions (Collins 2008b). Here, we focus our attention on what is arguably the most important avenue for influence: the ability of amicus briefs to shape the content of the Court's majority opinions.

A limited body of scholarship demonstrates that amici are capable of influencing the substance of the Court's opinions (Epstein and Kobylka 1992; Samuels 2004; Spriggs and Wahlbeck 1997). We make a unique contribution to this literature using computer assisted content analysis techniques to investigate the ability of amicus briefs to contribute to the content of the Court's opinions by comparing the language used in amicus briefs with the language used in the Court's majority opinions. We argue that when the Court borrows language from amicus briefs in its opinions, that is evidence that the amicus brief influenced the Court's opinion (e.g., Landes, Lessig, and Solimine 1998). Thus, when the Court's opinion uses the same language found in an amicus brief, that amicus brief has affected the substance of the opinion, contributing to the development of federal law. While this does not necessarily mean that the amicus brief has influenced the decision of the Court (Segal and Spaeth 1993), whether it be conceptualized in terms of its outcome (e.g., reverse or affirm) or its ideological direction (e.g., conservative or liberal), it nonetheless demonstrates that the amicus brief has influenced the behavior of the justices, as revealed in the Court's most significant policy output: its majority opinion.

To understand how amicus briefs can influence the content of Supreme Court majority opinions, it is important to recognize two key features of this relationship. First, amicus briefs seldom contain a recitation of the facts of a dispute, focusing instead on providing the justices with persuasive argumentation that gets to the substance of the legal and policy issues involved in a case (Collins 2008a; Ennis 1984; Lynch 2004). Thus, when language from amicus briefs is incorporated into majority opinions, it overwhelmingly appears in the substance of the opinion, not in the treatment of facts of that typically opens an opinion.

Second, the justices rarely adopt language from amicus briefs into their opinions for the purpose of criticizing that language. This separates amicus briefs from litigant briefs and the opinions of the lower courts who initially handled a case, which are often disparaged in Supreme Court opinions (Corley 2008; Corley, Collins, and Calvin 2011). The reason for this distinction is straightforward: the justices are expected to engage the arguments—positively and negatively—of the litigants and lower courts, while no such expectation exists with respect to amicus briefs (e.g., Lynch 2004). Instead, if a justice finds fault with an argument in an amicus brief, the justice can disregard that brief altogether. Simply put, it is not an effective use of a justice's time and energy to dress down an amicus brief in a majority opinion when that brief can just be ignored.

To illustrate, we extracted a random sample of 60 amicus briefs from our data and investigated whether the language adopted in majority opinions was done so for the purpose of criticizing it.² To do this, we read the paragraph surrounding each matched phrase (in both the amicus brief and majority opinion) to establish the context in which the language appeared in the majority opinion. We then coded whether the phrase was integrated into the opinion for the purpose of rejecting the argument. Of the 1,032 matched phrases, only 0.68 percent (7) were negatively incorporated into majority opinions. Thus, it is clear that the justices rarely adopt language from amicus briefs for the purpose of criticizing the arguments made by the amici.

With this in mind, the language used in amicus briefs can be incorporated into Supreme Court opinions in a number of ways. A particularly persuasive argument might influence how the justices interpret the constitutional or statutory provision at issue, leading the Court to adopt the position or standard of review advocated in the brief. In addition, a justice might borrow the amicus’ treatment of existing precedent. Because the outcomes of cases are heavily influenced by governing precedent (e.g., Bailey and Maltzman 2011), this can be a particularly important form of influence, indicating that the amicus contributed not only to the content of the opinion but also to the disposition of the case. The justices might also adopt information in amicus briefs that speaks to the broader economic, legal, and policy implications of a decision (Epstein and Kobylka 1992). Further, amicus briefs often contain information regarding the preferences of other actors, such as the framers, state legislators, congress, and the president. To the extent that the justices view this information as beneficial (e.g., Bailey and Maltzman 2011), they might incorporate this material into their majority opinions. What is more, because amicus briefs are the primary source of social scientific information at the Court (e.g., Rustad and Koenig 1993: 94), the justices might engage those data in majority opinions, again providing evidence of amicus influence. Finally, the justices might incorporate language from especially well-written amicus briefs as a way to make an efficient use of their time and energy (e.g., Corley, Collins, and Calvin 2011). In this way, interest groups can subsidize the justice's opinion writing, while at the same time contributing to the content of Supreme Court opinions.

As these examples demonstrate, Supreme Court justices can incorporate the arguments made by amici into their majority opinions for a variety of purposes. Because the justices seldom adopt information from amicus briefs into their opinions for the purpose of criticizing that information, and because the language in amicus briefs overwhelmingly appears after the recitation of facts in the opinion, we do not differentiate between these various types of influence. This allows us to provide a large-scale investigation into amicus influence on opinion content that extends beyond the more limited studies that explore how amici contributed to the development of particular issue areas or cases (Epstein and Kobylka 1992; Samuels 2004) or focus on how repetition relates to amicus influence (Spriggs and Wahlbeck 1997).

Amicus Influence on Majority Opinion Content

We argue that Supreme Court justices are motivated by both legal and policy goals (e.g., Bailey and Maltzman 2011; Baum 1997). The justices seek to write legally sound opinions that promote their personal policy preferences and enhance the coherency of federal law. The primary means to pursue these dual goals is through the Court's majority opinions, which set precedents that constrain the behavior of lower courts, future Supreme Courts, and the elected branches of government. By authoring high quality opinions, these objectives are furthered in that sound legal opinions are more likely to be followed by lower courts and other actors (e.g., Owens and Wohlfarth 2012), in addition to contributing to a favorable view of the justice's jurisprudential legacy (e.g., Baum 2006).

To achieve these goals, the justices require information to assist them in maximizing their legal and policy preferences. As discussed above, amicus curiae briefs supply the justices with a wide array of such information, potentially improving the quality of their opinions. However, it is important to note that not all amicus briefs are created equal. Accordingly, we believe that the justices will look to attributes of the amicus briefs in determining the extent to which they will incorporate the arguments made in the briefs into their majority opinions. In particular, we argue that the justices will be interested in four key features of amicus briefs.

First, we believe that the justices will assess the quality of an amicus brief based on its cognitive clarity and use of plain language. To be sure, clarity is important in the law. A judicial opinion must explain its rationale to the bench and bar and establish guideposts for future cases. To do this most effectively, that opinion must be clearly written (Owens and Wedeking 2011; Owens and Wohlfarth 2012). Similarly, amici attempting to persuade the Court to endorse the position and reasoning advocated in the brief are best suited to do so by writing clear briefs. Cognitively clear amicus briefs are more compelling because they better enable the justices to ascertain the precise issue and logic advanced in the brief. Moreover, such briefs demonstrate that the amicus is knowledgeable, skilled at making the necessary legal analysis, and is able to analyze the issues “in a clear, logical, step-by-step way, such that the reader sees how the writer reached his or her conclusions and, ideally, agrees with them” (Baker 2012: 302). Indeed, Justice Scalia has highlighted the importance of clarity in brief writing, advising attorneys to “[v]alue clarity above all other elements of style” (Scalia and Garner 2008: 107). Since cognitively clear amicus briefs facilitate the justices’ comprehension and acceptance of the briefs’ arguments, we believe the justices will rely more on these briefs because they enable the justices to write clear opinions that promote their legal and policy preferences. Simply put, we propose that brief clarity promotes opinion clarity. Thus, we expect that the justices will incorporate more language from cognitively clear amicus briefs.

The use of plain language may also indicate a high quality brief. Since at least the time of Cicero, concerns over the use of verbose writing have been raised in a variety of quarters. Members of the legal community have been especially vocal, urging lawyers, and judges to avoid legalese in favor of writing that is designed to ensure that the reader understands the material as quickly and completely as possible (e.g., Baker 2012; Benson and Kessler 1987; Flammer 2010; Garner 2009). For example, Justice Ginsburg has stated that she “can't bear [legalese]. I don't even like legal Latin. If you can say it in plain English, you should” (Garner 2010: 141). Similarly, in response to excessive prose, Justice Thomas has advised that “the genius is having a ten-dollar idea in a five-cent sentence, not having a five-cent idea in a ten-dollar sentence” (Garner 2010: 100). Empirical studies support these sentiments, evincing that judges find briefs written in plain language to be substantively stronger and more persuasive than those written in legalese (e.g., Benson and Kessler 1987; Flammer 2010).³ Based on this, we posit that the justices will latch onto language in amicus briefs that are written in plain English since those briefs are likely to be viewed as more compelling than excessively verbose briefs. Thus, we hypothesize that the justices will incorporate more language from amicus briefs that rely on plain language.

Second, we expect that the justices will consider the degree to which the brief reinforces arguments raised by other entities or provides primarily novel information. On the one hand, practitioners, law clerks, and justices themselves make it clear that amicus briefs should avoid repeating information that is otherwise available to the Court (e.g., Ennis 1984; Lynch 2004; O'Connor 1996: 9). Supreme Court Rule 37 speaks exactly to this point: “An amicus curiae brief that brings to the attention of the Court relevant matter not already brought to its attention by the parties may be of considerable help to the Court. An amicus curiae brief that does not serve this purpose burdens the Court, and its filing is not favored.”

On the other hand, research on repetition and persuasion suggests that repetition can have positive effects. Studies show that repetition increases familiarity, which is used as a cue to judge the validity or truth of a statement (e.g., Arkes, Hackett, and Boehm 1989; Boehm 1994; Schwartz 1982). Indeed, even a single repetition can make an argument appear more valid (Hasher, Goldstein, and Toppino 1977). Additionally, research has shown that the persuasiveness of messages composed of weak arguments can be enhanced by the repetition of those arguments (Garcia-Marques and Mackie 2001; but see Cacioppo and Petty 1989). In essence, these studies demonstrate that decision makers are more likely to agree with arguments presented to them multiple times since the repetition of those arguments makes them appear more valid, irrespective of the objective validity of the argument. As it relates to amicus curiae briefs, this literature suggests that justices are more likely to view the arguments in repetitious amicus briefs as valid since they encounter those arguments in multiple information sources (i.e., lower court opinions, party briefs, and other amicus briefs). Insofar as the justices seek to make good law and policy, we believe that they will be especially likely to adopt repetitious arguments in their opinions since they will view those arguments as valid, thus, better enabling them to pursue their goals (see also Spriggs and Wahlbeck 1997). Thus, we expect that, as the level of repetition in an amicus brief increases, the justices will incorporate more language from that brief.

Third, we propose that the justices will examine whether the position advanced in the amicus brief corresponds to their ideological preferences. We expect that justices will adopt more language from ideologically congruent amicus briefs into their majority opinions for three reasons. First, research in social psychology indicates that decision makers are more likely to accept arguments consistent with their preferences as valid, while discounting argumentation that undermines their preexisting beliefs (e.g., Kunda 1990; Lord, Ross, and Lepper 1979). Insofar as judges are susceptible to the same cognitive biases as the rest of us (e.g., Guthrie, Rachlinski, and Wistrich 2007), we expect that they will view the arguments in ideologically congruent amicus briefs favorably and incorporate those arguments into their opinions. Second, because the justices are motivated by their ideological preferences (Segal and Spaeth 1993), and because majority opinions are the primary means for justices to etch their preferences into law, we believe they will seek out information in ideologically congruent briefs to ensure majority opinions reflect their preferences. Amicus briefs that are consistent with a justice's ideology can play a particularly important role by alerting the justice to the broader policy implications of a decision, helping ensure that the opinion will maximize the justice's preferences. Third, by relying on the language in ideologically congruent amicus briefs, a justice can make an effective use of his or her time. In adopting the arguments made in amicus briefs that are consistent with the justice's preferences, the need to engaging in additional research to craft the majority opinion is lessened. Such is the case because the justice is likely to be more confident in the validity of the amicus’ argument than he or she would be in an amicus brief that runs counter to the justice's ideology. Thus, we expect that the justices will incorporate more language from ideologically congruent amicus briefs into their opinions.

Finally, we anticipate that the justices will consider the identity of the amicus brief filer, paying particular attention to amicus briefs filed by the U.S. Solicitor General, the American states, and elite interest groups. As the lead attorney for the executive branch in the Supreme Court, the Solicitor General stands out from other amici in a number of ways. Formally, the Solicitor General has an office at the Supreme Court, wears a morning coat at oral argument (if male), and does not have to obtain the permission of the litigants or the leave of the Court to file amicus briefs. Informally, the Court grants substantial deference to the Solicitor General. This is reflected in the frequent invitations from the Court to participate as an amicus and the high rates of success the office enjoys at the certiorari and merits stages, as well as in its ability to influence the doctrinal content of the Court's opinions (Bailey, Kamoie, and Maltzman 2005; Black and Owens 2012; Caldeira and Wright 1988). Because the Solicitor General speaks on behalf of what is arguably the most significant interest in the county—the federal government—we believe the justices will pay particular attention to the arguments advanced in the Solicitor General's amicus briefs. By relaying the views of the executive branch to the Court, the Solicitor General can improve the quality of the justices’ opinions, enabling them to craft those opinions with an eye toward how they will impact the coherency of federal law and the interests of the nation as a whole. Given this, we expect that the justices will incorporate more language from the Solicitor General's amicus briefs.

Just as the justices are concerned with the views of the federal government as represented in the Solicitor General's amicus briefs, we expect they will likewise pay close attention to the perspectives of state governments as represented in state attorneys amicus briefs. Institutionally, this is reflected in the fact that, like the U.S. Solicitor General, states are not required to obtain the permission of the litigants or the Court to file amicus briefs. Through their amicus briefs, states are able to provide the justices with information regarding how the Court's opinions will impact state and local governments, who often play a primary role in the implementation of the Court's decisions (Clayton 1994). In addition, state amicus briefs often supply the justices with details involving state regulatory, law enforcement, and civil rights policies, allowing the justices to better understand the effect of their decisions at the subnational level (Waltenburg and Swinford 1999). Moreover, state attorneys general have a reputation for authoring high quality amicus briefs, particularly since the 1980s (Clayton 1994; Waltenburg and Swinford 1999), and Supreme Court law clerks have identified state amicus briefs as second only to those filed by the U.S. Solicitor General in terms of receiving special consideration (Lynch 2004: 48). Thus, insofar as state amicus briefs are capable of improving the justices’ understanding of legal disputes, and, therefore, potentially enhancing the quality of their opinions, we expect that they will incorporate more language from amicus briefs filed by state governments.

Lastly, we consider the influence of elite, private amici on majority opinion content. Although such organizations do not speak on behalf of government interests, they are nonetheless capable of attracting the justices’ attention as a result of having reputations as credible and reliable information sources (Galanter 1974; McGuire 1993). Due to this status, we expect the justices will carefully examine the content of their amicus briefs, viewing that content as more compelling than amicus briefs filed by other interests. Such is the case because the justices are likely to perceive that elite, repeat player amici possess accurate legal and political arguments that better enable the justices to pursue their legal and policy goals through their majority opinions. Indeed, there is evidence that the justices and their law clerks heed closer attention to amicus briefs filed by certain high status organizations that are respected across chambers, regardless of the ideological orientation of the amici (Lynch 2004). Thus, we expect that the justices will incorporate more language from amicus briefs filed by elite amici into their majority opinions.

Data and Methods

To investigate the factors that contribute to the extent to which the justices incorporate language from amicus curiae briefs into their majority opinions, we collected the texts of Supreme Court majority opinions and amicus briefs during the 2002–2004 terms.⁴ After initially identifying the Court's majority opinions in the Spaeth (2007) database using the docket as the unit of analysis, we then downloaded each majority opinion from Westlaw and secured the amicus briefs filed in each case from Findlaw, Lexis-Nexis, or Westlaw. The dataset includes 229 orally argued, signed, majority opinions and 2,016 amicus briefs. Majority opinions from cases without amicus curiae participation are excluded from our analysis.

The unit of analysis in the data is the amicus curiae brief. Each amicus brief is tied in an observation to the majority opinion in the case in which the brief was filed. Our dependent variable is the proportion of the majority opinion that is based directly on language found in an amicus brief. To calculate this, we relied on the plagiarism detection software, WCopyfind 4.1.1 (Bloomfield 2013), which allows us to compare two documents for the purpose of establishing the extent to which they share common words in phrases. Although originally created to investigate plagiarism by college students, WCopyfind has been successfully extended to a variety of social science applications, including investigations of media coverage of presidential speeches (Eshbaugh-Soha 2013), agenda setting in the senate (Grimmer 2010), and the content of Supreme Court opinions (Black and Owens 2012; Corley 2008; Corley, Collins, and Calvin 2011).

To promote comparability to these studies, we programmed WCopyfind's settings-based largely on its recommended (default) parameters, which is the established standard in the literature (e.g., Black and Owens 2012; Collins, Corley, and Hamner 2014; Corley 2008; Corley, Collins, and Calvin 2011; Eshbaugh-Soha 2013). We set the shortest string of words to match at six, meaning that the program does not record matches of five words or fewer. To permit the program to identify matches in spite of slight editing in the language, we set the minimum percentage of matches that a phrase can contain at 80; allowed it to bridge its way across up to two nonmatching words (“imperfections”) as it connects pieces of identically matched phrasing; and instructed the program to ignore letter case, numbers, outer punctuation, and nonwords (the latter of which has the effect of skipping over case citations).

Because WCopyfind is intended to capture the extent to which two or more documents share common words in phrases, it measures the repetition of language in amicus briefs and majority opinions. It cannot, however, capture instances in which the majority opinion adopted the argument of an amicus, but framed that argument in different language than the amicus. In addition, the plagiarism detection software cannot ascertain whether the opinion adopted language from an amicus brief in the substance of the opinion or in dicta, nor can it identify positive or negative treatments of amicus briefs in opinions.⁵ Nonetheless, we believe that the benefits of using WCopyfind outweigh its limitations. In particular, the program allows us to validly and reliably compare thousands of amicus briefs to majority opinions to determine the extent to which the justices incorporate the language from amicus briefs into their majority opinions, thus, shedding new light on the influence of amici curiae on the Court.⁶

An example of the repetitive language reported by the plagiarism software from Clackamas Gastroenterology v. Wells (2003) is shown below. We have italicized the language in the majority opinion that was adopted from the amicus brief.

Amicus Brief of the US and Equal Employment Opportunity Commission

In particular, a court should examine whether shareholder-directors operate independently and manage the business or instead are subject to the firm's control. If the shareholder-directors operate independently and manage the business, they are proprietors and not employees; if they are subject to the firm's control, they are employees.

Majority Opinion

[The EEOC] argues that a court should examine “whether shareholder-directors operate independently and manage the business or instead are subject to the firm's control.” According to the EEOC's view, “if the shareholder-directors operate independently and manage the business, they are proprietors and not employees; if they are subject to the firm's control, they are employees.” [citations omitted, italics added]

To gain a sense of the frequency with which the justices incorporate language from amicus briefs into their majority opinions, Figure 1 presents a box plot of the dependent variable. This figure is demarcated by justice, and also contains the overall proportion of majority opinions that adopt language from amicus briefs, labeled SC. The lower portion of the box is the 25 percentile and the upper portion represents the 75 percentile. The black dot in the middle of the box is the median value. The upper and lower adjacent values are indicated by the whiskers above and below the box, and outlier values are represented by the dots above the upper adjacent value.

Figure 1. Proportion of U.S. Supreme Court Majority Opinions from Amicus Curiae Briefs, by Justice (2002–2004 Terms). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Across all justices, the mean of the dependent variable is 0.027, with a s. d. of 0.035 (range = 0–0.35). This indicates that majority opinions incorporate less language from amicus briefs than party briefs and lower court opinions. For example, Corley (2008) reports that majority opinions borrow an average of 9.8 percent of language from party briefs, and Corley, Collins, and Calvin (2011) reveal that 4.3 percent comes from the opinions of the lower federal courts who initially handled the litigation (compared with an average of 2.7 percent for amicus briefs). This is an interesting finding as it indicates that majority opinions are more focused on the direct parties to litigation and the lower courts being reviewed than amicus briefs. This makes sense in that the litigants and lower courts are primarily responsible for establishing the case record, while amici typically enter the litigation only when it reaches the Supreme Court and, thus, have more limited opportunities to shape how the case is framed.

Figure 1 also reveals that there is some pronounced variation among the justices. Justice Thomas borrows most heavily from amicus briefs, with 0.044 of his majority opinions using language that appears in amicus briefs. Following Thomas, Justices Ginsburg (0.035), Scalia (0.034), O'Connor (0.031), Breyer (0.031), and Rehnquist (0.031) make the most use of language in amicus briefs, while Kennedy (0.025), Souter (0.021), and Stevens (0.018) adopt the least amount of language from amicus briefs. Looking at the outlier values, it is clear that the opinions that integrated the most language from amicus briefs are fairly evenly dispersed among the justices. The highest proportion corresponds to Rehnquist's opinion in Connecticut Department of Public Safety v. Doe (2003) (0.35), followed by Souter in Breuer v. Jim's Concrete (2003) (0.31), and Ginsburg in Intel Corporation v. Advanced Micro Devices (2004) (0.25). All of these amicus briefs were filed by the federal government. Among nongovernmental amici, briefs filed by the American Federation of Labor and Congress of Industrial Organizations in Desert Palace v. Costa (2003) (0.17), the American Bar Association et al. in Beneficial National Bank v. Anderson (2003) (0.16), and the National Automobile Dealers Association in Koons Buick v. Nigh (2004) (0.16) saw a good deal of language adopted in the majority opinions of Thomas, Stevens, and Ginsburg, respectively. Notably, 21 percent of amicus briefs witnessed no incorporation of their language into majority opinions.

To subject our hypotheses to empirical scrutiny, we have coded our independent variables as follows. First, to assess brief quality, we use two variables that tap into the caliber of the argumentation in the brief: Cognitive Clarity and Plain Language. To measure the cognitive clarity of each brief, we use the automated content analysis program, Linguistic Inquiry and Word Count (LIWC). LIWC is a dictionary-based program, meaning that it contains lists of words that correspond to separate dictionaries that represent a larger concept. LIWC was developed by psychologists to quantify a variety of concepts, such as the expression of emotions, cognitive thought processes, and other language dimensions (Tausczik and Pennebaker 2010). We adopt our measure from Owens and Wedeking (2011) using 10 LIWC categories that are directly connected to cognitive clarity: causation, insight, discrepancy, inhibition, tentativeness, certainty, exclusiveness, inclusiveness, negations, and the percentage of words containing six or more letters. We standardized and collapsed these categories into a single variable and then multiplied the values by negative one, meaning that higher values represent more cognitively clear argumentation. We expect this variable will be positively signed, indicating that the justices incorporate more language from cognitively clear amicus briefs.⁷

Our second measure of brief quality captures each brief's use of Plain Language. Using LIWC, we calculated the inverse of the average words per sentence in each amicus brief.⁸ The operationalization of this variable follows Garner's (2009: 297) advice that a key to plain language is to “[a]chieve a reasonable average sentence length,” as well as Justice Thomas’ position that it is preferable to convey ideas in shorter sentences (Garner 2010: 100). Moreover, it is consistent with Benson and Kessler's (1987) survey of judges and attorneys who rated appellate briefs with very long sentences with many embedded clauses as less convincing and less persuasive than plain language briefs. Because higher values indicate more use of plain language, we expect this variable will be positively signed, indicating that the justices adopt more language from briefs that use plain language.⁹

To measure the level of repetition in the amicus briefs under analysis, we include three variables. Using the WCopyfind settings outlined above with respect to the dependent variable, these variables are the percentage of each amicus brief that is based on other written information sources in the case. Litigant Repetition is the level of repetition between the amicus brief and the brief of the litigant the amicus supported. Amicus Repetition is the level of repetition between the amicus brief and the other amicus briefs that supported the same litigant in the case. Lower Court Repetition is the average level of repetition between the amicus brief and all lower court opinions in the case. We expect these variables will be positively signed, indicating that, as the amount of repetition increases, the justices will incorporate more language from the amicus brief. We also expect that the substantive effect of the Litigant Repetition and Lower Court Repetition variables will be largest based on previous research that demonstrates the extent to which the justices borrow from party briefs and lower court opinions (Corley 2008; Corley, Collins, and Calvin 2011).

To capture the ideological compatibility of the amicus brief with the Supreme Court's majority opinion, we measured the ideological congruence between the amicus brief and the majority opinion author, based on the opinion author's Martin and Quinn (2002) score and the ideological direction of the position advocated in the amicus brief (e.g., Bailey, Kamoie, and Maltzman 2005). If the amicus brief advocated for a conservative outcome, this variable is the positive value of the majority opinion author's Martin and Quinn score. If the amicus brief advocated for a liberal outcome, this variable is the negative value of the majority opinion author's Martin and Quinn score. We expect the Ideological Congruence variable will be positively signed, indicating that the justices adopt more language from ideologically compatible amicus briefs.

To examine whether amicus briefs filed by particular amici are especially influential, we include three binary variables in the model. Solicitor General Amicus is scored 1 if the amicus brief was filed by the Office of Solicitor General, and 0 otherwise. State Amicus is coded 1 if the amicus brief was filed by one (or more) of the American states, and 0 otherwise. To operationalize our Elite Amicus variable, we rely on Lynch's (2004) survey of Supreme Court law clerks who served from 1966 to 2001. This variable represents the private (i.e., nongovernmental) organizations that law clerks identified by name as those whose amicus briefs are “considered more carefully than others” (Lynch 2004: 51). Thus, this variable reflects the reputations of the amici based on the views of the law clerks who initially process the amicus briefs. This variable is scored 1 for amicus briefs filed by the American Civil Liberties Union, AFL-CIO, Brennan Center for Justice, Chamber of Commerce, Criminal Justice Legal Foundation, Lambda Legal, National Association for the Advancement of Colored People, NAACP Legal Defense and Educational Fund, and the Washington Legal Foundation, and 0 otherwise.¹⁰ We expect these variables will be positively signed, revealing that the justices incorporate more language from amicus briefs filed by these amici.

Finally, we include a series of control variables in the model to account for other factors that might influence the extent to which the justices incorporate language from amicus briefs into their opinions. First, we use a Case Salience variable to capture the possibility that justices will spend more time and effort crafting the content of opinions that dispose of salient cases, as compared with relatively trivial litigation, and will, therefore, rely less on the language in amicus briefs. This variable represents the log of the number of words spoken by the justices at oral argument, standardized by the number of justices participating in oral argument, and is, thus, an a priori measure of a case's importance (Black, Sorenson, and Johnson 2013). We anticipate that this variable will be negatively signed. Second, we include a Brief Length variable that is calculated as the log of the number of words used in each brief. Because longer briefs provide more opportunity for the justices to borrow from the brief, we expect this variable will be positively signed. Third, we control for the information environment at the Court. We expect that, in cases with a high volume of amicus briefs, the extent to which the justices incorporate the language from any one amicus brief will be decreased (Black and Owens 2012: 103). To test this, we use an Amicus Brief Volume variable, which is composed of the number of amicus briefs filed in the case. We expect this variable will be negatively signed. Fourth, we consider the degree to which the majority opinion adopts language from sources of information other than amicus briefs. We believe that an opinion that relies on lower court opinions and party briefs will incorporate more language from amicus briefs, signifying that the majority opinion author is devoting less effort to crafting such an opinion. Percent from Other Sources represents the average percentage of the Court's majority opinion adopted from the opinions (majority, concurring, and dissenting) of the lower courts who initially handled the case and the briefs of the litigants, based on the WCopyfind settings discussed above.¹¹ We expect this variable will be positively signed. Finally, to account for possible differences among the justices, we include a binary variable for each justice, save one (results not shown).

Results

Our dependent variable—the proportion of the majority opinion based on an amicus brief—ranges from 0 to 0.35 and values below 0 and above 1.0 are infeasible since this variable is a proportion, making the use of ordinary least squares regression inappropriate. Accordingly, we use a fractional logit model, which is a quasi-likelihood method estimated as a generalized linear model, that allows us to account for the nature of this variable (Papke and Wooldridge 1996). To account for the fact that majority opinions appear in the data multiple times, we use robust s. e.'s, clustered on docket number. Table 1 reports the results, revealing support for our hypotheses.

Table 1. The Influence of Amicus Curiae Briefs on U.S. Supreme Court Majority Opinions, 2002–2004 Terms

The unit of analysis is the amicus curiae brief. The dependent variable is the proportion of the Supreme Court opinion taken from each amicus curiae brief. Entries are fractional logit regression coefficients. Numbers in parentheses are robust s. e.'s, clustered on docket number. The expected direction of the coefficients of the independent variables appears in brackets. Percent change indicates the percentage increase in the dependent variable corresponding to a 0–1 increase in dichotomous variables and a one s. d. increase in continuous variables, holding all other variables at their mean or modal values. p values are based on one-tailed tests. The model includes eight justice-specific dummy variables (results not shown).

First, we find that the justices incorporate more language from higher quality amicus briefs into their majority opinions, based both on the brief's cognitive clarity and its use of plain language. Substantively, a one s. d. increase in cognitive clarity results in a 13 percent increase in the majority opinion's reliance on an amicus brief (from a baseline prediction of 0.018 to 0.021). With regard to an amicus brief's use of plain language, we find that a one s. d. increase leads to a 6 percent increase in the percentage of the majority opinion that incorporates language from the brief (0.018 to 0.019). Thus, it is clear that the justices assess the quality of amicus briefs, integrating more language from higher caliber briefs into their opinions.

We also find that the justices adopt more language from amicus briefs that repeat information contained in the opinions of the lower courts that initially handled the litigation, litigant briefs, and other amicus briefs. A one s. d. increase in an amicus brief's repetition of the litigant brief it supports produces a 30 percent increase in the extent to which the majority opinion incorporates language from the amicus brief (0.018 to 0.024). A one s. d. increase in repetition of other amicus briefs and lower court opinions results in a 5 percent (0.018–0.019) and a 9 percent (0.018–0.20) increase in the dependent variable, respectively. Moreover, statistical tests indicate that the justices incorporate more language from amicus briefs that repeat the briefs of the litigants they support and lower court opinions, as compared with amicus briefs that repeat other amicus briefs filed on the same side of the case. These are particularly interesting findings in light of the fact that judges and practitioners have admonished amici for providing repetitious information. It seems that, rather than ignore repetitious arguments, the justices are likely to view repeated arguments as valid and integrate those arguments into their opinions, and are especially likely to do so if the amici repeat arguments made in litigant briefs and lower court opinions. While inconsistent with what we might expect given the advice of judges and practitioners to avoid repetition, this finding is supported by research on repetition and persuasion. These studies demonstrate that repetition can increase the extent to which message recipients view a message as credible and valid (e.g., Arkes, Hackett, and Boehm 1989; Boehm 1994; Garcia-Marques and Mackie 2001; Hasher, Goldstein, and Toppino 1977; Schwartz 1982).

Table 1 also reveals that majority opinion authors evaluate the ideological position advocated in amicus briefs. A one s. d. increase in the ideological congruence between an amicus brief and the majority opinion writer results in a 4 percent increase in the adoption of language from that brief (0.018 to 0.019), although the p value for this variable is 0.06. This evinces the role of ideology in the opinion writing process at the Court.

Our results also provide evidence that the justices consider the identity of the amici. This is especially evident with regard to the influence of the Solicitor General on the content of majority opinions. Compared with other amici, majority opinions adopt 135 percent more language from amicus briefs filed by the Solicitor General (0.018 to 0.043). Moreover, the justices are more likely to embrace information from amicus briefs filed by state governments and elite organizational interests. Compared with other amici, the justices incorporate 13 percent more language from state amicus briefs (0.018 to 0.021) and 24 percent more language from amicus briefs filed by high status amici (0.018 to 0.023). These are especially significant findings in light of the fact that we control for the quality of the amicus briefs under analysis. This suggests that the Court does view certain elite interests differently than other amici and looks favorably on the information supplied by these high status amici.

Turning now to the control variables, we find that the justices incorporate less language from amicus briefs in salient cases. A one s. d. increase in the (standardized) number of words uttered by the justices at oral argument results in a 31 percent decrease in the majority opinion's reliance on an amicus brief.¹² Thus, it appears that the justices spend more time and energy crafting opinions in salient cases, relying less on amicus briefs. The justices also incorporate more language from longer amicus briefs. A one s. d. increase in the (logged) number of words used in an amicus brief leads to a 37 percent increase in the amount of language in that brief incorporated into the majority opinion. The information environment at the Court also plays a role. A one s. d. increase in the number of amicus briefs filed in a case produces a 19 percent decrease in the amount of information the justices integrate from any one amicus brief. Finally, Table 1 indicates that the more the justices borrow information from nonamicus sources (lower court opinions and litigant briefs), the more they borrow from amicus briefs. Substantively, a one s. d. increase in the majority opinion's reliance on these other sources results in a 25 percent increase in the percentage of the majority opinion taken from an amicus brief. This demonstrates that when the justices adopt language from one source of information, they are more likely to borrow from other information sources.

While the individual effects of certain variables are somewhat small, the cumulative effect of several variables can be quite large. To illustrate, consider how the Court's reliance on language from amicus briefs increases from the baseline predicted value of 0.018 when we vary aspects of the amicus briefs that are manipulable by interest groups. To do this, we have made one s. d. alterations in the brief's cognitive clarity, plain language, level of repetition, length, and ideological orientation—factors that are determined by the amici and their counsel. When we make these changes, the Court's adoption of language from the amicus brief increases to 0.045. Thus, it is clear that, under certain conditions, amicus briefs can have a substantial influence on Supreme Court opinions, evincing the ability of friends of the Court to contribute to the development of federal law. Moreover, this influence is substantively significant—an increase of 0.01 in the majority opinion's incorporation of language from an amicus brief is about 110 words on average—and within the control of organized interests and their attorneys.

Conclusions

Given the importance that Supreme Court opinions play in our understanding of the development of federal law, it is not surprising that scholars have dedicated a great deal of research to investigating the factors that shape the substance of those opinions. We have contributed to this significant line of inquiry by analyzing the influence of amicus curiae briefs on majority opinion content. The novel data we offer reveals important insights into this topic, demonstrating that the justices systematically incorporate language from amicus briefs into the Court's majority opinions based on their perceptions as to whether those briefs will enhance their ability to make effective law and policy. We find that the justices borrow more language from high quality amicus briefs that, in turn, better enable them to author high quality majority opinions. Justices also incorporate more language from amicus briefs that repeat arguments advanced in other information sources, suggesting they are more likely to view that information as credible. Moreover, the justices adopt more language from amicus briefs that correspond to their ideological preferences, and those filed by high status interest groups.

In addition to enhancing our understanding of Supreme Court opinion content, this research also illuminates the ability of interest groups to shape public policy. To be sure, there has been substantial research on organizational involvement at the Court. However, existing studies overwhelmingly focus on how interest groups influence case outcomes and the justices’ voting behavior (e.g., Black and Boyd 2013; Box-Steffensmeier, Christenson, and Hitt 2013; Caldeira and Wright 1988; Collins 2008a; Kearney and Merrill 2000). While significant, these studies do not inform our understanding of the chief manner in which the Court makes policy: its majority opinions. By investigating this topic, we add to our understanding of foundational questions about the power of interest groups in the American political system. For example, our evidence that brief quality and the status of amici factor into the extent to which the justices rely on amicus briefs speaks to long-standing concerns about the differential influence of organized interests in the political system (e.g., Schattschneider 1960). Indeed, it appears that elite interests are better able to marshal their resources and submit higher caliber amicus briefs, which are given more deference by the justices. This is, thus, one of the first studies to provide evidence of bias with respect to interest group influence on the content of Supreme Court opinions.

This research also provides guidance for the authors of amicus briefs. For example, our results reveal that the justices adopt more language from amicus briefs that repeat arguments and are written in less complex language. The former finding suggests that, despite criticisms of repetition in amicus briefs from the justices and their clerks (Lynch 2004; O'Connor 1996: 9), a reasonable amount of repetition can convince the justices of the merit of an argument. The latter finding supports the advice of those advocating for attorneys to use simpler language in their briefs (e.g., Garner 2009). However, these results should not be interpreted to mean that amici should submit briefs that entirely restate the arguments of the litigants they support or briefs that are written at a kindergarten level. Legal training and norms set boundaries on what is considered acceptable in an amicus brief and the amount of repetition and plain language appearing in the briefs under analysis here reflects these norms. Thus, our sample does not contain such briefs and, therefore, cannot speak to the extent to which the justices would adopt language from fully repetitive briefs or those written at a kindergarten level.

This article also corroborates the benefits of using automated content analysis to understand the behavior of legal and political actors. We are particularly excited about the application of plagiarism detection software to the study of law since it has the potential to help bridge the gap between large-n studies that tend to ignore the content of judicial opinions and case studies that focus specifically on the content of those opinions. Of course, computer assisted content analysis can never replace the insights that can be gained from the close reading of legal and political texts. But nor is its purpose to do so. Instead, techniques such as those advanced here can augment more qualitative research methods by providing a valid and reliable manner to analyze a large corpus of texts, thus, helping to better inform our comprehension of a host of substantively interesting topics.

While we believe there are many benefits of using computer content analysis techniques, it is also important to recognize their limitations. Although using an automated software program allows us to reliably and validly compare thousands of briefs with each other and with the Court's majority opinions, there is obviously no way to know exactly which arguments and language the majority opinion would have contained in the absence of the amicus briefs. Additionally, the software does not allow us to determine the percentage of a majority opinion that can be solely attributed to the content of amicus briefs.¹³ We have addressed this to some extent through our inclusion of the variables that capture the amount of repetition in the amicus briefs, which allows us to determine the degree to which repetition influences the Court's adoption of language from amicus briefs. Thus, as with other research strategies, there are both strengths and weaknesses of automated content analysis.

Finally, this research further demonstrates the value of adopting theories and methods developed in a variety of disciplines to tackle significant law and society research questions. Theoretically, we have drawn on insights from social psychology to understand how both argument repetition and ideology relate to persuasion. To better comprehend how judges view argument quality, we turned to research on plain language, which is informed by a variety of disciplines with a common goal of improving the quality of communication in a diverse array of contexts. Methodologically, we have adopted computer assisted content analysis techniques originally developed to understand the psychometrics of words and to detect plagiarism in our effort to understand how amicus briefs shape the content of Supreme Court opinions. Given their utility, we encourage future researchers to further explore the application of interdisciplinary theories and methods, which are capable of providing fresh insight into a wide range of questions that are important to legal scholars.