Introduction
Scientific information is deeply relevant to many judicial proceedings, but the extent to which judges incorporate this type of information into their decision-making is unclear. Judicial decisions in the United States are crafted to follow a commonly accepted structure, citing legal principles, doctrine, and authorities such as statutory language and precedent as the basis for the outcome reached (Federal Judicial Center 2013). Judges are trained to use sources outside this list—often referred to as secondary sourcesFootnote 1—“sparingly” (Federal Judicial Center 2013, 18).
The strictures of this traditional approach to opinion writing have led judges, like Supreme Court Justice Stephen Breyer, to advocate for their colleagues to more aggressively engage in building up a working understanding of scientific processes relevant to judicial decision-making (Malakoff Reference Malakoff2022). Breyer explains that it is vital for judges to engage with science, emphasizing “It is equally true that the law itself increasingly requires access to sound science. This need arises because society is becoming more dependent for its well-being on scientifically complex technology, so, to an increasing degree, this technology underlies legal issues of importance to all of us” (Reference Breyer1998, 537). Indeed, according to Breyer, “scientific issues permeate the law” and “science should expect to find a warm welcome, perhaps a permanent home, in our courtrooms” (Breyer Reference Breyer2000, 52). Breyer argues that, as the gatekeepers of evidence, courts handling criminal cases must consider the validity and reliability of scientific evidence, such as DNA results or statistical evaluations used to predict “future dangerousness” (Breyer Reference Breyer2000, 53). Although a judgeFootnote 2 is not a scientist and a courtroom is not a scientific lab, the law should aim for decisions grounded in scientifically reliable knowledge (Breyer Reference Breyer2000).
Breyer’s perspective on the need for an interdisciplinary exchange between science and law has been increasingly adopted by other judges. Judges apply science across diverse legal contexts, both procedurally in civil proceedings and evidentiary hearings and substantively in nearly all areas of law, including criminal, civil, administrative, and constitutional law (Faigman Reference Faigman2006). Additionally, Carlson (Reference Carlson2013, 844) argues that attorneys tend to rely more on science during pretrial evidentiary hearings because it is easier and “highly probative” for them to establish their case by presenting reputable texts and articles that support the expert’s testimony. Scientific claims hold significance due to science’s high status and cultural authority in the United States (Moyer Reference Moyer2022; Layzer Reference Layzer2012). Courts often require attorneys to present scientific literature to corroborate their claims of scientific reliability and the courts in turn will closely review and scrutinize the content of the publication (Carlson Reference Carlson2013). Hence, scientific information matters and publications presented to corroborate or challenge scientific evidence are discussed in judicial opinions.Footnote 3
For judges, the need to apply or engage with science is especially critical when considering forensic evidence in criminal cases. Individuals in the scientific community are working to make expert information more accessible to those in the legal community. For example, when forensic science methods and protocols are updated, this information must be translated for decision-makers, like judges, who are not trained to holistically understand scientific information (Garrett et al. Reference Garrett, Gardner, Murphy and Grimes2021; Breyer Reference Breyer1998). In the forensic science field, a common method of information sharing is through national high-level reports prepared by scientific and legal experts. Judges have taken note of these reports even going so far as to cite these secondary sources in their decisions, signaling their importance. For example, Brim v. State (2000) explains that “the supreme court itself reviewed and relied upon the 1996 report by the National Research Council (NRC), to determine that…calculation of population frequency statistics are no longer necessary.” In another case, State v. Marcus (1996), the opinion notes that “The general acceptance within the scientific community of DNA analysis…is clearly shown by authoritative scientific literature…importantly, a 3-year study by a National Research Council.” These examples suggest judges are using the information provided by national forensic reports to inform their understanding of scientific evidence and that this expert information is, at least some of the time, guiding legal decision-making.
Judges must regularly make consequential decisions in criminal cases while lacking training in relevant scientific areas like forensics. Most judges will need to handle criminal issues regularly, depending on the nature of their docket. But there is no norm or requirement that judicial hopefuls work in criminal law before they come to their seat on the bench. Very few judges have formal criminal law experience or have served as public defenders, but the number is growing (Harris and Sen Reference Harris and Sen2026). This lack of experience in criminal law and related scientific training can leave a decision maker with a gap in knowledge about forensic evidence commonly used in criminal cases (Cooper et al. Reference Cooper, Shooter, Scanlon and Yates2026).
Further, when judges need to make decisions in criminal cases specifically, they may face added pressure from the weight of the consequences for the defendant. Criminal consequences are not limited to a monetary penalty but their outcomes can carry lengthy jail sentences or, in more extreme cases, even the death penalty. Crime is also often an extremely salient policy issue in the American political context, adding to the pressure on judges. For the past three decades, Americans have believed that crime is increasing in the United States, despite official statistics from the Federal Bureau of Investigations showing a significant decline (Gramlich & Eddy, Reference Gramlich and Eddy2024; Morgan & Thompson, Reference Morgan and Thompson2022). Media outlets often run an outsized amount of stories about crime leaving the public to overestimate crime rates, feeding the impression that crime is high (Budd & Ghandnoosh, Reference Budd, Ghandnoosh, Budd, Dillaway, Lane, Muschert, Nair and Jason2024). As a result, judges have an incentive to approach decision-making in criminal cases with extra attentiveness and caution—leveraging all tools available to them—including the scientific expertise presented in national forensic reports.
In this article, we marshal an interdisciplinary-informed approach to the scholarly discourse on how judges reference expert-produced information in their decision-making process and specifically how expert information appears in opinion writing. Are judges following Breyer’s recommendation to engage seriously with scientific information? We find that they are. We analyze how judges engage with national forensic reports in their written opinions for criminal cases by considering the amount of authoritative language used when referencing reports and whether the use of authoritative language varies based on whether the reference supports one or more of the parties in the case. We find that judges are using authoritative language when engaging with national forensic reports in their written opinions. Specifically, our analysis indicates that judges write with more authoritative language when their writing benefits the defense. We see some evidence that judges also write with authoritative language when referencing forensic reports benefiting the prosecution, but this finding is mixed and the relationship is often negatively associated in our models. These patterns reveal that prosecutors may be at a disadvantage with how judges cite scientific information from national forensic reports and may want to take this informational pattern into account as they prepare their cases.
This study makes at least two key contributions to the literature. First, our findings shed light on the conditions under which judges directly engage with scientific information when they cite national forensic reports—a key example of an expert-produced secondary source—in their written opinions. We show that the example references to forensic reports in judicial decisions quoted above are not isolated instances. Rather, judges formally engage with scientific expertise in their opinions for criminal cases and do so in systematic ways, but our study’s approach is conditional on whether scientific evidence is central to the case and whether specific reports are referenced.
Second, the analysis further illuminates how judges explain and justify their decisions in criminal cases to the parties and the public more generally by referencing expert-produced information. Citing national forensic reports in judicial opinions may help communicate about the judicial decision-making process to the parties and the public, due to the high regard in which science is held in the United States (Moyer Reference Moyer2022; Layzer Reference Layzer2012). Traditional legal training emphasizes the primacy of sources of law (e.g., statutes) and precedent in judicial opinions. Given this context, the evidence we provide of judges citing national forensic reports and the types of language judges use when writing about these reports within opinions in criminal cases implies that judges may recognize the utility of scientific information and may extend deference to this outside expertise. In other words, this study provides a first step toward evaluating if judges are adopting Justice Breyer’s position that good judicial decision-making, at appropriate times, requires scientific information.
This article proceeds with a broad overview of the role forensic science plays in the criminal legal system. We then review perspectives from the policy process and legislative literatures on problem solving and information seeking in decision making and connect them to literature on judicial behavior. Third, we consider the factors that motivate judicial decision-making and the literature developed around how judges craft their opinions. Next, we lay out our hypotheses and provide a description of our research design and the novel dataset of forensic report references in criminal judicial opinions. Then, we test our hypotheses and discuss our results. Finally, we consider the future of a research agenda in judicial information-seeking behavior and the connection between the law and science.
The Role of Forensic Science in the Criminal Legal System
Forensic science is used extensively within the criminal legal system. First, forensic evidence (e.g., DNA, fingerprint, or ballistics evidence) frequently plays a critical role in criminal investigation and adjudication, with crime laboratories performing various analyses for federal, state, and local law enforcement agencies (Julian et al. Reference Julian, Kelty, Roux, Woodman, Robertson, Davey, Hayes, Margot, Ross, Sibly and White2011). And, the demand for these forensic techniques is high. The Department of Justice (2023) reports that publicly funded crime laboratories and multi-laboratory systems received more than 3.3 million requests for service in 2020. Second, forensic evidence can also shape the severity of the punishment convicted defendants receive. For example, Peterson et al. (Reference Peterson, Sommers, Baskin and Johnson2010) find that convicted defendants linked to a crime by forensic evidence are more likely to receive longer sentences.
Intuitively, judges recognize the critical role played by forensic science within the criminal legal system. Previous work demonstrates that judges view forensic evidence as credible and persuasive (e.g., Wechsler et al. Reference Wechsler, Cramer, Kehn, Wosje, Boccaccini and Varela2015; Nir and Liu Reference Nir and Liu2021). However, because judges are not typically trained as forensic scientists, their decisions are shaped by the general reputation of evidentiary forms as well as factors such as an expert witness’ ability to communicate (Nir and Liu Reference Nir and Liu2021). Judges may, therefore, find themselves needing expert information and contemporary guidance regarding new forensic techniques and ongoing debates about how to properly interpret forensic evidence.
Recognizing the importance of forensic science and its reach, as well as the debates surrounding its use, high-level national reports that aimed to focus on various or specific forensic science methods have been published by national science advisory entities. Importantly, these reports were constructed with and for the legal and scientific communities, further indicating the substantial connection between the two fields. The reports provide accessible explanations of the scientific principles and practices of various forensic disciplines. Some reports contained critiques of laboratory practices and methods, expert conclusions, and statistical methods, emphasizing technical issues and limitations and providing recommendations for improvement. Conversely, other reports endorse and provide justification for the use of specific practices and methods. Forensic science and legal experts compiled these critiques and endorsements and translated complex concepts into these reports in a way that is understandable and useful for legal actors. These reports were designed as a versatile tool for litigation involving forensic science that could be utilized by either the defense or prosecution. These reports have altered the way forensic testing is conducted and the way forensic evidence is presented in court, giving them, by extension, the potential to shape judicial decision-making in criminal cases. In what follows, we examine the extant literature on how decision makers engage in expert information seeking and consider how it applies to the realm of judicial decision making.
Theories of Political Decision Making and Expert Information Seeking Behavior
For our theory, we adopt the problem-solving model from Baumgartner and Jones’ (Reference Baumgartner and Jones2015) book, The Politics of Information. One major feature of the problem-solving model is that policymakers need to seek information from experts in order to make decisions. Complex problems are the norm in policymaking—and in the legal process. When a policymaker wants to try to solve a problem (especially a complex one), they will likely find themselves insufficiently prepared to do so (Workman Reference Workman2015). Baumgartner and Jones’ (Reference Baumgartner and Jones2015) model describes two types of information-seeking behavior—entropic and expert searches. Entropic searches occur when a policymaker devotes themselves to identifying the problems that will require more of their attention down the road and are ongoing in nature (Boydstun Reference Boydstun2013; McCubbins and Schwartz Reference McCubbins and Schwartz1984). Only after that problem is identified and defined does the search for expert information begin.
When a problem becomes both complex and important, policymakers will start to seek out someone (or someones) that has the requisite experience to provide this important information (Baumgartner and Jones Reference Baumgartner and Jones2015). Thus begins the “expert-search” stage of decision-making. Often the policymaker will not have to go far to find expert information or they might already find themselves bombarded with it by an interest group or lobbyist ready to provide a version of the sought-after information that supports their own institutional goals and serves their own group members (Hall and Deardorff Reference Hall and Deardorff2006; Jones and Baumgartner Reference Jones and Baumgartner2005; Baumgartner and Leech Reference Baumgartner and Leech1998). Also standing at the ready are less high-profile organizations that produce specialized information tailormade for policymakers faced with tough decisions like the Congressional Research Service (Fagan and McGee Reference Fagan and McGee2022). Though interest groups and lobbyists come with their own agendas, they can often provide expertise, data, and evidence useful to policymakers.Footnote 4
One important utility of expert information for policymakers at this stage is that it can break down key information into smaller pieces. This allows the policymaker to digest and use the information to understand the problem more clearly despite lacking expertise (Baumgartner and Jones Reference Baumgartner and Jones2015). Notably, when policymakers are presented with (sometimes too much) information, policymakers do not always seem to be weighing more expert-driven and higher quality information more heavily than other types of information (Patashnik and Peck Reference Patashnik and Peck2017). Our study examines what judges do with the information they have, focusing explicitly on the use of authoritative language when judges incorporate expert-produced scientific forensic information in their written opinions.
Extending the logic of the problem-solving model of decision making to the study of judges requires two assumptions. First, in the context of this study, the problem in a criminal case is defined by the prosecutor as they are the entity that chooses if and which charges will be brought against someone accused of a crime.Footnote 5 Second, once the problem is defined (i.e., the charges are filed and/or a trial or appellate process concluded), the judge must engage in the expert search stage of the problem-solving model when they work to write their opinion, should they determine there is a reason to use secondary sources.
Since the expert search stage of the problem-solving model happens after the problem has been defined, this stage is where information about how best to solve the problem is most valuable (Baumgartner and Jones Reference Baumgartner and Jones2015). Issue-area experts, like those who developed the national reports about forensic sciences, are usually ready and waiting with the necessary information to solve the problem, either in the form of testifying as an expert witness (which is very expensive for the litigating party employing them) or in the form of the educational forensic reports we utilize in this study (which can be accessed at no cost).
We extend the problem-solving theory to judicial behavior. When writing opinions, judges may engage in an information-seeking process with a similar goal as their legislative counterparts albeit with the mechanics operating a little differently (Hazelton, Hinkle, and Spriggs Reference Hazelton, Hinkle and Spriggs2019). Judges too have the power to solve a problem by deciding a case (in the example of a non-jury decision) and are making these decisions within institutional constraints like limited resources and full dockets. They must at least appear to consider evidence presented by both sides including witness and expert witness testimony. National forensic reports may have been cited by one of the parties and may, therefore, be consulted. Judges are not bound to only the information presented in the case by the parties, though. When writing opinions involving complex problems judges might want to go further and use a secondary source even if that source was not cited by the prosecution or the defense. When scientific information is particularly relevant to a case’s substance (like forensic DNA evidence in a sexual assault case), judges may feel especially motivated to seek out expert information like national forensic reports. Here, we argue judges can use their discretion and may seek out information to assist their understanding and assessment of forensic evidence from information presented within the case (e.g., through witnesses, briefs, etc.), as well as beyond. We think this is especially important when the criminal charges against defendants carry more arduous consequences not limited only to imprisonment.
Existing work in the judicial politics literature has already established the importance of how judges evaluate and use information in their decision-making process. A judge’s own level of expertise in an area of law will also influence how much information they seek. Information comes to the court primarily through 1) party briefs, 2) amicus briefs, or 3) secondary sources. When judges are more knowledgeable about an issue, they are also more likely to confidently identify when litigants present arguments opposed to their preferences (Nelson, Oxley, and Clawson Reference Nelson, Oxley and Clawson1997). When judges have low levels of issue-area expertise, they may be more persuadable and also more likely to rely on sources they deem credible (Szmer and Ginn Reference Szmer and Ginn2014). At the U.S. Supreme Court level, the justices rely on information from lawyers deemed more “capable” who are well-known to the Court (Corley Reference Corley2008; McGuire Reference McGuire1995) and who are more capable of writing persuasive arguments based on their knowledge of the judicial system and deeper knowledge of the factors that would sway judges (Brodie Haire et al. Reference Brodie Haire, Hartley and Lindquist1999; Szmer, Johnson, and Sarver Reference Szmer, Johnson and Sarver2007; Wahlbeck Reference Wahlbeck1997). Amicus briefs can provide useful and often unique information to judges (Collins, Corley, and Hamner Reference Collins, Corley and Hamner2014, Reference Collins, Corley and Hamner2015). Additionally, judges are more likely to turn to nonparty-supplied information (often from amici) when they feel the quality of parties’ legal representation was not strong (Simard Reference Simard2008). Further, amici view their own information as more likely to influence judicial decision-making when legal representation for the parties is weak (Hansford Reference Hansford2004). Studies evaluating persuasion more generally found when the information from highly credible sources is more persuasive generally (Tormala, Brinol, and Petty Reference Tormala, Briñol and Petty2007; Tormala & Petty Reference Tormala and Petty2004; Kumkale, Albarracin, and Seignourel Reference Kumkale, Albarracín and Seignourel2010) and subject-matter experts’ attempts to persuade their targets are more successful when targets are not subject-matter experts (Miller and Mauet Reference Miller and Mauet2011). Similar patterns are well-documented in judicial behavior (Hazelton and Hinkle Reference Hazelton, Rachael and Hinkle2022; Szmer and Ginn Reference Szmer and Ginn2014; Lane and Schoenherr Reference Lane and Schoenherr2025). We build on the literature of how judges use a variety of sources in their decision-making process to include how they write about scientific information in criminal law opinions.
It is important to note here that while the national forensic reports we consider in this study are prominent examples of secondary sources produced by issue-area experts, they are not the only source of expert information available to judges to guide their understanding of forensics. Expert witnesses are still commonly used to directly explain the relevant science to judges in individual cases. Additionally, judges may elect to attend trainings designed and developed by forensics experts specifically to help judges strengthen their knowledge of forensic methods and concepts (Garrett et al. Reference Garrett, Gardner, Murphy and Grimes2021). We opt to focus on national high-level forensic reports since they are created for and available to a broad audience across cases and jurisdictions. This allows us to include a wide range of criminal judicial opinions (e.g., state, federal, trial, and appellate) in our analysis and provides insight into the larger question of how judges engage with expert information when making decisions.Footnote 6
Writing Judicial Opinions and Citation to Expert Information
Judges are trained to take opinion writing seriously (Federal Judicial Center 2013). They are taught that opinions should “fairly, clearly, and accurately” present significant facts (Federal Judicial Center 2013). As a result, judges need reliable and accurate information. Lebovits, Curtin, and Solomon (Reference Lebovits, Curtin and Solomon2008) state that “what a judge writes enhances confidence in the judiciary and brings justice to the litigants, or not. The heart of a judge’s reputation and function rests with the use of the pen.” Judges strive to write sound and well-researched opinions as these documents are the only mechanism by which they can communicate to other legal actors, political actors, and the broader public (Corley and Wedeking Reference Corley and Wedeking2014). In order to craft these documents, judges can use or reference information supplied to them from lower courts, litigants, as well as from other experts and interest groups (Spriggs and Wahlbeck Reference Spriggs and Wahlbeck1997; Collins Reference Collins2008a; Collins Reference Collins2008b; Box-Steffensmeier, Christenson, and Hitt Reference Box-Steffensmeier, Christenson and Hitt2013; Hansford and Johnson Reference Hansford and Johnson2014; Hazelton and Hinkle Reference Hazelton, Rachael and Hinkle2022; Canelo Reference Canelo2022; Lane and Schoenherr Reference Lane and Schoenherr2025).
Judges motivate their decision-making process following not only their desire to get it right, but also their desires to maintain their own legacy and their position’s prestige (Baum Reference Baum1994; Reference Baum2008). Further, they want their decisions to be interpreted correctly and implemented beyond a singular case (Baum Reference Baum2008; Black et al. Reference Black, Owens, Wedeking and Wohlfarth2016b; Wedeking and Zilis Reference Wedeking and Zilis2018; Corley, Collins, and Calvin Reference Corley, Collins and Calvin2011). Relying on inadequate or inaccurate information may lead to a decision reversal, damage their reputation, and compromise their potential to be elevated to a higher judicial seat. Writing an opinion that is based on accurate and reliable facts works in the opposite direction, bringing positive evaluations from peers and higher courts. In fact, the adequacy of an opinion might be a reason why it is chosen for review by a higher court (Hitt and Lempert Reference Hitt and Lempert2024).
Writing a good opinion involves numerous decisions on the part of the opinion author. The judicial politics literature has been investigating how and why judges make different kinds of writing choices in their opinions for over a decade. Studies have considered textual clarity (Owens and Wedeking Reference Owens and Wedeking2011); emotional language (Bryan and Ringsmuth Reference Bryan and Ringsmuth2015; Black et al. Reference Black, Hall, Owens and Ringsmuth2016a; Boyd, Collins, and Ringhand Reference Boyd, Collins and Ringhand2023; Budziak, Hitt, and Lempert Reference Budziak, Hitt and Lempert2019; Hitt and Lempert Reference Hitt and Lempert2024; Owens, Wedeking, and Wohlfarth Reference Owens, Wedeking and Wohlfarth2013); readability—typically measured by a combination of length of words and the length of sentences (Black et al. Reference Black, Owens, Wedeking and Wohlfarth2016b); and even quality (Budziak and Lempert Reference Budziak and Lempert2023; Hitt and Lempert Reference Hitt and Lempert2024). Combined, these studies demonstrate the utility of close, systematic analysis of judges’ opinion writing.
While judges on lower courts are also at least somewhat constrained in their decisions by other factors, such as precedent (Bueno de Mesquita and Stephenson Reference Bueno de Mesquita and Stephenson2002; Landes and Posner Reference Landes and Posner1976; Jacobi and Tiller Reference Jacobi and Tiller2007), we can gain important insights by examining the manner in which they craft opinions (e.g., Maltzman, Spriggs, and Wahlbeck Reference Maltzman, Spriggs and Wahlbeck2000). One major component of a judge’s discretion when authoring an opinion is which authorities to cite and how to cite them. In doing so, judges consider their own goals and audiences (Corley, Collins, and Calvin Reference Corley, Collins and Calvin2011). For example, Supreme Court opinions are more likely to include lower court opinion language when the lower court judge is more “prestigious” (Corley et al. Reference Corley, Collins and Calvin2011). Or, when a federal circuit court judge cites a Supreme Court opinion that is currently preferred by the Supreme Court negatively in their own opinions, the circuit court judge’s writing complexity increases because the lower court judge is working to conceal the out-of-favor citation from the Supreme Court as to evade review (Boston Reference Boston2020). The converse is true as well. These types of patterns are at the core of our inquiry.
Drawing on the problem-solving model, we theorize that judges will engage in expert searches to contend with scientific information, which may be unfamiliar and technical (Baumgartner and Jones Reference Baumgartner and Jones2015). One potential way to capture if judges are engaging in expert search is by looking at the secondary sources they reference in their written opinions along with the language they use when doing so. Here, we use a text-as-data approach to evaluate the degree to which a judge uses “authoritative language” (following Corley and Wedeking Reference Corley and Wedeking2014) in conjunction with references to national forensic reports in criminal cases. Authoritative language is often conceptually associated with strong conviction, confidence, and even absolutism that one’s position is correct—even sometimes a “bravado” (Boyd et al. Reference Boyd, Collins and Ringhand2023). Thus, judges might turn to authoritative language to emphasize a point about which they are writing. The norm of citing secondary sources sparingly within judicial opinions suggests that judges will carefully choose how to reference national forensic reports. This then in turn suggests that we can gain insight into how judges think of and use this expert information.
Based on the literature, we expect that when judges reference expert information from national forensic reports in their written opinions, they may use authoritative language as a form of emphasis for a few possible reasons. First, judges may be using authoritative language as a justificatory device. In other words, the authoritative language judges employ surrounding the reference may signal the relevance or importance of a national forensic report, and/or justify the inclusion of expert scientific information contained in a secondary source. For example, many of the observations in our data use authoritative language when referencing a national report to support or challenge a premise stated by the prosecution or defense. This theoretical mechanism extends previous work indicating judges use their discretion in the opinion writing process to integrate arguments and citations that support their conclusion (Spriggs and Wahlbeck Reference Spriggs and Wahlbeck1997; Collins Reference Collins2008a; Collins Reference Collins2008b; Hazelton and Hinkle Reference Hazelton, Rachael and Hinkle2022; Canelo Reference Canelo2022; Spriggs and Hansford Reference Spriggs and Hansford2002) because such an approach facilitates clarity for those reading the opinion and applying its reasoning to subsequent cases and can increase the likelihood of compliance (Owens and Wedeking Reference Owens and Wedeking2011; Owens, Wedeking, and Wohlfarth Reference Owens, Wedeking and Wohlfarth2013). Second, judges may include authoritative language when citing sources like national forensic reports to signal their own credibility or the credibility of an argument or to demonstrate confidence in understanding the scientific concepts needed to make well-informed decisions (Hazleton and Hinkle Reference Hazelton, Rachael and Hinkle2022). Third, the judges may pair authoritative language with the reference to a secondary source—such as a national forensic report—because they are trying to persuade their audience (including other judges) that their opinion is correct, reliable, and sound (Baum Reference Baum2008, Reference Baum2009, Black et al. Reference Black, Owens, Wedeking and Wohlfarth2016b). While our data do not permit us to directly assess judges’ reasons for citing national forensic reports, we can investigate whether judges’ references to these reports are consistent with the potential motivations outlined above.
Importantly, we expect that the use of authoritative language will differ when the reference to a national forensic report benefits the defense. The preference embedded in the U.S. legal system to avoid falsely convicting an innocent person suggests that judges may approach writing differently when it supports the prosecution compared to the defense. The criminal justice system is designed to avoid both Type 1 (a false positive, or convicting an innocent person) and Type 2 (a false negative, or a not-guilty verdict when the defendant is guilty) errors. However, legal scholars have discussed at length the inherent tradeoff between the two types of error, where working to minimize one type typically opens the door to more of the other type of error (e.g., Radford Reference Radford1988; Givati Reference Givati2019). Standards like the presumption of innocence and due process protections (e.g., Radford Reference Radford1988) suggest Blackstone’s (Reference Blackstone1765-1769) frequently referenced adage that “it is better that ten guilty persons escape than that one innocent suffer” is reflected within the criminal legal system. Additionally, the high burden of proof required to find a defendant guilty “beyond a reasonable doubt” further demonstrates that avoiding Type 1 errors is the priority in criminal cases.
We argue that judges recognize this tradeoff and are attentive to the stakes for the criminally accused. In other words, the above structural features underline for judges the cost of a Type 1 error and the need for reassurance regarding the evidence supporting a defendant’s guilt given that the punishment for criminal offenses may involve deprivation of fundamental rights such as property, liberty, and life (e.g., Radford Reference Radford1988; Stith Reference Stith1990; Findley Reference Findley2018). We expect that an attentiveness to and awareness of the risks of Type 1 error may be reflected in a judge’s use of more authoritative language when referencing national forensic reports in a way that benefits a criminal defendant. While judges do not want to facilitate Type 2 error, we expect that the pull to use authoritative language when referencing a national forensic report in a manner that benefits the prosecution will be less. These arguments lead to two hypotheses:
Hypothesis 1: Judges will use more authoritative language in their opinions when referencing forensic reports in favor of criminal defendants.
Hypothesis 2: Judges will use less authoritative language in their opinions when referencing forensic reports in favor of prosecutors.
We focus our examination on the judicial opinion language surrounding citations to national forensic reports in criminal cases and whether the reference to the report benefits one, both, or neither of the parties. This is distinct from the ruling or outcome of a case. In other words, a judge might use a forensic report in a way that benefits the defense while still ultimately ruling against the defendant. We take this approach as a first step toward better understanding how judges reference expert scientific information. This approach paves the way for future studies to assess the role national forensic reports and the expert information they provide might shape rulings and final merits decisions.
Data and Methods
To assess how judges engage with national forensic reports, we collected and analyzed judicial opinions that reference one or more of the following six reports: DNA Technology in Forensic Science (National Research Council, 1992) (NRC I), The Evaluation of Forensic DNA Evidence (National Research Council, 1996) (NRC II), Forensic Analysis: Weighing Bullet Lead Evidence (National Research Council, 2004) (NAS 2004), Ballistic Imaging (National Research Council, 2008) (NAS 2008), Strengthening Forensic Science in the United States: A Path Forward (National Research Council, 2009) (NAS 2009), and Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods (President’s Council of Advisors on Science and Technology, 2016) (PCAST 2016) (Browne Reference Browne2025). We selected these reports because they are widely recognized for elevating national awareness of forensic science issues and shaping how courts evaluate forensic evidence (Shooter Reference Shooter2019). For example, federal judge, Jed S. Rakoff, and California Supreme Court judge, Goodwin Liu, acknowledge NAS 2009 as one of the three major developments in forensic science evidence and its use in courts (Reference Rakoff and Liu2023).
Five of the six reports were published by the National Research Council (NRC) within the National Academy of Sciences (NAS), a private nongovernmental institution established to advise the United States on science and technology issues. The creation of these reports was prompted and/or financially supported by the Federal Bureau of Investigation (FBI), the National Institute of Justice (NIJ), the National Institutes of Health’s National Center for Human Genome Research, the National Science Foundation, and other national entities. Although they received requests or financial backing from the FBI, NIJ, or other entities, the studies and their findings were not swayed by these organizations. Additionally, one of the reports (National Research Council, 2009) was created as a result of a congressional mandate. The five NRC reports were produced after the creation of independent committees, composed of individuals from the legal, forensic science, and other scientific fields. These committees were responsible for studying the general practicality, scientific validity, and reliability of specific techniques.
One of the reports was published by the President’s Council of Advisors on Science and Technology (PCAST), a body of advisors created to provide scientific, technological, and innovation recommendations to the President of the United States. This group is usually composed of leading experts in science, technology, and engineering. In 2015, President Obama requested PCAST to evaluate different forensic science fields and highlight areas requiring improvement. Following consultation with lawyers and judges, PCAST published their findings in the 2016 report.
We searched Westlaw and LexisNexis for all state and federal criminal judicial cases that cited one or more of the six NAS and PCAST reports between 1992 and 2023.Footnote 7 We identified and included all report references using the report name, date, or content of the report.Footnote 8 Report references were collected from majority, dissenting, and concurring opinions released beginning the year the oldest report (NRC I) was published in 1992.
The national forensic reports we examine consider scientific methods for analyzing various forensic evidence, such as firearms and toolmark, fingerprint, DNA, and bitemark evidence. National forensic reports are potentially directly relevant in cases where questions arose about the admissibility of forensic evidence and/or testimony about how the evidence was collected or presented at trial. This universe of cases includes claims that the forensic method was unreliable and lacked general acceptance, as well as constitutional violations under the Fourth, Fifth, Sixth, and Fourteenth Amendments. References also occur in post-conviction cases, where the national forensic reports are introduced as “newly discovered evidence” to support actual innocence claims or to justify defense requests for additional forensic testing, usually DNA testing. There were also citations to the national forensic reports to address or highlight claims of insufficient evidence, presented by a party (usually the defense) that argued the forensic evidence was legally insufficient to prove identity. Often, the cases we identified as having opinions with references to national forensic reports related to violent crimes like murder, rape, sexual assault, assault, and battery. For example, just over half of the cases we examine involved a murder, attempted murder, homicide, or similar charge. However, the universe of criminal cases we examine is not strictly limited to cases where these types of forensic evidence are directly relevant. Infrequently, judges also cited the national forensic reports in cases that discuss forensic evidence not mentioned in any of the reports (e.g., digital, serological, and cadaver dog evidence).
It is important to keep in mind that our findings can only speak to the subset of judicial opinions in criminal cases in which national forensic reports are cited. Excluded from this inquiry are cases that did not involve forensic evidence or cases where the forensic evidence was never questioned. Lastly, cases where other scientific reports or literature were cited to address forensic evidence, but the six national reports were not, are excluded from this dataset.
Judicial Engagement with National Forensic Reports
The searches yielded 621 state and federal criminal cases with opinions citing at least one national forensic report.Footnote 9 In this study, our unit of analysis is a written opinion. Over the period we examine, we report an average of 21.4 cases per year with an opinion citing a national forensic report though the total varied across the years we examined. This variation is at least in part explained by the fact that reports were issued throughout the time period we examined. Excerpts were collected from all instances within an opinion when a report was referenced. When a report was referenced more than once within an opinion, excerpts from all references were gathered and repeated invocations were recorded as a single entry. Each excerpt included the report reference as well as additional information needed to understand the judge’s reasoning. Additionally, when the report reference was linked to a particular argument, the judge’s full discussion and conclusion of that argument was also captured. We note that our approach of using direct citations to national forensic reports likely underestimates the extent to which judges engage with scientific information in their decision-making because we suspect that over time, the information contained in these reports is subsequently utilized by citing past precedent that referenced one or more forensic report(s). However, starting our inquiry with data about the most direct version of judges engaging with these reports allows us to establish an initial pattern of judicial engagement with scientific information.
The first takeaway from the data is that judges are engaging with national forensic reports—through citation and discussion—in a non-trivial number of their written opinions in criminal cases. Any engagement with these national forensic reports is noteworthy given the guidance judges receive about citing secondary sources sparingly. While the number of cases we identified that contain references to national forensic reports is only a small part of the judiciary’s criminal docket in a year,Footnote 10 examining these cases provides analytical leverage toward our goal of understanding how judges engage with forensic reports in their written opinions. Our original dataset suggests judges are attentive to the expert information contained in these national forensic reports and that judges may see these secondary sources as authoritative and useful for making decisions. In our theoretical framework, it also demonstrates that judges have engaged in at least a cursory search for this expert information in some cases.
When this expert information search occurs, what do references to reports in judicial opinions look like? Judges cite national forensic reports as secondary sources in their criminal opinions in a variety of ways. One common way is to explain a forensic method or principle and/or provide background information. Judges may use this type of reference to describe or define a forensic science concept and to showcase the judge’s understanding of the methods. For example, in State v. Anderson, 853 P.2d 135 (N.M. Ct. App. 1993), Justice Baca of the New Mexico Supreme Court included a section of the majority opinion with the heading, “DNA Background.” Justice Baca begins this section with the following:
A basic understanding of the scientific principles and techniques underlying DNA typing is essential in order to understand the legal issues relating to its admissibility. Unless otherwise stated, we derive our scientific explanation of DNA and DNA profiling from testimony given at the Frye hearing, from a report entitled “DNA Technology in Forensic Science,” which the National Research Council published in April 1992, and from Government of Virgin Islands v. Penn, 838 F. Supp. 1054, 1057-73 (D.V.I. 1993) and State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 486 (N.H. 1992).
In the above example, Justice Baca acknowledges in writing that the NRC I is scientific, expert information. He uses the report as the foundation for explaining key principles of forensic evidence alongside other highly respected authorities, such as the record in the case and precedent.
Another example of a judge referencing a report in a written opinion to explain a forensic science concept or principle appears in State v. Quatrevingt, 617 So. 2d 484 (La. Ct. App. 1992) where Judge Klees, writing in the majority opinion, discusses how NRC I provides insight into the views of the scientific community, noting that:
The scientific community is in agreement that, when conducted within accepted procedures and standards, the results of DNA analysis are very reliable. See National Research Council, DNA Technology in Forensic Science (1992); Report of New York State Forensic DNA Analysis Panel (1989).
This example is representative of judges’ recognition of the expertise contained in NRC I and, by extension, their willingness to apply the guidance offered by national forensic reports. Notably, these forensic reports do not always provide clear-cut answers. They may also explain existing biases in the forensic science system, highlighting challenges associated with forensic evidence. Melendez-Diaz v. Massachusetts, 557 U.S. 305 (2009) is one such example. Justice Scalia, writing in the majority opinion, notes that:
Forensic evidence is not uniquely immune from the risk of manipulation. According to a recent study conducted under the auspices of the National Academy of Sciences, “[t]he majority of [laboratories producing forensic evidence] are administered by law enforcement agencies, such as police departments, where the laboratory administrator reports to the head of the agency.” National Research Council of the National Academies, Strengthening Forensic Science in the United States: A Path Forward (2009) (hereinafter National Academy Report). And “[b]ecause forensic scientists often are driven in their work by a need to answer a particular question related to the issues of a particular case, they sometimes face pressure to sacrifice appropriate methodology for the sake of expediency.” Id., at 23–24. A forensic analyst responding to a request from a law enforcement official may feel pressure—or have an incentive—to alter the evidence in a manner favorable to the prosecution.
The above example illustrates that judges recognize the complexity of gathering and using forensic evidence in criminal cases. They may then seek expert information from reports like NAS 2009 in order to provide explanation, highlight the limitations of forensic analysis, and discuss guidance on best practices.
Sometimes national forensic reports play a larger role in a judicial opinion. For example, in People v. Genrich, 471 P.3d 1102 (Colo. Ct. App. 2019), Judge Taubman, writing for the majority, discusses how the defendant’s argument in a post-conviction appeal rests at least in part on a forensic report:
Defendant, James Genrich, appeals the district court’s denial of his Crim. P. 35(c) motion for postconviction relief. He contends that the district court erred in denying him an evidentiary hearing to prove allegations set forth in his motion and incorporated affidavit. In support of his argument, he points to a 2009 report, commissioned by Congress and published by the National Academy of Sciences, Nat’l Research Council of the Nat’l Acads., Strengthening Forensic Science in the United States: A Path Forward (2009), https://perma.cc/8H3Q-S9SU (hereinafter NAS Report), that found toolmark identification evidence—which served as a linchpin in the prosecution’s case against him—had not been scientifically validated. He also alleges that the district court violated his right to due process by admitting such evidence to support his conviction. In addition, he contends that the opinions of a forensic scientist, premised on extensive scholarship, review of the evidence, knowledge of contemporary scientific consensus, and authorship of the NAS Report, constitute newly discovered evidence that undermines confidence in the jury’s verdicts.
In the above example, the defense leverages NAS 2009 to challenge the reliability of evidence used against the defendant. Taubman’s opinion then engages substantively and significantly with this report to evaluate the defendant’s claim and addresses the challenge presented. Combined, these examples illustrate that judges are attentive to these national forensic reports, and we see that judges engage with the forensic reports in a variety of ways in criminal judicial opinions, ranging from simple informational references to instances where reports serve as the basis of one party’s argument.
We make explicit assumptions to connect our dependent variable, authoritative language, to judicial writing as follows. Given the guidance to use secondary sources “sparingly,” we infer that judges’ citation to these reports suggests that the reports may be seen as valuable resources that contain expert information relevant for understanding forensic methods and principles. We also assume that a citation to any secondary source occurs intentionally because the judge found value in engaging with the citation’s source in some way. In our current research design, we cannot distinguish empirically between the different ways judges might reference national forensic reports showcased in the examples above. On the minimal side, a judge could be using the citation to explain a method. At maximum, the judge could be referencing a national forensic report as an important part of how the judge decided the outcome of the case. Readers should therefore be cautious about using the analysis as evidence of judges’ intent.
Dependent Variable
For our dependent variable (the outcome variable), we rely on computer-assisted text analysis software from the most recent update to the Linguistic Inquiry Word Count program called LIWC-22 to measure authoritative language which was developed based on scholarship on social psychology (Boyd et al. Reference Boyd, Collins and Ringhand2023). This software uses a “bag of words” dictionary approach which counts words associated with a concept (here, authoritative language) within the text excerpt we identified as containing a reference to a national forensic report or reports described above. Bag-of-words dictionary approaches consider words separately from their context in a sentence, paragraph, or larger text and count the number of words included in the corresponding dictionary for the concept measured. All words are weighed equally, but only the words that appear in the dictionary figure into LIWC’s score. The score LIWC computes is based on the proportion of words in a text that we provided to the software that are associated with a given category, such as authoritative words. The bag-of-words approach does not incorporate context, syntax, or legal meaning, and can only imperfectly capture rhetorical features of authoritative language. In this study, the degree of authoritative language a judge uses when referencing forensic reports provides insight into how a judge contends with scientific information in judicial decision-making. For example, following Corley and Wedeking (Reference Corley and Wedeking2014) and Stobb and Scalera-Elliott (Reference Stobb and Scalera-Elliott2023), an opinion excerpt written with a high level of authoritative language suggests that the writer may be trying to convince a reader that the information included is correct and reliable. In light of the advice judges are given about using secondary sources sparingly, we argue that authoritative language could be one way in which judges justify their choice to cite a forensic report in an opinion.
We operationalize our dependent variable, authoritative language, by adding together LIWC-22’s “certitude” and “all or nothing” measures. LIWC’s certitude dictionary (certitude) includes words that social psychology scholarship found to indicate a speaker or author may be trying to convince an audience of his or her position. The certitude dictionary includes but is not limited to words like: really, actually, of course, real. LIWC’s all-or-nothing thinking dictionary (allnone) includes words conveying absolutism including but not limited to words like: all, no, never, always, and unquestionably (Boyd et al. Reference Boyd, Collins and Ringhand2023).Footnote 11 Both measures of authoritative language reveal when a text has a higher proportion of words that convey a strong conviction, confidence, and even absolutism that one’s position is correct. Examining the conditions under which judges use more authoritative language when citing national forensic reports helps us understand how this scientific information contributes to judicial decisions.Footnote 12
In our dataset of opinion excerpts with citations to national forensic reports, authoritative language ranges from 0 to 4.05 with a mean of 0.78. In the following example of an excerpt from United States v. Boyd, 686 F. Supp. 2d 382 (S.D.N.Y. 2010) from our data, we indicate which words appeared in the authoritative language measure in bold font.Footnote 13 This excerpt scored a 3.23 in authoritative language based on the LIWC dictionaries indicating just over three percent of the words in the excerpt register as authoritative language. The judge uses the words “extremely,” “must be,” “the one,” and “simply” as they highlight information about DNA evidence found in the NAS 2009 report when addressing the need to cross-examine a DNA expert witness.
Even DNA testing, though providing the “gold standard” for forensic testing, is not immune from error or falsification, see NATIONAL RESEARCH COUNCIL OF THE NATIONAL ACADEMIES, STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES: A PATH FORWARD 130-33 (2009), and, thus, even if the overall error rate in properly conducted DNA testing is extremely low, a defendant must be given, consistent with the mandates of the Constitution, a reasonable opportunity to determine through cross-examination if any such error or falsification is present in any DNA testing admitted into evidence. That opportunity cannot be boundless, however. If, on the one hand, a technician’s role at any stage involves a material exercise of judgment or analysis, then a defendant’s inability to cross-examine the witness on the results obtained from that exercise may rise to an injury of constitutional magnitude. But if, as here, the DNA testing in its preliminary stages requires the technician simply to perform largely mechanical or ministerial tasks, then, absent some reason to believe there was error or falsification, the need to call such a technician as a witness may not become a constitutional necessity.
Since this measurement strategy focuses on word usage, this analysis should not be used on its own to definitively claim judges are relying on or valuing scientific expertise because context can (and does) vary. For example, because this method does not fully capture the context of a reference to a forensic report, a judge’s use of the authoritative language may be associated with a fact-based statement and may not reflect an effort to persuade other judges. For example, words identified by LIWC’s dictionaries can be found in Commonwealth v. Fowler, 2 Mass. L. Rep. 168 (Mass. Ct. Super. 1994):
The first group of scientists has criticized the 1992 NRC report as being scientifically regressive …The second group of scientists has taken the position that the ceiling principle will not always result in a conservative statistical calculation. The third group of scientists includes those who believe that under the proper circumstances, the ceiling principle can provide the corrective conservative statistical calculation…
In this example, the judge is simply explaining a division that existed at the time and is not trying to persuade the audience in any particular direction, nor is the judge clearly relying on the report to make their decision. Additionally, while the reference itself may benefit the defense, the actual statement containing the authoritative language could be referencing the prosecution and vice versa. Despite these limitations, we argue that this well-established text-as-data method (Corley and Wedeking Reference Corley and Wedeking2014; Stobb and Scalera-Elliott Reference Stobb and Scalera-Elliott2023) provides an opportunity to systematically assess the manner in which judges use these reports in their opinions. By examining the extent to which judges deploy authoritative language when they cite and discuss national forensic reports in an opinion in a criminal case, we gain insight into how judges engage with and utilize scientific information in their decision-making. This is especially important to understand since guiding resources for judges, such as the Judicial Writing Manual (Federal Judicial Center 2013), the Judicial Opinion Writing Handbook (George Reference George2007), and The Redbook: A Manual On Legal Style (Garner Reference Garner2023), do not offer much insight into when and how scientific information should be incorporated into judicial decisions.
Independent Variables
Our key independent variables capture whether references to a national forensic report within a criminal judicial opinion benefit the prosecution and/or the defense (or neither). To assess this relationship, we coded whether each report reference within an excerpt benefited the prosecution, the defense, neither, or both. An excerpt was designated as a “1” on Prosecutor Benefit or Defense Benefit if it contained a reference that supported or elaborated on the prosecutor’s or defense’s argument, respectively, and “0” otherwise. An example of this distinction can be found in Carmon v. State, 2022 WL 3593977 (Conn. Ct. Super. 2022):
The State further argues that the firearms evidence offered by the petitioner is not newly discovered because doubts and questions over the validity of firearm and toolmark identification were being considered and discussed prior to the petitioner’s trial. The state relies on a 1959 study and a 1984 study cited in the National Research Council of the National Academy of Sciences, “Strengthening Forensic Science in the United States: A Path Forward” (National Academies Press 2009) (“2009 NAS report”) as support for its position.
In the above example, the state referred to articles that were cited in NAS 2009 to support their argument. In contrast, a defense-benefiting example can be found in Commonwealth v. Lykus, 20 Mass. L. Rep. 598 (Mass. Ct. Super. 2005):
With respect to the bullet analysis, the National Research Council has concluded that there is no scientific basis for the testimony of the FBI agent that it was a “remote” possibility that the bullet found in the victim came from a source other than the box of bullets found in the defendant’s apartment.
In this example, NAS 2004 was used to further challenge the testimony of the prosecutor’s expert witness.
When looking at the two variables Prosecutor Benefit and Defense Benefit descriptively, judges more commonly reference forensic reports in ways that benefit only the defense more (85.2% of references) when compared to solely benefiting the prosecution (11.6% of references). Importantly, a reference may benefit both parties or may not benefit either party. In our dataset, there are 3.1% of cases citing forensic reports in ways that support both the defense and the prosecution. We also account for report references that support neither the defense or the prosecution explicitly with the dummy variable Neutral. Since a judge may reference national forensic reports in multiple ways within the same case, the variables Prosecutor Benefit, Defense Benefit, and Neutral are not mutually exclusive.
A judge may cite one national forensic report or may choose to cite more than one, and it is possible that this decision relates to the opinion language used. We account for this possibility with the variable, Reports Cited Count, which indicates the number of forensic reports cited in a case. This variable ranges from 1 to 4 with a mean of 1.3. Approximately 25% of cases in our data cite two or more forensic reports.
Additionally, Years Since Report accounts for the number of years since the most recent report cited in that opinion was published. It is likely that discussion of national forensic reports in judicial opinions evolves over time. For example, the novel expert information contained in newly released reports might lead judges to cite reports more near the time when they are originally published. This variable helps control for this potential for time-based variation in how judges engage with forensic reports.
We also include several case-level control variables. First, we create a series of dummy variables to control for the type of crime(s) involved. This allows us to account for the possibility that judges use national forensic reports differently in an opinion based on the crime at issue. We include dummy variables for Murder, Rape/Sexual Assault, Burglary/Robbery, Kidnapping, as well as Other Violent and Other Nonviolent crimes.Footnote 14 Second, since our dataset includes judicial opinions from multiple courts, we control for the type and level of court. More specifically, we include several dummy variables capturing whether the report citation occurred within a judge’s opinion for an Appellate Court, a High Court, and/or a Federal Court. This controls for the different approaches taken by judges whose courts play distinct roles within their respective judicial hierarchies. For example, a judge on a court of last resort might use different opinion language compared to a lower court judge who must guard against the possibility of reversal by a higher court.
Results and Discussion
We use ordinary least squares regression to estimate the relationship between a judge referencing a national forensic report in a way that favors a particular party and the judge’s use of authoritative language (our dependent variable) across four models displayed in Table 1. Our unit of analysis is at the opinion level. In all models we use standard errors clustered on the year the opinion was released.
Linear regression predicting authoritative language.
Table 1. Long description
The table presents linear regression results with coefficients and robust standard errors in parentheses. The columns are labeled 1 Neutral, 2 Defense benefit, 3 Prosecutor benefit, and 4 All.
Key independent variables and their coefficients across models 1 through 4 include:
* Neutral: 0.015 in model 1 and 0.387*** in model 4.
* Defense Benefit: 0.167*** in model 2 and 0.466*** in model 4.
* Prosecutor Benefit: -0.118** in model 3 and 0.160 in model 4.
* Reports Cited Count: 0.224***, 0.191***, 0.239***, and 0.088*.
* Murder: 0.226***, 0.190***, 0.225***, and 0.132**.
* Rape Sexual Assault: 0.082, 0.091, 0.106*, and 0.012.
* Burglary/Robbery: 0.087, 0.087, 0.089, and 0.033.
* Kidnapping: -0.061, -0.079, -0.053, and -0.062.
* Other Violent: 0.196**, 0.182*, 0.191*, and 0.135.
* Other Nonviolent: 0.374***, 0.362***, 0.369***, and 0.270***.
* Appellate Court: 0.143*, 0.123, 0.137*, and 0.025.
* High Court: -0.047, -0.055, -0.041, and -0.052.
* Federal Court: 0.149**, 0.120*, 0.145**, and 0.079.
* Years Since Report: 0.011***, 0.012***, 0.012***, and 0.008*.
Summary statistics at the bottom show 618 observations for all models. R-squared values are 0.628, 0.635, 0.629, and 0.647 respectively. Significance levels are indicated by asterisks: * p < 0.10, ** p < 0.05, *** p < 0.01.
Note: Robust standard errors clustered by year of decision in parentheses.
*p < 0.10, **p < 0.05, *** p < 0.01
In Model 1 we consider the extent to which judges write with authoritative language in conjunction with the variable, Neutral, which compares excerpts that benefit neither the prosecution nor the defense with those that benefit one or both parties. In Model 2, we shift to considering Defense Benefit, which compares excerpts that only benefit the defense with all others (i.e., those that are neutral, benefit only the prosecution, and those that benefit both parties). Model 3 considers Prosecution Benefit, which compares excerpts that benefited only the prosecution with all others (i.e., those that are neutral, benefit only the defense, and those that benefit both parties). Finally, Model 4 considers all three variables together.
We find that across model specifications a strong and statistically significant relationship exists between an excerpt with a reference to a national forensic report benefiting the defense and a judge’s use of authoritative language. We see a consistent positive relationship between references benefiting the defense and judges’ use of authoritative language in the excerpt in Model 2. When an excerpt benefits the defense, it is associated with a 0.176 increase in authoritative language. This means that when a judge references a national forensic report in a manner that benefits the defense, this reference tends to be used in conjunction with more language associated with the concept of authoritativeness—at least as measured by LIWC’s approach.
The relationship is less straightforward for excerpts containing references benefiting the prosecution. When Model 3 tests excerpts with references only benefiting the prosecution, we see a strong negative relationship that is statistically significant. When an excerpt benefits the prosecution, it is associated with a 0.118 decrease in authoritative language. This negative relationship indicates judges are using less language associated with authoritativeness when the reference benefits the prosecution. The prosecution still does not seem to be attracting authoritative language from the judges—even though we model separately for excerpts benefiting the prosecution (Model 3). We believe this interesting finding may be related to the concerns judges have about Type 1 and Type 2 errors and because the criminal justice system is set up such that the burden of proof is on the prosecutor to prove guilt rather than on the defendant to prove innocence. When we consider the excerpts benefiting either party as well as neutral categories together in Model 4, we see the strong negative relationship between invocations benefiting the prosecution and authoritative language disappear entirely.
Combined, the results suggest that when judges author opinions and cite expert information from forensic science as a secondary source, the language they employ is more authoritative when the reference benefits the defense. Thus we find consistent evidence supporting our first hypothesis, but mixed for our second. More specifically, we do not see a consistent finding that judges tend to use less authoritative language when citing forensic science reports in a way that benefits the prosecution (Model 4). We submit that the finding regarding references that benefit the defense likely reflects the extreme caution baked into the legal system to prevent falsely convicting an innocent person of a crime (e.g., Radford Reference Radford1988).
These findings provide important context for both prosecutors and defense attorneys as they contemplate their legal strategies. In our dataset, references to reports are being used authoritatively in ways that benefit the defense. These references within judicial opinions indicate that judges are engaging with the expert information contained in national forensic reports, suggesting that prosecutors may be underutilizing the reports. Importantly, this does not mean that national forensic reports are biased against prosecutors, but rather could be related to the fact that there are overall more references that benefit the defense. This pattern likely stems from the public defense’s relative lack of resources. Public defense attorneys may be more likely to rely on scientific literature due to their relative lack of resources. Prosecutors benefit from having more resources, as they can afford to pay expert witnesses to testify directly to the court and they have closer relationships with state-sponsored forensic labs (Wright Reference Wright2004). Additionally, as prosecutors’ practice is more localized than generalized criminal defense attorneys, prosecutors are building their status as repeat players with the same judges more swiftly than defense attorneys (Bellin Reference Bellin2019). Therefore, prosecutors are less dependent on scientific literature than defense attorneys. As a result, judges may be more likely to review or use forensic reports through the lens of the defense or to evaluate the defense’s argument. Even though prosecutors have more systemic access to scientific information in other forms, this should not lead them to ignore that judges appear to value these reports and the expert information they contain.
While we did not have expectations for a relationship between excerpts containing neutral references—those that do not benefit the prosecution or the defense—we note that in Model 4 we observe that excerpts with neutral references are positively and statistically significantly associated with the use of authoritative language. This may be because the neutral excerpts tend to be more technical and objective explanations, which can be used authoritatively. The relatively small number of neutral excerpts also could contribute to the relationship between neutral excerpts and authoritative language in Model 1. We also suspect that this relationship is weak because a neutral reference does not serve the same purpose as justifying why a judge might think the evidence supports one party over the other.
Additionally, we note that several control variables shape a judge’s use of authoritative language when referencing a national forensic report. First, we find some evidence that as the number of forensic reports cited by a given judge in a given opinion increases, so too does the use of authoritative language. This suggests that judges express more conviction when leveraging more than one forensic report. Further, we see some indication that federal court judges are more likely to deploy authoritative language when referencing a national forensic report. We find this relationship to be an interesting puzzle. Future research could consider the differences between the writing patterns of federal and state judges when citing expert scientific information.
The evidence suggests that type of crime is at least somewhat related to the use of authoritative language, as well. For example, we find that judges use more authoritative language in opinion excerpts from cases involving murder. This supports the notion that judges are attentive to the stakes of a given criminal case and the potential costs associated with error for the defendant as well as for the victim’s loved ones.
Conclusion
When judges write an opinion to explain a decision in a criminal law case, they may be inclined to cite scientific information when forensic evidence is being used to connect the accused to a criminal act or refute the defendant’s involvement. Since judges are not typically formally trained in forensic science, they may choose to seek out expert-produced information—such as a national forensic report—that has been vetted by the scientific community. Such scientific information could help a judge correctly interpret and explain forensic evidence in a case.
Our data demonstrate that judges engage with national forensic reports by formally citing them in their written opinions in criminal cases. Since the typical advice on judicial opinion writing privileges case citations over secondary sources, the affirmative decision by a judge to use a secondary source like a national forensic report in written opinions is noteworthy. Our analysis examines the extent to which judges use authoritative language when citing forensic reports in judicial opinions, testing whether this propensity relates to whether the reference(s) benefits the prosecution or the defense. We find a positive relationship between an opinion author’s use of authoritative language and references to a national forensic report benefiting the defense. We do not observe a clear pattern for judges’ use of authoritative language when referencing a national report in a way that benefits the prosecution.
While we cannot directly assess judges’ reasons for citing national forensic reports, our findings are consistent with the notion that judges view national forensic reports as important and informative expert sources that are relevant to resolving the legal issue at hand. This possibility is important for at least two reasons. First, criminal cases carry serious consequences for the accused, not limited to fines and imprisonment—the death penalty is on the table in many jurisdictions. If the potential consequences for the accused are so high, it is imperative that judges take seriously the scientific community’s process and expertise for confirming a person’s guilt beyond a reasonable doubt. Higher levels of certainty in the evidence and in how judges craft the opinion may also deliver some level of certainty and closure for the victims of these crimes and their loved ones. Second, there is currently a buzz of criticism regarding bias in the legal system and that judges are not a) trying to and b) doing a good job of taking seriously the scientific community and their expertise. We argue that the evidence presented here—that judges cite directly to expert-produced national forensic reports—lends support to the idea that judges can and do incorporate scientific information into their decision-making and may encourage judges to take into account the expertise of other scientific communities when appropriate.
While this article breaks new ground by examining references to national forensic reports in judicial opinions, it is constrained by data limitations and raises many unanswered questions. For example, we only analyze the subset of judicial opinions in criminal cases in which national forensic reports are cited. Additionally, it is unclear whether and how citations to national forensic reports shape outcomes in criminal cases, how judges interpret the science undergirding forensic reports, and what exactly motivates judges to reference forensic reports and the conditions under which these motivations apply. Beyond answering these important questions, there are several additional avenues ripe for further inquiry.
First, future work could build on this analysis by exploring other aspects of the language judges use when referencing forensic reports in their opinions (e.g., emotional content, tone, etc.), how the language used compares to the opinion as a whole, as well as examine alternative facets relating to these citations outside of linguistic analysis.
Second, we group judicial opinions of all court types and levels together and include controls for this variation. This decision is unconventional, as it would be ideal to separately examine the opinions of different types of courts to assess the extent to which federal judges, for example, interact with national forensic reports compared to state judges and similarly judges on courts of last resort compared to intermediate appellate courts. The current number of report citations and data structure precludes this approach, but future research could analyze judges’ engagement with forensic reports from a different vantage point that would facilitate such comparisons.
Third, we do not distinguish between citations to the six different national forensic reports. Future work could delve more deeply into judicial citation to certain types of forensic reports, such as those considering DNA analysis. A more focused approach would provide more detailed insights into how judges interact with expert information in different topic areas.
Fourth, we do not currently have information about how the litigants are presenting forensic evidence to judges. Further inquiry into how litigants choose to present scientific information in pre-trial documents, written briefs, and/or oral arguments would present another fascinating angle on how forensic science is used in the legal system.
Finally, we also urge future scholars to consider the impact of the resource disparity between criminal defense attorneys and prosecutors on their use of national forensic reports. For example, future research could examine briefs to assess the extent to which defense attorneys are more likely to rely on a general forensic report given their relative lack of resources compared to the prosecution. As discussed earlier in this article, prosecutors have access to state-sponsored forensic labs. For defendants, procuring similar access to forensic analysis and evidence is often prohibitively costly.
Our approach can be generalized to other areas of law by examining a wider variety of secondary sources, but for now, we focused exclusively on the subset of criminal cases where forensic evidence is relevant. The analysis above is a crucial first step that sheds light on the mechanism judges use to explain and justify their decisions both to the parties and the public—written opinions—and how judges use expert forensic science information to do so. Our findings showing the use of authoritative language when referencing national forensic reports in judicial opinions and the variation in language when referencing citations that benefit the defense further suggest that there is more to learn about how judges engage with and utilize forensic reports in decision-making.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/jlc.2026.10028.
Author Contribution
All authors contributed equitably to this project.
Funding statement
No funding was receoved to conduct this research.
Competing interests
The authors report no competing interests in connection to this research.
Data availability statement
Replication data are available in the Journal of Law and Court’s Harvard Dataverse at https://doi.org/10.7910/DVN/JMARDN.
Ethical standards
No ethical approvals are required for this study.
Open access
