George A. Mocsary
In their groundbreaking and provocative article, From Theory to Doctrine: An Empirical Analysis of the Right to Keep and Bear Arms After Heller,Professors Eric Ruben and Joseph Blocher seek to advance Second Amendment scholarship in two important ways. First, they aim to describe Second Amendment doctrine by quantifying aspects of that constitutional provision’s common law. Second, they challenge the view, which I have unabashedly asserted, that lower courts have deliberately and systematically undercut both the right to bear and the right to keep arms as articulated by the U.S. Supreme Court in District of Columbia v. Heller and officially declared fundamental and applicable to the states in McDonald v. City of Chicago.
Professors Ruben and Blocher present a “distant reading”of post-Heller Second Amendment doctrine: An examination of Heller’s impact—and the particular proposition that courts are defying that landmark decision —based on scientific analysis of aggregate hard data on judicial rulings, rather than a parsing of the language of individual opinions. Scholars and practitioners alike would do well to pay close attention to the authors’ descriptive findings. They perform a valuable service by both confirming conventional wisdom and upending common misapprehensions.
I remain unconvinced, however, by the authors’ secondary,albeit generally tentative, implication that courts are not underenforcing Heller. My reluctance stems primarily from the authors’ decision not to measure the final outcomes of Second Amendment challenges. It is heightened by the limitations of empirical studies, the knowledge that judges with even a modicum of self-control are capable of masking their hostility to arms rights, and my desire for information to complement that which is reported in the study. Professor Ruben and Professor Blocher’s study also presents evidence of judicial defiance that I would like to investigate further.
This Comment discusses these intertwined items. Part I discusses ways in which Professors Ruben and Blocher have provided value to the world of Second Amendment study. Part II discusses some features of empirical study and, in proposing that the evidence of judicial underenforcement may be stronger than the authors suggest, offers items worthy of further research.
Professors Ruben and Blocher thoroughly catalog every available federal and state-appellate Second Amendment challenge from Heller’s announcement through February 1, 2016.They present a range of findings that focuses on success rates, but which includes a wealth of descriptive information about the methods employed by courts to perform their analyses.
The authors analyze individual challenges rather than opinions, cases, or final outcomes.Their sample includes interlocutory outcomes, like remands on a given issue, even if that issue returned to the remanding court for adjudication. Professors Ruben and Blocher exclude outcomes found in noncontrolling opinions like dissents, magistrate recommendations, and opinions vacated by its issuing court, and outcomes found in cases “in which the Second Amendment analysis was merely incidental to other legal issues.” They count as a “success” an outcome at any stage that does not reject entirely a plaintiff’s right-asserting challenge.
The Professors’ findings both confirm some of my preconceptions and surprise me. They confirm, for example, that the failure rate of Second Amendment challenges is so high in part because so many of them are “objectively weak claims,” like those brought by felons despite Heller’s sanction of felon disarmament.They also confirm that may-issue-carry licensing regimes and bans on so-called “assault weapons” are nearly always upheld, while total bans on public handgun carriage have universally fallen.
Interestingly, tried federal civil actions that involve the Second Amendment appear to be appealed at a lower-than-average rate than both the universe of federal trials and other civil rights cases.The article provides a wealth of other cuts of the authors’ data, including success rates over time, the doctrinal tests employed by courts, and a Part discussing characteristics of successful challenges.
The authors’ findings also buck some of the conventional wisdom. They report, for example, that challenges subject to ever-malleable intermediate scrutiny succeed at a higher rate than the data set as a whole. This is somewhat contrary to the perception that courts employ that test to dispense with claims to which they are hostile;this should be a self-contained figure mostly unaffected by which opinions are counted as successes, making it a relatively reliable measure.
Other data, although less informative in the absence of additional data on final outcomes, are valuable and intriguing. For example, contrary to the locus communis, the Two-Part Test adopted by most courts was employed in a mere minority of challenges in their sample.This may be partially the result of some interlocutory decisions not lending themselves to analysis under this framework, but it seems doubtful that omitting these opinions would reveal that a supermajority of challenges were subject to the Test.
The authors report that the federal circuits considered most hostile to arms rights, the Second, Fourth, and Ninth, are home to both more challenges to arms restrictions and higher success rates for challengers.They postulate—correctly, I believe—that these findings result, at least in part, from these circuits being home to the nation’s most restrictive arms regulations, and thus the best litigation targets. Litigants have more to gain from defeating harsher restrictions, and should, in theory, have greater chances of success as the restrictions become more likely to overstep Heller.
Relatedly, the rate of successful Second Amendment challenges has increased steadily. All but one full year of the study period has seen an increase, starting at 4 percent in 2009 and concluding at 15 percent in 2015.
These success-rate data would be unusual in an environment characterized by judicial hostility to the Second Amendment. And doubly so if earlier judicial defiance, coupled with Supreme Court complacency, should be expected to embolden later hostile judges to underprotect arms rights.It may mean, for instance, that some judges perceived to be hostile to the Second Amendment simply make a more memorable impression upon readers of their opinions, a phenomenon exacerbated by the heated legal and public discourse surrounding the issue. But I believe that this conclusion is premature, for reasons set forth in Part II.
Quantitative analysis limits, but does not eliminate, the “subjectivity inherent in qualitative analysis.”It may “provide a better basis for broad Second Amendment claims than intuition or cherry-picked case law.” But because empirical studies cannot read minds or compare what is to a counterfactual reality, few claims can truly be proven or falsified.
In the context of illuminating judicial treatment of Heller, Professor Ruben and Professor Blocher’s study is—unavoidably—itself a proxy (for perfect knowledge about judges’ thought processes) that contains proxies (for example, counting whether judges spent three or fewer paragraphs on their Second Amendment analyses to gauge whether they are seriously considering such claims).Proxies are by nature imperfect. This Part sets forth why I find that the authors’ proxies are insufficient grounds on which to reject the hypothesis that lower courts are underenforcing arms rights.
A. Personal Probabilities Matter
The Bayesian approach to statistics shows that intuition can substantially enhance the predictive power of a quantitative measurement. Two types of probabilities are relevant to the instant question: The “empirical probability” of an outcome is determined by observing an event’s frequency; it is based on objectively verifiable results of experiments, but is subject to random chance and confounding factors. A “personal,” or “subjective” or “Bayesian,” probability is based on “prior knowledge,” which includes one’s intuition, personal qualitative knowledge of circumstances, and experience; it is subject to one’s biases and the limits of past experience. Objective knowledge, like that obtained in an empirical test, can be combined with prior knowledge external to a study to arrive at a more accurate conclusion.
Two examples illustrate the point.One might test whether a coin is fair by flipping it 100 times. If it comes up fifty heads and fifty tails, one might be satisfied that it is fair, but if comes up 100 heads and no tails, one might reject the notion that it is fair. These are reasonable assumptions, even though a fair coin could conceivably come up 100 heads in 100 flips, and a biased coin could similarly come up exactly fifty heads and fifty tails. For a result between these two extremes, the decision maker must either pick a threshold number of heads above which he or she will reject the idea that the coin is fair, or simply refrain from deciding. One might also observe that another flipping of a coin 100 times came up with fifty-five heads and forty-five tails. An ordinary onlooker might believe that this is close enough to an even number of outcomes to infer that the coin is fair. Now imagine that the person flipping the coin (1) has placed a bet with another onlooker that more heads would come up than tails and (2) has previously been thrown out of a number of casinos for cheating. An onlooker with this prior knowledge would have a different assessment of the coin’s fairness than an onlooker without that knowledge, and with good reason.
Now project the example to judicial enforcement of the Second Amendment. One might start with the assumption that Justice John Paul Stevens performed an intellectually honest analysis of the Second Amendment’s history to determine, as he wrote in dissent, that it does not protect an individual right to keep a firearm independently of “certain military purposes.”One might begin to doubt his sincerity upon reading an opinion piece in which he writes that pro–gun-control demonstrators should “demand a repeal of the Second Amendment” because it is “a relic of the 18th century.” Similarly, one might believe, based on the success rates measured by Professors Ruben and Blocher, and in the absence of other information, that federal appellate judges are not underenforcing Heller. One might legitimately question that belief upon reading articles by two appellate judges decrying Heller, or passages in judicial opinions that evince open hostility to gun rights, one of which was written by a judge who wrote that “there is no recognized duty of candor in judicial opinion writing.”
An empirical probability can, in other words, legitimately be different from a final modified probability. Normally sufficient statistical evidence—here, of high success rates—can become insufficient when paired with information that forms the basis for a prior estimate. “Rational belief formation . . . is holistic; it is the amalgamation of frequency data with the other (theoretical or observational) grounds of belief that produces experts’, assessors’, and regulators’ Bayesian probabilities.”As judgment is eliminated from an analysis, the reasoning process is truncated by removing relevant facts from consideration. Stated differently, potential judicial underenforcement is an attitudinalist enterprise (by judges) that purely quantitative study can only partially measure.
The remainder of this Part applies and supplements the material discussed thus far. Section B discusses why I believe that reporting on final outcomes, in addition to the important findings presented by Professors Ruben and Blocher, would shed meaningful light on judicial treatment of the Second Amendment. Section C suggests additional areas of measurement to complement to the authors’ data in answering the defiance question.
B. Final Outcomes Matter
Professors Ruben and Blocher include a broad range of nonfinal outcomes in their dataset to better gauge whether courts are “reflexively rejecting” Second Amendment challenges. They acknowledge that this results in more “successes” than counting final rulings would because it includes prevailing opinions in challenges where the Second Amendment claim ultimately failed. A count of nonrejections across courts may be useful for approximating the treatment of a right by, presumably, a larger sample of judges than a count of final outcomes would. And although a lower-court decision that is appealed provides an important measure of that court’s behavior , it can bias the across-court measure of success in either direction by including success or failures that ultimately mean little to litigants and others with a strong interest (one way or another) in the robustness of a right or doctrine. Within courts, it double-counts judges who issue multiple opinions on the same challenge.
I also question whether judges feel a need to reflexively reject a disfavored challenge given that they can do it carefully and systematically, and thus with the thicker veil of legitimacy (and more suspicion-dispelling rulings in favor of plaintiffs) that going through the procedural and analytical motions provides.One way to read, for example, a decision that remands a challenge for fact-finding is as a serious attempt to gather information for later case adjudication. But it may be giving the government a second chance to meet its burden of proof. It may also be the court waiting for the most opportune time to deny the claim while eliminating some legitimate grounds for appeal—the court has no reason to choose an obvious, and perhaps procedurally suspect, time to deny a claim improperly. It also fits with Professor Shay Lavie’s finding that cases with more procedural effort put into them are less likely to be overturned.
Judges are likewise capable of couching their subjective attitudes in ostensibly objective, or, at least, legal, analyses.Given that some decisions denying Second Amendment claims have been reversed, including one summarily vacated by the Supreme Court (albeit after the authors’ study period), savvy lower-court judges wishing to deny Second Amendment claims know that they should not give them cursory treatment. They will be sure legitimately to rule in favor of a law’s challengers as often as possible before spending several paragraphs dishonestly distinguishing precedents and ruling against the claim. It is in judges’ interest to be in secret, or at least unstated, rather than “open,” rebellion against Heller.
The structure of the judicial system also makes final outcomes relevant. A given sample of judges who are assigned Second Amendment cases may not reveal an institutional bias present among judges. Professor Barry Friedman describes “process majoritarianism” as the ways in which an institution’s accountability may be influenced by its structure.A great deal of legislative work, for example, is often done in small compartments within the legislative apparatus and never sees sunlight. But judicial decisions, at least on “matters of serious public moment, . . . are made by the judiciary as a whole” as cases “bubble up through the . . . system.” A system that allows repeated appellate review, including sua sponte en banc review, provides a majority of an appellate court’s judges significant ability to hear cases until that majority’s desired outcome is reached. The potential for nonrandom case assignment further enables outcome manipulation. Final decisions are the public manifestations of the judiciary’s inner workings. They reveal information on “case outcomes and prevailing doctrine” that an aggregation including only nonfinal decisions does not.
It should be possible to extract final outcomes from Professors Ruben and Blocher’s data. The next Section briefly offers additional starting points for future work that could be combined with the study’s findings to paint a clearer picture of potential underenforcement.
C. Measuring “Should”
In evaluating whether the Second Amendment is underenforced, Professors Ruben and Blocher note that the task involves both “a conclusion about how stringently it should be enforced, and an assessment of how it actually is enforced in practice,” and they note that they focus on the latter. Yet the former should question can likewise be split into two components: a comparison to one’s view of what level of enforcement is socially optimal, and another comparison compared to what precedent “requires,” especially as expressed by the Supreme Court, but also by other “binding” decisions. This (like any study) is not a clear objective-subjective dichotomy, but the second component can be done objectively enough to fall within the authors’ goals of remaining as normative-free as possible. A few examples follow.
A relatively objective and quantifiable, but relatively indirect, test of adherence to precedent is a comparison of success rates of Second Amendment claims to those asserting other rights under a given form of heightened scrutiny. Strict and intermediate scrutiny, after all, are judicially created tests that courts typically purport to apply in unaltered form.An imperfect, but likely indicative, comparison of Professor Ruben and Blocher’s findings to similar studies suggests that Second Amendment claims are subjected to a substantially weakened form of heightened scrutiny. They find that plaintiffs’ success rates are 19 and 10 percent for strict and intermediate scrutiny, with an overall rate of 11 percent for heightened scrutiny. These are generally lower than the success rates found in other studies and are in accord with the intuition of the market for information on judicial treatment of the Second Amendment.
Professor Adam Winkler, in his groundbreaking study of strict scrutiny, found that 70 percent of all strict scrutiny cases, and 69 percent of First Amendment cases (to which courts frequently analogize the Second Amendment) succeed.Relatedly, Professor Winkler finds that overall success rates have increased from 62 percent in 1990 to 82 percent in 2003, suggesting that strict scrutiny, although “survivable in practice,” has become “more fatal” over time. Jennifer L. Greenblatt finds an overall success rate of 73 percent for heightened scrutiny claims generally, and 88- and 74-percent success rates in strict and intermediate scrutiny cases. A recent study of free-exercise cases modeled on Professor Winkler’s study finds a 52-percent strict-scrutiny success rate.
A more direct, but more subjective and less easily quantifiable, test of adherence to precedent is a comparison of an opinion’s reasoning on its specific point to what binding precedent says on that point. The potential for subjectivity is obvious: How, for instance, does one define what precedent says, what the examined case reasons, and what the “point” at issue is? Yet even in this realm, some cases should be easy. It is not difficult, for example, to discern that a court is not following precedent when, with Heller explicitly stating that “the Second Amendment ‘extends . . . to . . . arms . . . that were not in existence at the time of the founding,’”that court holds that certain weapons “are not protected because they ‘were not in common use at the time of the Second Amendment’s enactment.’” Harder cases, like those deciding the constitutionality of a may-issue concealed-carry regime in a jurisdiction banning open carry, would require more nuanced analyses on which honest scholars could differ. But even here cases could be checked in various ways for whether they meaningfully address the arguments on either side of a question. Related tests, somewhere near the middle of the objective-subjective and direct-indirect spectra, would be to code opinions for whether they apply the standards that they purport to, checking, for example, whether the government (in usual cases) and plaintiff (for “presumptively lawful” regulations) are held to their burdens. Such work would afford scholars on different sides of this issue an opportunity to work closely together on mutual projects, which would have its own benefits for the area of study.
Two of Professors Ruben and Blocher’s findings suggest that courts are not enforcing Heller as they should. First, at least seven percent of challenges to which courts applied levels-of-scrutiny analysis were decided under rational-basis, or the “equally deferential” reasonableness, review.This is contrary to Heller’s instruction that Second Amendment regulations, like others touching enumerated constitutional rights, are not to be subjected to rational-basis review. Second, the authors note that decisions on a large majority of challenges involved neither original historical analysis nor use of judicial precedent. This suggests that courts are neither following Heller’s methodology nor following previous cases that have.
Professors Ruben and Blocher have gathered a wealth of data on judicial treatment of the Second Amendment. They have both challenged and reaffirmed my prior beliefs. Although their study has not convinced me that courts are properly enforcing the Second Amendment, it provides meaningful data that informs, and invites further examination of, the issue. I thank the authors for inviting this Comment, and I look forward to working with them and their data to further examine this distinctly “diverse, nuanced, and interesting”
Copyright© 2018 George A. Mocsary.
† Associate Professor, Southern Illinois University School of Law. Fordham University School of Law, J.D., summa cum laude, 2009; University of Rochester Simon School of Business, M.B.A., 1997. I would like to thank John Anderson, William Baude, Nicholas J. Johnson, and Jeremy Kidd for their valuable insights and feedback. I am grateful to Mark Collins, Travis Hillenbrand, Jessica Hostert, and Skylar Petitt for their research.
“Trifles, light as air, are, to the jealous, confirmations strong.” 6 William Shakespeare, The Dramatic Works of Shakespeare 487 (Joseph Rann ed., 1791). I am acutely aware that we all suffer from confirmation bias. I thank Professors Ruben and Blocher for inviting me to comment on their work. I hope that, one day, they will oblige me with their favor of the same.
 Eric Ruben & Joseph Blocher, From Theory to Doctrine: An Empirical Analysis of the Right to Keep and Bear Arms After Heller, 67 Duke L.J. 1433 (2018).
 District of Columbia v. Heller, 554 U.S. 570 (2008); McDonald v. City of Chicago, 561 U.S. 742 (2012); Nicholas J. Johnson, David B. Kopel, George A. Mocsary & Michael P. O’Shea, Firearms Law and the Second Amendment: Regulation, Rights, and Policy 1010–11 (2d ed. 2017); Robert J. Cottrol & George A. Mocsary, Guns, Bird Feathers, and Overcriminalization: Why Courts Should Take the Second Amendment Seriously, 14 Geo. J.L. & Pub. Pol’y 17, 30–33 (2016); George A. Mocsary, Defying the Supreme Court in Kolbe v. Hogan, Libr. L. & Liberty (Dec. 20, 2017), http://www.libertylawsite.org/2017/12/20/defying-the-supreme-court-in-kolbe-v-hogan. [https://perma.cc/DU78-59XZ] I have also criticized the “U.S. Supreme Court’s delegitimizing refusal to review blatant affronts to its authority,” id., and shared these views in several addresses before diverse audiences, e.g., Heller and Public Carry Restrictions, 40 Campbell L. Rev. 431, 441–46 (2018) (presentation transcript).
 “Distant reading” is a relatively new form of literary criticism that seeks to understand literature by aggregating and analyzing data from, rather than by reading and interpreting, texts. See Franco Moretti, Distant Reading (2013); Kathryn Schulz, What Is Distant Reading?, N.Y. Times Book Rev., June 26, 2011, at 14.
 See Ruben & Blocher, supra note 1, at 1443, 1447–51, 1475, 1489, 1497, 1507.
 See infra Part I.
 Id. at 1438 (“Our goals are primarily descriptive and analytic.”).
 Id. at 1487.
 See id. at 1475, 1489, 1497, 1507. The authors are careful, as is appropriate, not to make unwavering conclusions on this point.
 See infra notes 15–18 and accompanying text (citing the authors’ study).
 But cf. Kolbe v. Hogan, 849 F.3d 114, 150 (4th Cir. 2017) (en banc) (referring to ownership of so-called “assault weapons” as “conduct that has visited . . . communal bereavement across America”); Kolbe v. Hogan, 813 F.3d 160, 198–99 (4th Cir. 2016) (King, J., dissenting) (stating his preference for banning so-called “assault weapons” and suggesting that the majority will be “‘responsible for some unspeakably tragic act of mayhem,’” but “find[ing himself] outvoted”); id. at 184 (stating that “inferences of this nature have no place in judicial opinions”); Heller v. District of Columbia, 801 F.3d 264, 281 (2015) (LeCraft Henderson, J., concurring in part and dissenting in part) (expressing hostility to both Heller and its “new . . . definition of what the Second Amendment protects”); John Paul Stevens, Opinion, John Paul Stevens: Repeal the Second Amendment, N.Y. Times, Mar. 27, 2018, at A23.
 See infra Part II.
 As opposed to “proof.” A great deal of information can evidence a claim, but relatively little, or none, can definitively prove it. See George A. Mocsary, Statistically Insignificant Deaths: Disclosing Drug Harms to Investors (and Patients) Under SEC Rule 10b-5, 82 Geo. Wash. L. Rev. 111, 113 n.7, 138–56 (2013).
 See infra Section II.C.
 Ruben & Blocher, supra note 1, at 1437, 1456, 1458.
 Id. at 1455–56. One case may challenge “more than one law or policy,” and there may be multiple opinions in each case. Id. at 1455; accord id. at 1496 n.249. There may be multiple prevailing opinions for a given challenge that ultimately fails. Id. at 1462–63.
 Id. at 1462–63.
 Id. at 1460–61.
 Id. at 1462–63, 1466 & n.159. For example, a ruling that dismisses some, but not all, defendants from a claim is counted as a success. Id. at 1466 & n.159.
 Id. at 1447, 1478.
 Id. at 1482–83 (“assault weapons”), 1484–85 (public carriage).
 Id. at 1473 & n.185 (reporting an appeal rate of “as many as 28 percent of federal trial court civil decisions”); Theodore Eisenberg, Appeal Rates and Outcomes in Tried and Nontried Cases: Further Exploration of Anti-Plaintiff Appellate Outcomes, 1 J. Empirical Legal Stud. 659, 659, 664, 666, 672–73 (2004) (noting that the article reports the results of civil appeals; reporting an overall appeal rate of 39.6 percent for all trials, 40.9 percent for all trials with judgments for plaintiff or defendant when using raw data; 28.2 percent and 29.6 percent when employing stricter methods to match trial and appellate data; and an appeal rate of “about 40 percent” for civil rights cases, presumably based on the study’s raw, rather than more-strictly-matched, data); Kevin M. Clermont & Theodore Eisenberg, Plaintiphobia in the Appellate Courts: Civil Rights Really Do Differ from Negotiable Instruments, 2002 U. Ill. L. Rev. 947, 952 (2003) (reporting an overall appeal rate of 28.73 percent).
The authors reach the opposite conclusion, comparing Professor Theodore Eisenberg’s finding of a 19 percent appeal rate of nontried civil cases resulting in judgment with their finding that “28 percent of federal trial court civil decisions involving the Second Amendment are appealed.” Ruben & Blocher, supra note 1, at 1472–73 (citing Eisenberg, supra at 660); see also Eisenberg, supra at 664. If I understand the authors’ statement, their data measures, or, at least, more closely resembles, the appeal rates in tried civil cases. Ruben & Blocher, supra note 1 at 1473 (referring to “decisions” and “opinion[s]”).
 See Ruben & Blocher, supra note 1, at 1471–1506; see also id. App. C, https://dlj.law.duke.edu/wp-content/uploads/sites/2/2018/04/Ruben-and-Blocher-App-C-1.pdf [https://perma.cc/9PQG-7WY6].
 Ruben & Blocher, supra note 1, at 1495–96 (reporting a success rate of 10 percent for intermediate scrutiny versus 9 percent for all claims, and properly acknowledging that the difference may be slightly greater if the former successes are subtracted from the latter). But see infra notes 72–77 (describing studies reporting higher rates of success in heightened scrutiny challenges in other doctrinal areas).
 This statement is based on my untested assumption that nonfinal, interlocutory opinions, supra notes 16, 18 and accompanying text, are unlikely to apply levels-of-scrutiny analysis. It would be valuable to know what the success rate is for final determinations applying intermediate scrutiny. See infra notes 49–53, 73 and accompanying text. It should be possible to determine this from the data collected by the authors. Accord infra text following note 66.
 Ruben & Blocher, supra note 1, at 1490 (reporting a 41-percent uptake); see Nicholas J. Johnson, David B. Kopel, George A. Mocsary & Michael P. O’Shea, Firearms Law and the Second Amendment: Regulation, Rights, and Policy 903–09 (2d ed. 2017). “At Step One, the government has the burden of proving that the activity or thing at issue is outside the scope of the Second Amendment, as traditionally understood. If the government fails to meet its burden on Step One, then the case proceeds to Step Two, which by definition involves some form of heightened scrutiny. Under all forms of heightened scrutiny, the government always bears the burden of proof.” Id. at 907 (citations omitted).
 A motion to dismiss on standing grounds, for example.
 Ruben & Blocher, supra note 1, at 1474–75.
 Id. at 1475–76.
 Id. at 1486.
 Cf. Richard H. McAdams, The Expressive Powers of Law: Theories and Limits 22–135 (2015) (discussing the power of law to serve as a focal point to future actors); A “focal point” is a result toward which decision makers naturally drift. See Thomas C. Schelling, The Strategy of Conflict 57–63 (1980).
 See, e.g., supra note 10.
 See Ruben & Blocher, supra note 1, at 1445–53 (citing sources highlighting this heated rhetoric); Amos Tversky & Daniel Kahneman, Availability: A Heuristic for Judging Frequency and Probability, 5 Cognitive Psychol. 207, 230 (1973) (describing the well-known “availability heuristic,” under which decision makers tend to assign heightened probabilities to the occurrence of events that they have heard about more frequently and recently).
 Ruben & Blocher, supra note 1, at 1445.
 Id. at 1451.
 See Mocsary, supra note 12, at 138–56.
 Ruben & Blocher, supra note 1, at 1451, 1459, 1489.
 Following the mind-reading and three-paragraph examples, I believe that judges are generally capable of underenforcing a right while writing opinions that mask their motives, and it is very likely that a case dealing with multiple Second Amendment issues will spend more paragraphs on Second Amendment analysis. See supra text accompanying note 36; infra notes 54–60.
 Mocsary, supra note 12, at 139.
 Id. at 140. A third type of probability, “theoretical,” or “true” probability, is “based on mathematical models that, if the inputs are correct, are always right.” Id. at 139. Technically, a prior probability can also be a precise theoretical, or even empirical, probability. See Charles Yablon, The Meaning of Probability Judgments: An Essay on the Use and Misuse of Behavioral Economics, 2004 U. ILL. L. REV. 899, 913, 927–28 & n.167. In practice, they are subjective because studies are done in the first instance because theoretical or empirical data are unavailable.
 Mocsary, supra note 12, at 153. Subjective probabilities are updated as more potentially relevant information becomes available. Id.
 The first example is taken nearly verbatim from my article, Mocsary, supra note 12, at 139, 140. To enhance readability, I have not marked it here as a quote.
 District of Columbia v. Heller, 554 U.S. 570, 636 (2008) (Stevens, J., dissenting).
 Stevens, supra note 10. One might also doubt the sincerity of Justice Stevens’ opinion that history did not support Dick Anthony Heller’s right to keep a revolver in his home for defense given that two related cases cited by the Heller majority and Justice Stevens protect an “unqualified right to keep” such a weapon “for the ordinary purposes to which they are adapted,” including “us[ing] such arms at home or on [one’s] own premises.” Heller, 554 U.S., at 608, 613, 614, 623, 629, 648 n.10, 663 n.30, 678 (citing Aymette v. State, 21 Tenn. (3 Humph.) 154 (1840) and Andrews v. State, 50 Tenn. (3 Heisk.) 165 (1871)); Andrews, 50 Tenn. (3 Heisk.) at 178, 186; Aymette, 21 Tenn. (3 Humph.) at 160. This is a rather nuanced point, however, which would require a lengthier discussion than this Comment allows.
 See Harvie J. Wilkinson, III, Of Guns, Abortions, and the Unraveling Rule of Law, 95 Va. L. Rev. 253 (2009); Richard A. Posner, In Defense of Looseness: The Supreme Court and Gun Control, New Republic, Aug. 27, 2008, at 32.
 See, e.g., supra note 10 (citing passages).
 Richard Posner, Some Realism About Judges: A Reply to Edwards and Livermore, 59 Duke L.J. 1177, 1182 (2010). It is not surprising that many question Judge Posner’s motives in striking Illinois’s ban on public firearm carriage in Moore v. Madigan, 702 F.3d 933 (7th Cir. 2012). See Johnson et al. supra note 2, at 993 (question drafted by J. David Sanders stating that “[s]ome have privately speculated that Judge Posner drafted Moore contrary to his own personal views in order to provoke the U.S. Supreme Court to grant certiorari and revisit the Court’s Heller decision.”). Does Judge Posner’s Moore opinion cut in favor of or against judicial defiance? I leave it to the reader to decide. See also Nat’l Rifle Ass’n of America, Inc. v. City of Chicago, 567 F.3d 856 (7th Cir. 2009) (Judge Posner joining an opinion that did not extend Heller to the states), rev’d sub nom. McDonald v. City of Chicago, 561 U.S. 742 (2012).
 Matthew D. Adler, Against “Individual Risk”: A Sympathetic Critique of Risk Assessment, 153 U. Pa. L. Rev. 1121, 1216 (2005).
 Mocsary, supra note 12, at 156.
 Ruben & Blocher, supra note 1, at 1462; see supra notes 15–18 and accompanying text.
 Ruben & Blocher, supra note 1, at 1462–63.
 But see infra notes 61–66 and accompanying text.
 The authors have reported this data, see Ruben & Blocher App. C, supra note 22, at xvii–xxii, though it includes interlocutory decisions, see supra text following note 16 and accompanying text.
 See Adam Winkler, Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts, 59 Vand. L. Rev. 793, 811 (2006) (measuring final outcomes in testing the robustness of strict scrutiny protection). Professor Winkler potentially excludes some final outcomes, like preliminary injunction decisions that Professors Ruben and Blocher correctly note may result in success for a challenger. Ruben & Blocher, supra note 1, at 1462–63; Winkler, supra, at 811. A case-coding project that measures final outcomes should include such decisions.
 See supra notes 46, 49 and accompanying text. But see supra note 10 and accompanying text (citing an apparent lack of judicial self-control).
 For example, intermediate scrutiny permits, but does not require, remand to allow the government to gather and present more evidence. Compare Turner Broad. Sys. v. FCC, 512 U.S. 622, 664–68 (1994), with Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 721–27, 730 & n.16 (1982).
 Shay Lavie, Are Judges Tied to the Past? Evidence from Jurisdiction Cases, 43 Hofstra L. Rev. 337, 347 (2014).
 To their credit, Professors Ruben and Blocher acknowledge this possibility. See Ruben & Blocher, supra note 1, at 1459 & n.118, 1470, 1487. I believe, however, that they underestimate its importance. For example, although a binding opinion’s text certainly “constitute[s] precedent,” id. at 1459 n.118, it carries less weight if a future judge defies it. Indeed, that is a key point presently under discussion vis-à-vis Heller.
 See Caetano v. Massachusetts, 136 S. Ct. 1027, 1027–28 (2016) (vacating Commonwealth v. Caetano, 26 N.E.3d 688 (Mass. 2016), because it thrice applied reasoning that Heller explicitly rejected). It may have helped that Ms. Caetano was a sympathetic plaintiff.
 See supra note 36 and accompanying text (discussing the authors’ use of the number of paragraphs spent by judges on Second Amendment analyses as a proxy for whether the judges seriously considered such claims); infra Section II.C. Doctrinal questions may have yielded lower intercoder reliability because coders had difficulty untwisting obfuscatory language. Ruben & Blocher, supra note 1, at 1487.
 Ruben & Blocher, supra note 1, at 1497. Less cynically, judges may not perceive their own biases. See supra note † (discussing confirmation bias).
 Barry Friedman, Dialogue and Judicial Review, 91 Mich. L. Rev. 577, 609–13 (1993).
 See id. at 610–11; see also Louis D. Brandeis, Other People’s Money and How the Bankers Use It 92 (1914) (“Sunlight is said to be the best of disinfectants.”).
 Id. at 613.
 Order at 1, Peruta v. Cty. of San Diego, 824 F.3d 919 (9th Cir. 2016) (No. 10-56971). Despite a victory before a Ninth Circuit panel, Peruta v. Cty. of San Diego, 742 F.3d 1144, 1179 (9th Cir. 2014), and the San Diego Sheriff’s decision not to seek further review, Press Release, San Diego Cty. Sheriff’s Dep’t, San Diego Sheriff’s Decision Regarding Ninth Circuit’s Opinion on CCWs (Feb. 21, 2014), the court ordered en banc review and reversed Mr. Peruta’s Second Amendment challenge, Peruta v. Cty. of San Diego, 824 F.3d 919, 925, 942 (9th Cir. 2016) (en banc) cert. denied sub nom. Peruta v. California, 137 S. Ct. 1995 (2017).
 See Marin K. Levy & Adam S. Chilton, Challenging the Randomness of Panel Assignment in the Federal Court of Appeals, 101 Cornell L. Rev. 1, 5, 38 (2015).
 It may be that counting final decisions would reveal that Courts of Appeals are not systemically biased against Second Amendment challenges. See Ruben & Blocher, supra note 1, at 1473 (noting higher overall success rates in federal appellate courts than in federal trial courts).
 Ruben & Blocher, supra note 1, at 1449.
 The quotes are mine. See supra note 2 and accompanying text; see also Richard M. Re, Narrowing Supreme Court Precedent from Below, 104 Geo. L.J. 921, 960 (2016) (describing courts’ narrowing binding precedent, sometimes as a form of “resistance”).
 The authors’ goal is to measure the factual predicates of the underenforcement claim. Ruben & Blocher, supra note 1, at 1453. They correctly acknowledge that “even the most cautious and rigorous coding project will involve difficult judgment calls” and diligently note where they have performed subjective synthesis of their data. E.g., id. at 1459 (inferring the level of scrutiny applied where it is unstated), 1461 (excluding challenges “too tentative to count as . . . Second Amendment holding[s]”), 1462 (discussing “varying degrees of definiteness” in and deciding to include interlocutory challenges), 1466 (“definitional challeng[es],” including “ambiguous” victories), 1467 (“ambiguities and judgment calls are unavoidable”) (emphasis removed), 1478 (excluding figures from Illinois because they skew data), 1486 (“translating subjective questions into objective ones”). These choices are appropriate, both to enhance precision with prior knowledge, see supra Section II.A, and to approach the claim from the angle of most interest to the authors. Later scholars are free to build on their work as, indeed, this invited Comment does. The authors also recognize that normative claims are “unavoidably intertwined” with the analysis. See Ruben & Blocher, supra note 1, at 1438, 1449, 1453.
 But see N.Y. State Rifle & Pistol Ass’n v. Cuomo, 804 F.3d 242, 259 n.89 (2d Cir. 2015) (applying a “mild form of intermediate scrutiny”); Ezell v. City of Chicago, 651 F.3d 684, 708 (7th Cir. 2011) (applying “not quite ‘strict scrutiny’”); but cf. Darrell A.H. Miller, The Second Amendment and Second-Class Rights, Harv. L. Rev. Blog (Mar. 5, 2018), https://blog.harvardlawreview.org/the-second-amendment-and-second-class-rights [https://perma.cc/UFA2-Q3PY] (observing that judges frequently “rank rights,” and implying that the practice is nonproblematic).
 Compare supra notes 15–18 (describing the authors’ counting methods), 23–24 (discussing the likely reliability of those methods with respect to tiers-of-scrutiny analysis) and accompanying text, with infra notes 74, 76, 77 (describing the counting methods used in comparison studies).
 Ruben & Blocher, supra note 1, at 1495.
 The intuition point is important here. See supra Section II.A. Professors Ruben and Blocher are correct when they say that “the state of scholarly art rarely demands more than a ‘See, e.g.,’ signal followed by three case citations.” Id. at 1453. Yet the studies that follow suggest that, as in other areas, the Second Amendment market does a respectable job processing information that trickles into it as opinions are issued. See Mocsary, supra note 12, at 150–56, 158–73; Donald C. Langevoort, The SEC, Retail Investors, and the Institutionalization of the Securities Markets, 95 Va. L. Rev. 1025, 1047, 1043 (2009) (“rational actors weigh all available information in a Bayesian search process” in which they update their views as more information becomes available).
This conclusion is weakened if the “presumptively lawful” regulations identified by Heller, like bans on felon arms possession, District of Columbia v. Heller, 554 U.S. 570, 626 & n.26 (2008), are frequently analyzed under heightened scrutiny. I suspect that they, like issues in interlocutory opinions, see supra note 24, are not. Relatedly, even an overall (under any type of analysis) 13- to 16-percent success rate, arrived at by excluding all “Who” and pro se cases, is low compared to the rates reported by the following studies. Ruben & Blocher, supra note 1, at 1507; id. App. C at xxii, xxiv. That said, there may be a higher proportion of objectively weak claims even after removing these categories. But, at some point, weakness becomes subjective. And so on.
 Johnson et al. supra note 2, at 906; Winkler, supra note 53, at 815. Professor Winkler counts only final outcomes in federal court from 1990 to 2003. Id. at 810–11.
 Id. at 824–25, 871.
 Jennifer L. Greenblatt, Putting the Government to the (Heightened, Intermediate, or Strict) Scrutiny Test: Disparate Application Shows Not all Rights and Powers are Created Equal, 10 Fla. Coastal L. Rev. 421, 444–73, 481–88 (2009). Rights tested under varying forms of scrutiny succeed 68 percent of the time and are excluded from the figures for strict and intermediate scrutiny reported in the text. A spreadsheet showing my calculations is on file with the Duke Law Journal. Ms. Greenblatt counts all Supreme Court cases explicitly applying heightened scrutiny. Id. at 441.
 Caleb C. Wolanek & Heidi Liu, Applying Strict Scrutiny: An Empirical Analysis of Free Exercise Cases, 78 Mont. L. Rev. 275, 295 (2017). The study counts all federal cases from 1990 through 2015. Id. at 291–92.
 Caetano v. Massachusetts, 136 S. Ct. 1027, 1028 (2016) (quoting Heller, 554 U.S. at 582).
 Id. at 1027 (quoting Commonwealth v. Caetano, 26 N.E.3d 688, 693 (Mass. 2016); see also Mocsary, supra note 2, at 31–33 (criticizing Drake v. Filko, 724 F.3d 426 (3d Cir. 2013), for defying both Supreme Court and Third Circuit precedent); see also supra note 58 (discussing Caetano).
 E.g., Peruta v. Cty. of San Diego, 824 F.3d 919, 927–28, 941–42 (9th Cir. 2016) (en banc). I have criticized this decision for ignoring California’s having banned open carry during the pendency of the litigation, and that Heller “strongly implied” that the Second Amendment required the legality of either open or concealed may-issue carry. See Mocsary, supra note 2, at 30, 31.
 This would involve the judgment-intensive task of collecting these arguments. See, e.g., Mocsary, supra note 2, at 29–30 (offering five aspects of Heller suggesting that the Second Amendment protects public carriage); supra note † (discussing confirmation bias). Other parts of this Comment discuss these methods’ strengths and weaknesses.
Murphy v. Guerrero, No. 1:14-CV-00026, 2016 WL 5508998 (D.N. Mar. I. Sept. 28, 2016), is an exemplar of a case applying controlling precedent. Chief Judge Ramona V. Manglona adopts precisely the reasoning and conclusions of on-point Ninth Circuit case law, while providing original analyses and rulings on matters of first impression.
 Heller, 554 U.S. at 626 & n.26.
 Ruben & Blocher, supra note 1, at 1495 (noting also that an unspecified level of review was applied to eight challenges; in no event, therefore, was rational-basis or reasonableness review applied to more than nine percent of challenges); Adam Winkler, Scrutinizing the Second Amendment, 105 Mich. L. Rev. 683, 688 (2007).
 Heller, 554 U.S. at 628 & n.27.
 Ruben & Blocher, supra note 1, at 1491–93 (suggesting also that the latter may be susceptible to undercoding).
 Id. at 1507.