Printer Friendly

Confronting Crawford v. Washington in the lower courts.


In order to assess the potential ambiguity of the Crawford line, I randomly sampled and coded lower-court decisions (188) and quantitatively analyzed the facts and outcomes of those decisions. This Part details the empirical methodology I employed and the results that this investigation produced.

Previous work has evaluated Crawford in the lower courts by closely analyzing select decisions. (189) While this is the standard method of legal reasoning, there are important reasons to prefer systematic empirical analysis of judicial opinions when surveying an area of law. First, empirical methods help to counteract the cherry-picking phenomenon of conventional case analysis. This problem arises when scholars develop theories from a "small series of cases." (190) While those theories can be thought provoking, they rarely provide a "robust explanation of how the law works." (191) A handful of deliberately chosen cases cannot be considered a representative sample of the law. Conclusions are at best tentative or uncertain when drawn from an insufficiently large sample. (192) By randomly sampling cases, or by analyzing the entire pool of available cases on a subject, researchers can negate the biases of traditional case-analysis methods. (193)

Second, rigorous empirical methods help to discipline the researcher in reading selected cases. Proper quantitative analysis requires the researcher to develop a coding scheme before she reads the opinions. (194) A researcher develops a set of variables and criteria for valuing each variable for a given case. (195) When coding the cases, the researcher is forced to consider each variable as it applies to each case. Thus, the researcher is less likely to suffer from tunnel vision--picking and choosing the factors that subjectively stand out in a given opinion (196)

A. The Data Set

I compiled a random sample of cases from Westlaw, decided before October 17, 2011, that cite Davis. (197) These cases could be from any federal district court, federal court of appeals, or state court. The sample includes both reported and unreported decisions, as well as decisions that had been overturned on appeal. Of the 300 cases I sampled, I excluded 77. Those 77 were nonbinding magistrate judge recommendations and cases in which the court did not evaluate a Confrontation Clause challenge under Crawford's testimonial framework. The remaining 223 cases included several instances in which a single court evaluated multiple statements under the Crawford framework. Thus, the final sample consisted of 278 statements from 223 cases.

For each statement, I coded nineteen independent variables and one dependent variable--whether the reviewing court found the statement to be testimonial or not. The coding rules are described in detail in Appendix A. Table 3 lists the name of each variable and gives a description. I chose variables that both captured the important factors cited in the Crawford line and could be coded objectively.

The variable for measuring an injured party involved some subjective judgment on my part about whether an injury was serious enough to warrant medical attention. To address this, I included an alternate, more objective definition of an injury. The results of the regression analysis are almost identical between the two definitions, but the descriptive statistics do show some differences between the definitions, which are discussed below.

B. Descriptive Statistics

This Section presents descriptive statistics--statistics that summarize and explicate the data. Two conclusions were immediately obvious from the data. First, the government won the majority of cases. Overall, courts held that only 90 of the 278 challenged statements were testimonial. (198) Even this figure--that the defendants prevailed in only one-third of challenges--overestimates the success of defendants. In many of those 90 cases, courts ultimately gave judgment for the government on the basis of harmless error. (199) Second, the data showed a stark contrast in how courts evaluated statements made to different categories of recipients. For all of the graphs and data reported below, I did not include statements that were contained in forensic laboratory reports because the doctrine surrounding forensic analyses has become more refined than the general Crawford doctrine. This exclusion should have the effect of exaggerating any ambiguity or uncertainty in Confrontation Clause cases. Figure 1 plots the frequency with which statements to certain recipients were held to be testimonial.

Figure 1 shows that statements are more likely to be held to be testimonial when made to police officers or other government officials. They are less likely to be held to be testimonial when made to health professionals or 911 operators and very unlikely to be held to be testimonial when made to private citizens. Whereas half of the challenged statements made to police officers and nonpolice government officials were held to be testimonial, only around five percent of statements to private citizens were held to be so. The most curious feature of the data is the relatively high frequency, almost forty percent, with which statements to health professionals were held to be testimonial. This result is partly driven by statements to Sexual Assault Nurse Examiners (SANE nurses) and other health practitioners who specialize in treating victims of sexual assault and sexual abuse. These individuals are mandatory reporters who are trained to collect evidence as well as provide medical treatment. As one court explained, SANE nurse questioning is the "functional equivalent of police questioning." (200) Statements made to SANE nurses thus draw great scrutiny from reviewing courts.

Figure 2 combines the categories listed in Figure 1 to show the dramatic difference in how courts treat statements to state actors and statements to nonstate actors.

Without controlling for other factors, a statement is over three times more likely to be held to be testimonial when it is made to a state actor. Section II.C shows that the results are not quite so pronounced when one controls for other variables. Nonetheless, considering that the Crawford line has not placed a great deal of explicit emphasis on state action, these results are striking.

Less strikingly, the data also reflect different outcomes in cases where an injured party is involved in the events surrounding the statement. Figure 3 shows the outcomes in injured-party cases relative to non-injured-party cases.

Figure 3 shows that statements are, without controlling for other factors, less likely to be held to be testimonial when an injured party is involved. The results are most pronounced under Specification A, where an injury is defined as a shooting, stabbing, or other injury requiring immediate medical assistance. The presence of an injury should matter if courts are following the Crawford line. Davis stressed response to an ongoing emergency, (201) presumably in many cases a medical emergency. Bryant made this point explicit-emphasizing that physical injury plays a significant role in evaluating the testimonial status of statements. (202)

The reader might notice that the results in Figure 3 vary depending on how an injury is defined. When one includes not only shootings, stabbings, and other injuries requiring medical attention, but also kicks and punches, the results are weaker. Figure 4 explores the result further, showing the results by type of injury.

Crawford provides some reason to view different injuries differently. If the police see a victim with a black eye, they might respond more calmly and with greater investigatory emphasis than if they had found a victim with a gunshot wound. Statements made under those circumstances look more like substitutes for live testimony produced through the kind of ex parte examinations that troubled the Framers and the Crawford Court.

The results in Figure 4 suggest something different--when there is a slight injury, statements are more likely to be held to be testimonial than when there is no injury at all. This seemingly odd result is plausibly explained by the role of state action. Statements made around a slight injury are more likely to be made to state actors than statements made when there is no injury. (203) The role of state action may be overwhelming the effect of an injured party. To more fully disentangle the relative effects of state action and injury requires something more than description. To control for state action and injury--to fully evaluate the interaction of all the variables at play--I employed logistic regression.

C. Stepwise Logistic Regression Analysis

Regression analysis is a method for sorting and analyzing data. (204) The researcher assumes that a very general form of relationship exists between variables and attempts to estimate the parameters in that relationship. For example, a researcher might believe that an individual's salary depends on her level of education, gender, and height. Regression analysis allows the researcher to estimate, from a set of data about individuals' salaries, education levels, genders, and heights, how a change in one independent variable will affect salary.

Logistic regression is a particular type of regression. (205) Logistic regression is useful when the researcher hopes to explain the relationship between several explanatory variables and a binary dependent variable. (206) A binary variable is one that takes only two states. Whether someone has completed high school, for example, is a binary variable. A person either completes or does not complete high school. There is no third option. I employed logistic regression because the dependent variable in nay analysis--whether a statement is found to be testimonial--takes only two values.

Stepwise regression adds a final nuance to regression analysis. Imagine a scenario where a researcher is addressing a new problem without background knowledge. For example, the researcher has collected a great deal of data about death penalty defendants and wishes to know which of several hundred characteristics affect the probability of an execution. The researcher may not have enough data to run a useful regression analysis with all the variables. Stepwise regression takes several iterative steps that narrow down the set of variables to those explanatory variables which have a statistically significant relationship with the dependent variable. There are several different methods of stepwise regression. Some methods select variables into the final specification. These methods, known as forward selection, add variables to the specification in descending order of significance. The selection process ends when all statistically significant variables have been added. Other methods select variables out of the specification until all statistically insignificant variables have been eliminated. Stepwise regression thus helps avoid sample size problems by reducing the number of independent variables in the final regression specification. With fewer variables, fewer observations are required to have confidence in the resulting estimates. (207) Stepwise regression has been used frequently in empirical legal scholarship. Without using stepwise regression, the sample size here would be relatively small. There would be fewer than five events per variable--i.e., fewer than five occurrences of a statement being found testimonial for each independent variable. A sample that small would dramatically reduce confidence in the results. Stepwise regression eliminated statistically insignificant variables, resulting in more events per variable--more occurrences of a testimonial statement for each independent variable--in the final regression specification, thus allowing me to more accurately estimate the true relationship between the variables.

Table 4 presents the results of the regression analysis. I did not include in the regression analysis statements contained in forensic laboratory analyses for the same reasons that I excluded those statements from the descriptive statistics. The stepwise regression process eliminated several additional variables that were not statistically significant. (208) The middle column is the most important. (209) It shows how a change in the associated variable changes the probability that a statement will be found to be testimonial by the court, holding other variables constant. For example, other factors equal, the fact that a statement is made to a police officer makes it about 75% (between 60% and 89%, with 95% confidence) more probable that the statement will be found to be testimonial. Similarly, other factors equal, the fact that a statement is made while one of the parties is injured decreases the probability that it will be found to be testimonial by about 18% (between 8% and 28%, with 95% confidence).

What conclusions can we draw from the regression results? First, state action matters. A lot. A given statement is almost twice as likely to be found to be testimonial, all else equal, when it is made to a state actor. The fact that the recipient of a statement is either a police officer or a nonpolice government official (for example, a fire marshal, prosecutor, or judge) substantially increases the likelihood that a court will find that statement to be testimonial. This is consistent with the descriptive statistics in Section II.B. It is also consistent with the scholars and lower courts who have emphasized the Confrontation Clause as a protection against abuses of state power. (210) It is even consistent with the subtext of Crawford and the history of confrontation abuses in England. (211) State action, however, has only been explicitly emphasized by the Court in Bryant. (212) And even Bryant only made a comparative statement of priority; it did not state that the Confrontation Clause is exclusively or nearly exclusively concerned with statements made to state actors.

State v. Parker, a case decided by Tennessee's Court of Criminal Appeals, demonstrates the important role of state action in Confrontation Clause cases. (213) There, the victim of a home invasion was sexually assaulted and fled to her neighbor's house. (214) She recounted her attack to her neighbor, who dialed 911. (215) The police arrived several minutes later and questioned the victim, whose statements were "parallel[]" to those she had initially made to her neighbor. (216) Hours later, the victim told a nurse at the hospital that her son's friend had attacked her. (217) The Parker court, after reciting an eight-factor test for testimonial statements, concisely noted that "[s]tatements describing past events to law enforcement officers are testimonial" except "when [those] statements ... are meant to assist the officers in meeting an ongoing emergency." (218) On that basis, the court held that the victim's statements to her neighbor, a nonstate actor, were nontestimonial while nearly identical statements made by the same victim minutes later were testimonial because they were made in response to police questioning. (219)

Second, some of the results are quite consistent with Davis's emphasis on emergency response. For example, statements made during a crime--in the midst of an emergency with uncertain outcomes--are slightly less likely to be found to be testimonial. The same is true of statements held to be excited utterances and those made while a party is injured.

Several results are more ambiguous. For example, statements made during 911 calls are forty percent more likely to be testimonial. This makes some sense, because 911 operators are state agents, although they are not police officers. (220) Without controlling for other factors, 911 operators confront courts with conflicting indicators. They are state actors, yet they act, almost by definition, during emergencies. Regression isolates the effect of a 911 operator, independent of the emergency during which the operator acts. The data suggest that courts recognize 911 operators as state actors. This fact can be hidden because so many 911 calls occur during medical emergencies, but when one controls for that effect, the emphasis that courts place on state action becomes clear. (221)

The results for health professionals are also somewhat ambiguous. The abuse-counseling results seem to fit the state-action theory. Courts are often critical of social workers' interviews of children because they view these interviews as means to collect evidence, (222) and because the interviews sometimes occur at the instigation of the police. That skepticism might also partly explain the result for other health professionals. Health professionals include not only EMTs and doctors but also SANE nurses. As explained above, these professionals draw greater scrutiny from courts. (223)

State v. Steele, decided by the Ohio Court of Appeals, illustrates the complex interplay of medical emergencies, police interrogations, and discussions with health professionals. (224) Police arrived at the apartment of Virginia Austin to find her lying on the ground, partially disrobed and crying. (225) She had also been struck on the head. (226) Austin told the officers that that she had been assaulted. (227) The police called an ambulance, and Austin told the arriving EMT that she had been attacked (228) and repeated that statement to the police. (229) Finally, Austin gave a more detailed report to police at the hospital. (230) The court applied the "primary purpose" test to hold that Austin's initial statement to the police was nontestimonial; there was an ongoing emergency when she first spoke. (231) Her subsequent statements to police, however, were made after the medical emergency had passed and were therefore testimonial. (232) Although Austin's statement to the EMT was nearly contemporaneous with her second statement to the police, the court applied the "objective witness" test to the nonpolice EMT and thus held that statement to be nontestimonial. (233)

To summarize, the regression results show that two considerations matter-and in a specific, ordered way. First, statements made to nonstate actors are almost never testimonial statements. If a statement is not made to a police officer, government official, 911 operator, or mandatory reporter (such as a doctor or SANE nurse), the Confrontation Clause usually does not apply.

Second, assuming the statement is made to a state actor, courts consider factors showing an emergency. The greater the indication of a medical or public safety emergency, the more likely courts are to find a statement to be nontestimonial. Those two factors--a state actor and the presence of an emergency--do not perfectly predict every case outcome. But they explain most outcomes.

D. Limitations of the Data

Like all regression analysis and content analysis of judicial opinions, my data are subject to limitations. As explained above, using stepwise regression eased sample-size problems, but the sample size was still less than ideal. (234) A related concern, multicollinearity of the independent variables, was not a problem. (235) Multicollinearity refers to high correlation among independent variables. As an example, multicollinearity might confound a researcher attempting to estimate the effects of GPA and LSAT scores on law school admissions. Because higher LSAT scores and higher GPAs may frequently occur together, teasing out the effect of a change in one factor or the other could be difficult. The effect of multicollinearity is the same as the effect of a small sample size: it reduces the statistical significance of the regression coefficients for a given sample size and specification. (236) Both multicollinearity and small-sample-size problems do not bias the results in one direction or another. Rather, their effects are to reduce the confidence a researcher can have in her estimates. No variable here had a variance inflation factor (VIF) greater than three, while multicollinearity is usually a problem only with VIF scores of greater than ten. (237)

Sample-selection bias is also unlikely to be a problem. Sample-selection bias occurs when the sampling procedures yield a sample that is not representative of the population being analyzed. (238) In this case, the population is lower-court decisions in Confrontation Clause cases. The only criterion for selection was citing Davis. I chose Davis rather than Crawford for two reasons. First, Davis was decided two years after Crawford. Thus, lower courts had had time to adjust to the Supreme Court's new Confrontation Clause framework. Second, Davis helped resolve some common but ambiguous fact patterns (911 calls and excited utterances), but it also heightened important ambiguities in the Crawford line. (239) Because Davis, unlike Crawford, dealt with domestic violence and 911 calls, it is possible that domestic violence cases could be overrepresented in the sample because courts would be less likely to cite Davis in other types of cases. This possibility seems unlikely (but not completely improbable). In my experience reading cases and developing the coding mechanisms, most courts seriously considering Confrontation Clause cases cite both Davis and Crawford. There is one small possibility of selection bias, stemming from the difference between the sampling unit and the unit of analysis. While I sampled cases, I analyzed statements. Thus, statements from cases where the court evaluated several statements would be slightly underrepresented. This would only matter if statements contained in cases with multiple statements systematically differed from other statements. I do not believe there is any particular reason that would be true.

A more substantial concern is coding reliability. The coding rules are explained in Appendix A. These rules are generally objective. For example, a court says a statement is either testimonial or is not testimonial--if the court does not give any indication on that issue, then the case was excluded. Similarly, a statement was or was not made during a 911 call. At the same time, I did have to make a few judgment calls during the coding process. (240) The only variable that raised a concern of consistently requiring subjective evaluation was the "injured party" variable. This variable was too important to drop, so I developed an alternative and more objective coding method. Instead of judging whether an injury was medically serious, the alternative coding treated stabbings and shootings one way, and other physical injuries another way. Although reliable and consistent coding is important, some inconsistency or uncertainty is not fatal. For example, seventy-five to eighty percent agreement between coders is generally considered acceptable in content analyses. (241) After I completed my analysis, several students, unfamiliar with my coding or results, recoded a sample of the cases. Their results indicated that the initial coding is relatively reliable. The details and results of this reliability check are explained in Appendix B.

A final troubling problem is endogeneity. Endogeneity could refer to one of two problems. First, reverse causality could occur if an independent variable not only causes a dependent variable (for example, a 911 call causes an increased probability that the statement will be found to be testimonial) but is also caused by the dependent variable. That is unlikely here for the simple reason of time. The properties of a statement occur first, and the statement is classified as testimonial later by a reviewing court. That review cannot change the prior events. A second form of endogeneity, omitted-variable bias, is more troubling. There is no test for omitted-variable bias. While I have attempted to test for the factors reflected in the Crawford-line decisions, it is possible that other facts about each case--the characteristics of the offender and victim, the heinousness of the charged crime, and the strength of other evidence--have a role to play. It is quite difficult to code for some of these variables. To the extent that one has a coherent theory about why they would matter, evaluating their effects is a project for future researchers.


This Part offers a normative take on the data presented in Part II. I offer two primary arguments. First, critics have exaggerated the incoherence of Crawford in the lower courts. Even if the Court has failed to clarify its doctrine, lower courts have responded admirably, bringing consistency to Confrontation Clause doctrine. Second, lower courts have been not only consistent but also correct in how they have approached the Confrontation Clause. By focusing on the state's investigatory power, lower courts have captured not only the concerns that animated the adoption of the Clause, but also the deep concerns of Crawford itself.

A. The Confrontation Clause, Consistently

Crawford and its progeny have a problem of doctrinal vagueness. Part I explained how certain statements might or might not be testimonial depending on the definition applied by the courts. Scholars have argued the same point. (242) Courts have emphasized those concerns as well. (243) The result, according to critics, is that lower courts are left unable to make heads or tails of the Confrontation Clause. (244) This problem is exacerbated, according to a much smaller group of scholars, by the politically motivated desires of some courts to seize on any excuse for diminishing the scope of the Confrontation Clause. (245)

The results presented in this Note suggest that problems with Crawford's vagueness, coherence, and consistent applicability have been exaggerated. The evidence presented in Part II indicates that lower courts effectively engage (246) in a bifurcated analysis. The first step is relatively clear. Courts ask whether the recipient of a challenged statement was a state actor. If the answer is no, then the Confrontation Clause is almost never relevant. If the answer is yes, and the recipient is a state actor, then step two applies. In step two, lower courts consider the circumstances surrounding the challenged statement. The most important circumstance appears to be the existence of an emergency, especially a medical emergency, a result entirely consistent with Davis. The objective factors present in any case thus have significant predictive power. Knowing whether a statement is made to a police officer or a friend (247) or whether a nearby party is injured (248) tells us a lot about how a court will classify that statement. If Crawford were truly the mess that critics claim, (249) no factors should have much success in predicting case outcomes.

Critics might respond that there is another way to read my results. We can predict a great deal just from the involvement of a state actor and a medical emergency, but we cannot predict everything. Or nearly everything. And if many cases are unexplained by the data, doesn't that mean Crawford is doctrinally problematic?

I think this response demands too much. Judges are not computers, and judging is not computing. Crawford challenges involve guessing about subjective motivation (why did the police officer ask a question?), grading the seriousness of a problem (just how bad was the victim's injury?), and assessing the objective risk to a community. Since subjective assessment cannot be quantitatively measured, one cannot use it to statistically predict case outcomes. But that sort of uncertainty is not unfair to litigants. Litigants need to know what questions the court will ask and what evidence is relevant. The results in Part II provide litigants with good answers to those questions. They should argue about the role of a state actor in procuring a statement and the existence of a medical emergency.

I do not deny that outliers exist. In a few cases, courts probably got it wrong. (250) But a focus on selected individual cases can obscure the important large-scale patterns. (251) Lower courts sometimes flout the Supreme Court's stated doctrine, but practitioners and scholars should be concerned with the grand run of cases rather than curious outliers. (252) When one considers that larger population of Confrontation Clause cases, the results are more encouraging. Critics overstate the issue when they call the state of affairs "arbitrary" (253) or "confusing." (254)

B. Lower-Court Decisions and the Meaning of Confrontation

Accepting my results, a different set of critics might respond that if lower courts have converged, they have converged at the wrong site. Lower courts may have adopted a common framework for Confrontation Clause challenges, but this framework is wrong--either as a matter of history or when measured against Crawford's reasoning.

As explained above, (255) courts applying Crawford look first for the presence of a state actor, and second for the presence of a medical emergency in deciding whether a statement is testimonial. In other words, the Confrontation Clause applies most stringently to statements made to state actors acting in their investigative capacity. It applies less stringently to state actors acting in their emergency-response or public-safety capacity, and rarely to nonstate actors.

This theory already has scholarly support. (256) Proponents argue that without confrontation, state actors can manipulate evidence in both production and presentation to obtain a politically motivated or otherwise false conviction. (257) This theory also has a historical foundation: the concern of the Framers with the abuses of inquisitorial-style prosecutions in England, such as the Star Chamber and the trial of Raleigh. (258) Moreover, this interpretation is textually plausible. The Confrontation Clause is situated among other constitutional protections that can be read, as a whole, to form a bulwark against government abuses. (259) And the Clause applies only to criminal trials as a protection for the accused, not for the government. (260)

This theory reflects the underlying principles of Crawford. The Crawford majority extensively detailed the history of English abuses and the early American concern with the civil ex parte mode of examination. (261) And, "[i]t is only after this discussion that the Court set forth its three formulations of testimonial hearsay." (262) Crawford's three definitions and Davis's definitions thus play a distinctly secondary role in how we should properly read the decision. They are illuminative or exemplary. They should be read through the lens of the history. The definitions "should [be] construe[d] [as] three formulations ... directed at the same concern: preventing the government from using hearsay statements at trial if the statements were obtained through a method that resembles a civil-law mode of interrogation." (263)

Scholars, and sometimes courts, have a tendency to overread Crawford's definitions. For example, if one adopts the "objective witness" definition, it is natural to view almost every interaction between private citizens and law enforcement as testimonial. After all, anything a person says to the police that is uncoerced can potentially end up in court. (264) For this reason, defense attorneys often advise clients to say nothing to the police. And one would therefore think that every statement to the police would, objectively, be expected to be used in a criminal trial. However, if one reads the "objective witness" definition purposively, with the purpose being the exclusion of ex parte examinations, then lower-court decisions make more sense. All statements to police officers might end up in court. But not all statements to police officers are ex parte examinations.

A person not speaking to a state actor is almost never participating in anything like a civil ex parte examination. Thus the distinct role of state action in lower-court decisions. Moreover, a police officer who arrives at the scene of an emergency and tends to a wounded victim has little resemblance to an interrogator. By contrast, SANE nurses are trained to collect evidence and assess sexual assault. Their structured questioning has much more in common with the ex parte examinations that concerned the Framers than does the conduct of a police officer who arrives along with the ambulance. Lower courts, by excluding testimony from SANE nurses, or by excluding post-Miranda statements but including statements to first-responding police, are hewing closely to Crawford's contours.

Scholars often want Crawford to stand for a favored principle--a generically strong Confrontation Clause, (265) the need for evidence to be reliable, (266) or a repudiation of judicial discretion in confrontation decisions (267)--without allowing the decision to stand on its own terms. The Court was not trying to compromise competing interests or gesture nebulously in the direction of robust or weak confrontation rights. Crawford provided a specific protection to defendants to confront, live and in court, accusers who make testimonial statements against them. Defendants can ask for no more, and the government can provide no less. This may or may not provide more protection on the whole than the pre-Crawford regime. But it is a mistake to attack lower courts for failing to provide defendants with robust protection. The courts are, by and large, providing defendants exactly the protection the Constitution requires. (268)


This Note is one of the first attempts to understand empirically how lower courts have applied Crawford. The Crawford line of cases made a clear doctrinal break with Roberts. Post-Crawford doctrine, however, has not been terribly clear. Indeed, the doctrine has appeared contradictory at times. Lower courts could theoretically achieve different outcomes by stressing different elements of the doctrine. Scholars have cited these facts to argue that lower courts can effectively ignore Crawford and pursue their own policy preferences. Moreover, even lower courts that act in good faith and try to comply with Crawford can still issue inconsistent outcomes.

This Note investigated the lower courts empirically. I presented data showing that lower-court decisions are not random or arbitrary, as Crawford's critics would suggest. Instead, two factors-whether a statement is made to a state actor and whether a statement is made during a medical emergency-under-gird the actions of lower courts. Not every lower-court decision can be predicted entirely by absolutely objective variables. But the grand run of lower-court Confrontation Clause decisions is far from arbitrary or unpredictable.

Finally, I argued that lower courts are not only deciding Confrontation Clause challenges consistently, but they are also doing so in a manner that squares with the text and history of the Confrontation Clause, and with the Crawford Court's theory of that Clause. Lower courts are regulating the state's coercive power, and they are doing so in those circumstances in which Crawford calls for them to do so.


Table 5.

VARIABLE             CODED IF ...              CODED IF ...

Irrelevant           (1) The opinion did not   The opinion evaluated a
                     address the               Confrontation Clause
                     Confrontation Clause,     challenge, including if
                     (2) the opinion was a     the opinion stated how
                     nonbinding magistrate     the court viewed the
                     judge recommendation,     matter but ruled
                     or (3) the opinion did    primarily on harmless
                     not rule on or offer an   error or another
                     opinion on a              procedural question.
                     Confrontation Clause

Testimonial          The statement was held    Otherwise.
                     by the court to be
                     testimonial, including
                     if the court said the
                     statement was likely
                     testimonial but did not
                     resolve the issue with

During               The statement was made    Otherwise, including
                     during commission of      statements made between
                     the charged crimes, or    overt acts of a
                     immediately after the     conspiracy.
                     crimes were completed.

Before               The statement was made    Otherwise.
                     before the commission
                     of any of the charged

911                  The statement was made    Otherwise, including if
                     during a q11 call or      (1) the statement was a
                     was a Computer-Aided      classification of a 911
                     Dispatch report of a      call by a 911 operator
                     911 call.                 or (2) the statement
                                               was a report of a 911
                                               dispatcher to a police

Police               The statement was made    Otherwise.
Station/Court        at (t) a police
                     station, (2) a
                     courthouse, or (3)
                     another government

Abuse                The statement was made    Otherwise, including if
                     at a facility for         the statement was made
                     treatment of sexual       at a hospital or other
                     abuse or domestic abuse   general medical
                     during an interview by    establishment.
                     a trained professional.

Medical              The statement was made    Otherwise.
                     (i) in a medical
                     record, (2) in medical
                     notes that do not
                     formally qualify as a
                     medical record, or (3)
                     during medical

Recorded             The statement was         Otherwise.
                     audio- or video-
                     recorded. 9i1 calls
                     were coded as 1 unless
                     the opinion explicitly
                     indicated to the

Dying Declaration    The statement was found   Otherwise.
                     to be a dying
                     declaration. *

Excited Utterance    The statement was found   Otherwise.
                     to be an excited
                     utterance. *

Present Sense        The statement was found   Otherwise.
Impression           to be a present sense
                     impression. *

Government/          The statement was found   Otherwise.
Business Record      to be a government or
                     business record. *
                     Forensic analyses and
                     laboratory reports were
                     coded as 1 unless the
                     opinion explicitly
                     indicated to the

Cop                  The statement was made    Otherwise, including
                     to a police officer       business or government
                     other than a 911          records not kept with
                     operator or an            specific police
                     undercover agent.         involvement (such as
                     Certificates of no        DMV records).
                     report and forensic
                     analyses were coded as
                     1 unless the opinion
                     explicitly indicated to
                     the contrary.

Doctor               The statement was made    Otherwise, including if
                     to a health               the statement was made
                     professional including    to a veterinarian or
                     a doctor, nurse, or       social worker.
                     EMT. If multiple
                     parties were present,
                     the statement was coded
                     as a 1 if a health
                     professional was

Government           The statement was made    Otherwise.
                     to a government
                     official other than a
                     police officer or 911

Victim Speaking      The declarant of the      The declarant of the
                     statement was the         statement was not the
                     victim of the crimes      victim of the crimes
                     charged in the case,      charged in the case,
                     where the crime charged   including crimes that
                     had a defined victim.     were treated as having
                     Those crimes included     no identifiable victim:
                     murder, assault,          drug crimes, financial
                     battery, robbery, and     crimes, record-keeping
                     criminal sexual           violations, and crimes
                     conduct.                  not resulting in bodily
                                               contact with another

Injured-             Declarant or a known      Otherwise, including if
Specification A      nearby party was (1)      a party was injured but
                     shot or stabbed, or (2)   not seriously enough to
                     treated by emergency      warrant medical
                     medical technicians or    treatment. Absent other
                     treated at a hospital     physical injury, sexual
                     for injuries, and those   assaults and rapes were
                     injuries were medically   coded as 0.

Injured--            Declarant or a known      Otherwise, including in
Specification B      nearby party was shot     cases of sexual assault
                     or stabbed. Coded as      or rape without other
                     0.5 if declarant or a     injury.
                     known nearby party was
                     physically battered but
                     not shot or stabbed.

* For all hearsay codings, I applied several rules: (1) absent some
indication in the opinion, briefing, or lower-court opinions that the
court applied an exception, I coded the variables as 0; (2) I coded
the variables as 1 if the trial court held the statement to fall
within the relevant hearsay exception so long as the reviewing court
did not overturn that holding; and (3) I coded closely labeled state
equivalents for the same variable (for example, California's
Spontaneous Utterances were treated as Excited Utterances).


During the editing process, I checked the reliability of my initial coding of the cases. Approximately eighty statements from the sample, or slightly less than one-third, were recoded by individuals who did not know the original results. I then compared the results of the recoding to my initial coding, using Cohen's Kappa (269) to measure the level of agreement. Cohen's Kappa is one of several measurements designed to account for chance agreement between coders. (270) Accounting for chance is particularly important in this case, because certain fact patterns, such as a dying declaration, occur infrequently. Thus, agreement by chance will be quite common even if coders are guessing (provided that they are guessing with roughly the same distributions of responses).

Table 6 lists values of Cohen's Kappa for each variable. Unfortunately, there is no general agreement about the level of agreement required. Whereas some scholars suggest that only values greater than 0.7 or 0.8 are acceptable, others accept values as low as 0.4. (271)

Variables listed in the left-hand column are identical to those listed in Table 3, but are listed in descending order based on the value of Cohen's Kappa. Values in the right-hand column have been rounded to three digits.

Several conclusions can be drawn from the reliability analysis. First, coding for many of the important variables is fairly reliable. Those variables most important to the analysis--for example, the dependent variable or that a statement is made to a police officer--had high Cohen's Kappa values.

Second, some variables are less reliable, and a few (the temporal variables during and before, and the abuse counseling variable) are downright untrustworthy. Fortunately, the least reliable variables were relatively unimportant to the descriptive analysis of Section II.B. Several more important variables, however, were also less than ideally reliable. In particular, the government actor (nonpolice) and injury variables both had Kappa values less than 0.6. These low values are not fatal to my analysis, but they do make the results a bit more tentative.

Finally, it is worth noting that the alternative specification of an injury was in fact far more reliable than the initial definition with a Cohen's Kappa value of nearly 0.8, versus a value of 0.6 for the initial definition.

There are also reasons to be somewhat cautious about the reliability analysis itself. First, the recoders had fairly little training. It is possible that some of the disagreements between their recoding and the original coding may be due to misinterpretation of a rule. Second, certain variables occurred relatively infrequently in the data. Thus, the reliability analysis would show chance agreement to be common, and even a low number of disagreements could produce a very low Kappa. This, for example, is the case with the nonpolice government agent variable and the abuse counseling variable.
Table 6.

VARIABLE                             VALUE OF COHEN'S KAPPA

Present                              1
Dying                                1
Habeas                               0.940
Excited                              0.904
Testimony                            0.892
Victim Speaking                      0.847
Taped                                0.824
Injured, Alternative Specification   0.794
911                                  0.789
Forensic                             0.749
Co                                   0.747
Doc                                  0.707
Inured                               0.588
Medical                              0.586
Gov                                  475
Station                              0.431
Before                               0.410
Record                               0.304
During                               0.204
Abuse                                -0.017

(1.) 541 U.S. 36 (2004).

(2.) Compare id. at 51 ("The constitutional text, like the history underlying the common-law right of confrontation, thus reflects an especially acute concern with ... 'testimonial' statements...."), Davis v. Washington, 547 U.S. 813, 821 (2006) ("In Crawford v. Washington, we held that [the Confrontation Clause] bars 'admission of testimonial statements....'" (citation omitted) (quoting 541 U.S. at 53-54)), and Whorton v. Bockting, 549 U.S. 406, 420 (2007) ("Under Crawford ... the Confrontation Clause has no application to [out-of-court nontestimonial] statements and therefore permits their admission even if they lack indicia of reliability."), with Ohio v. Roberts, 448 U.S. 56, 66 (1980) (holding that when a hearsay declarant is unavailable for trial, "his statement is admissible only if it bears adequate 'indicia of reliability'"), abrogated by Crawford, 541 U.S. 36.

(3.) U.S. CONST. amend. VI ("In all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him....").

(4.) See infra Section I.B.

(5.) See infra Sections I.B-D.

(6.) See, e.g., United States v. Brito, 427 F.3d 53, 55 (1st Cir. 2005) ("Crawford ... effected a sea change in the jurisprudence of the Confrontation Clause...."); Anderson v. Jackson, 567 F. Supp. 2d 973, 982 (E.D. Mich. 2008) (characterizing Craw ford's interpretation of the Confrontation Clause as a "radical transformation"); People v. Pirwani, 14 Cal. Rptr. 3d 673, 688 (Ct. App. 2004) ("Crawford radically alters the way we analyze claims of error under the Confrontation Clause...."); State v. Snowdeu, 867 A.2d 314, 322 (Md. 2005) (stating that Crawford "fundamentally altered" Supreme Court Confrontation Clause jurisprudence); Robert P. Mosteller, Crawford v. Washington: Encouraging and Ensuring the Confrontation of Witnesses, 39 U. RICH. L. REV. 511, 626 (2005) ("After Crawford, the world of confrontation law has been radically altered."); Richard D. Friedman, Adjusting to Crawford: High Court Decision Restores Confrontation Clause Protection, CRIM. JUST., Summer 2004, at 4, 5 (stating that Crawford "radically transformed" Confrontation Clause doctrine).

(7.) See, e.g., Charles F. Baird, The Confrontation Clause: Why Crawford v. Washington Does Nothing More than Maintain the Status Quo, 47 S. TEX. L. REV. 305, 320-24 (2005) (arguing that Crawford "guarantees the status quo--the continued admission of statements that the Confrontation Clause was meant to exclude," because it fails to define "testimonial statements" while vesting discretion in judges to decide which statements reach the jury); Mark Dwyer, Crawford's "Testimonial Hearsay" Category: A Plain Limit on the Protections of the Confrontation Clause, 71 BROOK. L. REV. 275, 277 (2005) (arguing that based on a careful parsing of the case, "nothing in Crawford justifies this hope that virtually all hearsay statements are still within the reach of the Confrontation Clause"); John R. Grimm, Note, A Wavering Bright Line: How Crawford v. Washington Denies Defendants a Consistent Confrontation Right, 48 AM. CRIM. L. REV. 185, 205-11 (2011) (using the examples of 911 calls and medical diagnoses to argue that Crawford's amorphous test will yield largely similar results to those arising under Ohio v. Roberts); cf. Lisa Kern Griffin, Circling Around the Confrontation Clause: Redefined Reach but Not a Robust Right, 105 MICH. L. REV. FIRST IMPRESSIONS 16, 19-21 (2006) (arguing that although Crawford facilitates much broader application of confrontation rights, those rights are hollow because other decisions have weakened the meaning of confrontation itself). Other commentators have argued that while Crawford held the potential to provide defendants with more robust confrontation rights, subsequent decisions have neutered that potential. See, e.g., Michael D. Cicchini, Dead Again: The Latest Demise of the Confrontation Clause, 80 FORDHAM L. REV. 1301, 1312-17 (2011) (arguing that the Court's decision in Michigan v. Bryant, 131 S. Ct. 1143 (2011), undermined Crawford by reintroducing issues of reliability into Confrontation Clause analysis through the use of the rules of evidence to determine whether statements are testimonial).

(8.) Many critics of Crawford argue that the decision will undermine domestic violence prosecutions. See, e.g., Tom Lininger, Prosecuting Batterers After Crawford, 91 VA. L. REV. 747, 749-50 (2005) ("[W]ithin days--even hours--of the Crawford decision, prosecutors were dismissing or losing hundreds of domestic violence cases that would have presented little difficulty in the past.... In a survey of over 60 prosecutors' offices ... 63 percent of respondents reported the Crawford decision has significantly impeded prosecutions of domestic violence." (footnotes omitted)). But see Michael Baxter, Note, The Impact of Davis v. Washington on Domestic Violence Prosecutions, 29 WOMEN'S RTS. L. REP. 213, 226 (2008) ("While the Davis decision is not the most satisfying for those who seek strong domestic violence prosecutions, it certainly does not terribly hinder the use of most effective and traditional forms of hearsay for domestic violence, namely the 911 call and excited utterance."). Other critics argue that Crawford's originalist analysis of the Confrontation Clause is inaccurate. See, e.g., Thomas Y. Davies, Not "The Framers' Design": How the Framing-Era Ban Against Hearsay Evidence Refutes the Crawford-Davis "Testimonial" Formulation of the Scope of the Original Confrontation Clause, 15 J.L. & POL'Y 349, 434 (2007) ("Hence, however 'reasonable' the originalist inferences that Justice Scalia drew in Crawford might have appeared when viewed in isolation, they collide head-on with the evidentiary doctrine that actually shaped the Framers' understanding of the confrontation right. Admitting unsworn, 'nontestimonial' hearsay was not part of 'the Framers' design.'" (footnote omitted)). But see Stephanos Bibas, Originalism and Formalism in Criminal Procedure: The Triumph of Justice Scalia, the Unlikely Friend of Criminal Defendants?, 94 GEO. L.J. 183, 192 (2005) ("Crawford was a successful blend of originalism and formalism.... [Crawford's] formalistic rule is not only clear, but also rooted in the historical record, giving it objective legitimacy. It serves the historical goal of constraining judicial discretion and testing evidence before jurors' eyes. And there is no easy way to evade the rule....").

(9.) Crawford v. Washington, 541 U.S. 36, 51-52 (2004); see infra Section I.B.

(10.) Davis v. Washington, 547 U.S. 813, 822 (2006) (stating the "primary purpose" definition of a testimonial statement); see infra Sections I.C-D; see also Bryant, 131 S. Ct. at 1153 n.2 ("We noted in Crawford ... that '[j]ust as various definitions of "testimonial" exist, one can imagine various definitions of "interrogation," and we need not select among them in this case.' Davis did not abandon those qualifications; nor do we do so here." (citation omitted)); Crawford, 541 U.S. at 68 ("We leave for another day any effort to spell out a comprehensive definition of 'testimonial.'").

(11.) Josephine Ross, After Crawford Double-Speak: "Testimony" Does Not Mean Testimony and "Witness" Does Not Mean Witness, 97 J. CRIM. L. & CPIMINOLOGY 147, 149 (2006). Some scholars have argued that this ambiguity undermines Crawford's effectiveness in protecting criminal defendants. See, e.g., Grimm, supra note 7, at 190, 205-08, 210-11.

(12.) Michael D. Cicchini, Judicial (In)Discretion: How Courts Circumvent the Confrontation Clause Under Crawford and Davis, 75 TENN. L. REV. 753, 764-67 (2008) (arguing that the ambiguity of the Crawford tests has left them open to manipulation by lower courts).

(13.) Andrew Etter, Comment, Embracing Crawford: The Rights of Defendants and Children Under the Confrontation Clause, 77 U. CIN. L. REV. 1167, 1176 (2009).

(14.) Clifford S. Fishman, Confrontation, Forfeiture, and Giles v. California: An Interim User's Guide, 58 CATH. U. L. REV. 703, 750 (2009).

(15.) Deborah Ahrens & John Mitchell, Don't Blame Crawford or Bryant: The Confrontation Clause Mess Is All Davis's Fault, 39 RUTGERS L. REC. 104, 105 (2012).

(16.) Myrna S. Raeder, Domestic Violence Cases After Davis: Is the Glass Half Empty or Half Full?, 15 J.L. & POL'Y 759, 775-76 (2007).

(17.) Deborah Tuerkheimer, Exigency, 49 ARIZ. L. REV. 801, 832 n.173 (2007).

(18.) Cicchini, supra note 7, at 1316-17 (quoting Michigan v. Bryant, 131 S. Ct. 1143, 1167 (2011) (Thomas, J., concurring)).

(19.) People v. Rawlins, 884 N.E.2d 1019, 1026 (N.Y. 2008) (distinguishing the Crawford line's varying "formulations" of testimonial statements from the "dues" to help determine whether a statement is testimonial).

(20.) Compare People v. Moscat, 777 N.Y.S.2d 875 (Crim. Ct. 2004) (holding a 911 call to be nontestimonial), with People v. Cortes, 781 N.Y.S.2d 401 (Sup. Ct. 2004) (holding a 911 call to be testimonial).

(21.) Compare State v. Bobadilla, 709 N.W.2d 243, 255-56 (Minn. 2006) (concluding that under the objective observer test a child witness's statement was nontestimonial), with People v. Sisavath, 13 Cal. Rptr. 3d 753, 758 n.3 (Ct. App. 2004) (concluding that, in the case of a child witness's statement, the Crawford "objective witness" is not "an objective four year old").

(22.) Compare Martin v. State, 936 So. 2d 1190 (Fla. Dist. Ct. App. 2006) (holding a laboratory report was testimonial), with Commonwealth v. Verde, 827 N.E.2d 701 (Mass. 2005) (holding certificates of analysis were nontestimonial).

(23.) 131 S. Ct. 1143.

(24.) 132 S. Ct. 2221 (2012).

(25.) There are a few empirical analyses of the post-Crawford Confrontation Clause. See Donald A. Dripps, On Reach and Grasp in Criminal Procedure: Crawford in California, 37 N.C.J. INT'L L. & COM. REG. 349 (2011) (classifying recent California Confrontation Clauses cases by outcome and reasoning); Josephine Ross, Crawford's Short-Lived Revolution: How Davis v. Washington Reins in Crawford's Reach, 83 N.D.L. REV. 387, 460 (2007) (classifying post-Crawford, pre-Davis cases by the line of reasoning courts used to classify excited utterances as nontestimonial); Eleanor Simon, Confrontation and Domestic Violence Post-Davis: Is There and Should There Be a Doctrinal Exception?, 17 MICH. J. GENDER & L. 175, 185-99 (2011) (conducting an empirical analysis of 137 statements in eighty-two cases to determine whether there was a de facto "Domestic Violence 'Exception'" to the Crawford/Davis line). This Note is distinct from each of these analyses. Dripps focuses exclusively on California courts. His chief concern is federalism and state interests, rather than compliance by lower courts. Most importantly, his method depends on a subjective evaluation of the quality of each opinion's reasoning. Rather than classifying and evaluating opinions based on objective facts about each case, Dripps applies his own subjective judgment about whether each opinion is consistent with Crawford. See Dripps, supra, at 373-81. Ross's article categorizes state court decisions by their Confrontation Clause doctrine--for example, whether the decisions employ a primary purpose or a formality test--but Ross neither analyzes how the facts of those cases influence outcomes nor conducts any statistical analysis. See Ross, supra, at 460. Simon's analysis is empirical, large-scale, quantitative, and objective. But Simon focuses exclusively on domestic violence cases, and she does not present an overarching theory of the Confrontation Clause.

(26.) See infra Sections I.B-D.

(27.) See infra Parts II, III.

(28.) See infra Part III.

(29.) U.S. CONST. amend. VI.

(30.) See Richard D. Friedman, Confrontation: The Search for Basic Principles, 86 GEO. L.J. 1011, 1011-12 (1998) ("For the most part, however, the boundaries of the confrontation right as applied to trial witnesses are tolerably clear.").

(31.) See Delaware v. Fensterer, 474 U.S. 15, 18-19 (1985) (noting that while the "literal right to confront the witness at the time of trial ... forms the core of the values furthered by the Confrontation Clause," the Confrontation Clause also guarantees a right to cross-examine the witness (quoting California v. Green, 399 U.S. 149, 157 (1970)) (internal quotation marks omitted)); see also Dwain White, Note, Barber v. Page, 390 U.S. 719 (1968), 47 TEX. L. REV. 331, 334 (1969) ("There has always been general agreement among courts and commentators that the primary right encompassed within the constitutional guarantee of confrontation was that of cross-examination.").

(32.) See Akhil Reed Amar, Sixth Amendment First Principles, 84 GEO. L.J. 641, 647 (1996) (discussing the meaning of a "witness" in the Sixth Amendment).

(33.) See Crawford v. Washington, 541 U.S. 36, 42 (2004) ("The Constitution's text does not alone resolve this case."); Friedman, supra note 30, at 1012 ("The more pervasive perplexity arises [when] ... the declarant herself ... does not testify at trial.").

(34.) See Crawford, 541 U.S. at 42-43 (citing Woodsides v. State, 3 Miss. 655, 664-65 (1837)). Crawford, of course, rejects this view. Id. at 50-51 ("Accordingly, we once again reject the view that the Confrontation Clause applies of its own force only to in-court testimony....").

(35.) See id. at 42-43; cf. Amar, supra note 32, at 692 ("In ordinary language, when witness A takes the stand and testifies about what her best friend B told her out of court, A is the witness, not B."). Professor Amar ultimately endorses a slightly broader definition of "witness," one that covers, for example, "videotapes, transcripts, depositions, and affidavits when prepared for court use and introduced as testimony." Id. at 693.

(36.) See Crawford, 541 U.S. at 42-43 ("One could plausibly read 'witnesses against' a defendant to mean ... those whose statements are offered at trial...." (citing 3 JOHN HENRY WIGMORE, A TREATISE ON THE ANGLO-AMERICAN SYSTEM OF EVIDENCE IN TRIALS AT COMMON LAW 1397, at 104 (2d ed. 1923))). This view is arguably quite close to the views of the Framers. See Davies, supra note 8, at 352 ("[F]raming-era sources indicate that the confrontation right itself prohibited the use of hearsay statements as evidence of the defendant's guilt.").

(37.) See Simon, supra note 25, at 177 ("The Court has always understood witness statements [subject to the Confrontation Clause] to include in-court testimony as well as some, but not all, out-of-court hearsay statements."). Compare Bourjaily v. United States, 483 U.S. 171, 182 (1987) (rejecting the view that "the Confrontation Clause could bar the use of any out-of-court statements when the declarant is unavailable" as "unintended and too extreme" (quoting Ohio v. Roberts, 448 U.S. 56, 63 (1980))), with Sweta Patel, Comment, The Right To Submit "Testimony" via 911 Emergency After Crawford v. Washington, 46 SANTA CLARA L. REV. 707, 710 (2006) ("The U.S[.] Supreme Court has repeatedly rejected the proposition that the Confrontation Clause applies only to in-court testimony....").

(38.) See Crawford, 541 U.S. at 42-43 ("The Constitution's text does not alone resolve this case.... We must therefore turn to the historical background of the Clause to understand its meaning.").

(39.) GEORGE FISHER, EVIDENCE 567 (2d ed. 2008) ("In the Mattox era, the Supreme Court issued a number of ad hoc judgments to resolve particular controversies, but made little attempt to systematize the Confrontation Clause's impact on the admission of hearsay. The Court first undertook this task in earnest in Ohio v. Roberts." (citation omitted)).

(40.) Roberts, 448 U.S. at 66 (summarizing the two requirements of unavailability and reliability for admitting hearsay without violating the Confrontation Clause).

(41.) Id. ("[W]hen a hearsay declarant is not present for cross-examination at trial, the Confrontation Clause normally requires a showing that he is unavailable."). This requirement was gradually weakened following Roberts. See Friedman, supra note 30, at 1016 ("[T]he Court has cut back drastically on the unequivocal application of the unavailability requirement."); Randolph N. Jonakait, Restoring the Confrontation Clause to the Sixth Amendment, 35 UCLA L. REV. 557, 561 (1988) ("United States v. Inadi effectively abandoned Roberts's unavailability branch." (footnote omitted)).

(42.) Roberts, 448 U.S. at 66. ("[W]hen a hearsay declarant is not present for cross-examination at trial ... [r]eliability can be inferred without more in a case where the evidence falls within a firmly rooted hearsay exception. In other cases, the evidence must be excluded, at least absent a showing of particularized guarantees of trustworthiness.").

(43.) See Bibas, supra note 8, at 189 ("Roberts's tests for reliability proved to be murky, subjective, inconsistent, and unworkable."); Friedman, supra note 30, at 1022 ("[The Supreme Court's] approach [under Roberts] devalues the Confrontation Clause, treating it as a constitutionalization of an amorphous and mystifying evidentiary doctrine, the continuing value of which is widely questioned.").

(44.) See Jonakait, supra note 41, at 622 (stating that under the Roberts line, "evidence law now controls the content of the confrontation clause, and the clause now offers an accused little protection"); Lininger, supra note 8, at 760 ("[T]he 'Roberts test' was not much of a test at all.... Under either [a firmly rooted hearsay exception or the indicia-of-reliability test] Roberts seemed to abdicate the Supreme Court's responsibility for regulating the admission of hearsay that could violate a defendant's confrontation rights.").

(45.) See Amar, supra note 32, at 690-97 (arguing that the Court's Confrontation Clause jurisprudence has overlooked the plain meaning of the Sixth Amendment by conflating the word "witness" with an out-of-court declarant under hearsay rules).

(46.) See Randolph N. Jonakait, The Origins of the Confrontation Clause: An Alternative History, 27 RUTGERS L.J. 77, 167-68 (1995) (attacking the Roberts Court's theory that the Confrontation Clause is chiefly concerned with producing reliable evidence).

(47.) Justice Scalia's Crawford opinion refers to the Roberts reliability test as "amorphous," 541 U.S. 36, 63 (2004), "unpredictab[le]," id., and "[v]ague," id. at 68. Justice Scalia also argues that the reliability test is contrary to the intent of the Framers. Id. at 61 ("Where testimonial statements are involved, we do not think the Framers meant to leave the Sixth Amendment's protection to the vagaries of the rules of evidence, much less to amorphous notions of 'reliability.'"). Finally, Justice Scalia argues that because of its inherent vagueness, Roberts's reliability test fails to provide "any meaningful protection" in politically charged cases. Id. at 68 (using the foil of Sir Walter Raleigh's trial).

(48.) Although the majority does not state explicitly that it overrules Roberts, Chief Justice Rehnquist's concurrence states that the majority does just that. Id. at 69 (Rehnquist, C.J., concurring). Any doubt was surely put to rest in Whorton v. Bockting, where the Court explicitly stated three times that Crawford overruled Roberts. 549 U.S. 406, 413, 416, 419 (2007).

(49.) Crawford, 541 U.S. at 38-41.

(50.) Id. at 38-39.

(51.) Id. at 39-40.

(52.) Id. at 40.

(53.) State v. Crawford, 54 P.3d 656, 658 (Wash. 2002), rev'd and remanded, 541 U.S. 36.

(54.) Id.; see Crawford, 541 U.S. at 40.

(55.) Crawford, 541 U.S. at 41.

(56.) Id. at 41-42.

(57.) Id. at 67.

(58.) See id. at 50-56 (discussing the principles that emerge from the test and history of the Confrontation Clause, the concerns of the Framers, and the case law at the time of the Founding).

(59.) Id. at 43-47.

(60.) Id. at 47-50.

(61.) Id. at 45-50.

(62.) Id. at 48.

(63.) See id. at 50 ("It was these practices that the Crown deployed in notorious treason cases like Raleigh's ... and that the founding-era rhetoric decried."); infra notes 256-259.

(64.) Crawford, 541 U.S. at 44.

(65.) Id.

(66.) Id.

(67.) Id.; see also FISHER., supra note 39, at 360-62 (quoting various transcripts of Raleigh's trial).

(68.) See Erwin N. Griswold, The Due Process Revolution and Confrontation, 119 U. PA. L. REV. 711, 712 (1971) (noting that the Confrontation Clause "is said to find its historical origin in the trial of Sir Walter Raleigh"); Todd H. Neuman, Note, A Child's Well Being v. A Defendant's Right to Confrontation, 93 W. VA. L. REV. 1061, 1072 (1991) ("[M]any legal historians cite the 1603 treason trial of Sir Walter Raleigh as the catalytic influence in the development of the right to confront one's accusers."). But see Kenneth W. Graham Jr., The Right of Confrontation and the Hearsay Rule: Sir Walter Raleigh Loses Another One, 8 CRIM. L. BULL. 99, 100 n.4 (1972) ("My research gives me no reason to suppose that [the link between Raleigh's trial and the Sixth Amendment] represents anything other than a convenient but highly romantic myth, and I adhere to it for this reason.").

(69.) See Margaret A. Berger, The Deconstitutionalization of the Confrontation Clause: A Proposal for a Prosecutorial Restraint Model, 76 MINN. L. REV. 557 (1992).

(70.) Crawford, 541 U.S. at 50.

(71.) Id. at 50-53.

(72.) Id. at 51.

(73.) Id.

(74.) U.S. CONST. amend. VI.

(75.) Crawford, 541 U.S. at 51 (quoting 2 NOAH WEBSTER, AN AMERICAN DICTIONARY OF THE ENGLISH LANGUAGE (1828)).

(76.) Id.

(77.) Id. at 51-52. As I use the term, a "definition" is a dispositive test. By applying the definition to any particular statement, one can determine whether the statement is or is not testimonial. By contrast, I refer to other important considerations as "factors." Factors play an important but not outcome-determinative role in establishing whether a statement is testimonial. For a similar dichotomy, see People v. Rawlins, 884 N.E.2d 1019, 1026 (N.Y. 2008), which distinguishes "formulations," which I refer to as "definitions," from the "additional clues," which I refer to as "factors."

(78.) Crawford, 541 U.S. at 51 (citation omitted).

(79.) Id. at 52-52 (quoting White v. Illinois, 502 U.S. 346, 365 (1992) (Thomas, J., concurring)). This definition was first proposed by Justice Thomas's concurrence in White. The Court seems implicitly to have rejected a literal reading of this definition in later cases. See Melendez-Diaz v. Massachusetts, 557 U.S. 305, 329-30 (2009) (Thomas, J., concurring) (describing how the majority approach inappropriately broadens the reach of the Confrontation Clause by not confining its scope to formal statements); Davis v. Washington, 547 U.S. 813, 830 n.5 (2006) (rejecting a formality-based definition of testimonial as easily circumvented).

(80.) Crawford, 541 U.S. at 52.

(81.) Id. at 61.

(82.) Id. at 69.

(83.) See, e.g., United States v. Hadley, 431 F.3d 484, 500 n.11 (6th Cir. 2005) (referring to the Supreme Court's third definition of testimonial as the "broadest"); State v. Mizenko, 127 P.3d 458,466 (Mont. 2006) (describing the objective witness definition as "the broadest of the extant formulations acknowledged by Crawford").

(84.) See Crawford, 541 U.S. at 52-53, 68.

(85.) Id. at 68.

(86.) Many interviews with child witnesses are formal and akin to the civil ex parte examination and are thus testimonial under Definition Two; these interviews are nontestimonial under Definition Three if and only if the objective witness used for comparison is an objective witness possessing the age and mental capacity of a child witness. Compare People v. Vigil, 127 P.3d 916, 925 (Colo. 2006) (en banc) ("If the Crawford Court had intended the objective witness test to be applied from the perspective of an objectively reasonable observer educated in the law, the Crawford Court would have labeled the co-conspirator's statement 'testimonial.' However, by labeling the statement non-testimonial, Crawford directs us to apply the objective witness test from the perspective of an objectively reasonable person in the declarant's position." (footnotes omitted)), with State v. Snowden, 867 A.2d 314, 328-29 (Md. 2005) ("[W]e are unwilling to conclude that, as a matter of law, young children's statements cannot possess the same testimonial nature as those of other, more clearly competent declarants.").

(87.) See Miller v. Stovall, 608 F.3d 913, 923-26 (6th Cir. 2010) (holding that a suicide note by a former police officer is testimonial under the third Crawford formulation, but declining to endorse a lower-court holding that the note was testimonial under the second Crawford formulation).

(88.) See, e.g., United States v. Feliz, 467 F.3d 227, 233-36 (2d Cir. 2006) (holding that although "Crawford at least suggests that the determinative factor in determining whether a declarant bears testimony is the declarant's awareness or expectation that his or her statements may later be used at a trial," autopsy reports do not qualify as testimonial because "this statement ... should [not] be read to have adopted such an expansive definition of testimonial" (quoting United States v. Saget, 377 F.3d 223, 228 (2d Cir. 2004))).

(89.) See Crawford, 541 U.S. at 51; State v. Mizenko, 127 P.3d 458, 467-68 (Mont. 2006) (citing numerous state court cases to support the proposition that statements made to a "friend, family member or acquaintance" are nontestimonial "unless the declarant had clear reason to believe that they will be used prosecutorially").

(90.) The Crawford majority acknowledged that its "refusal to articulate a comprehensive definition in this case will cause interim uncertainty." 541 U.S. at 68 n.10.

(91.) See, e.g., Triplett v. Hudson, No. 3:09-CV-01281, 2011 WL 976575, at *6 (N.D. Ohio Mar. 17, 2011) ("[T]he extent to which Crawford applied to forensic laboratory reports created great confusion among lower federal and state courts and remained an unresolved issue until Melendez-Diaz."); State v. Mason, 126 P.3d 34, 39 (Wash. Ct. App. 2005) ("Crawford addresses statements made to government officials during examinations or interrogations initiated by those officials. As soon as the focus moves to disputed out-of-court statements voluntarily made by the witness during witness-initiated contact, confusion arises." (footnote omitted)); see also sources cited supra notes 11-18 (critiquing the Crawford line as vague, unclear, and inconsistent in its application).

(92.) See Davis v. Washington, 547 U.S. 813, 822 (2006) ("[T]hese cases require us to determine more precisely which police interrogations produce testimony.").

(93.) Id. at 818.

(94.) Id. at 818-19.

(95.) Id.

(96.) Id. at 819.

(97.) Id.

(98.) Id.

(99.) Id. at 819-20.

(100.) Id. at 820.

(101.) Id.

(102.) Id. at 821.

(103.) Id. at 820-22, 828.

(104.) Id. at 834.

(105.) Id. at 822.

(106.) Id. at 823.

(107.) See id. at 823-26 ("Moreover, as we have just described, the facts of [Crawford] spared us the need to define what we meant by 'interrogations.' The Davis case today does not permit us this luxury of indecision.... When we said in Crawford, that 'interrogations by law enforcement officers fall squarely within [the] class' of testimonial hearsay, we had

(108.) See id. at 828.

(109.) See id. at 824-25.

(110.) Id. at 828 (quoting United States v. Inadi, 475 U.S. 387, 394 (1986)).

(111.) Myrna S. Raeder, Confrontation Clause Analysis After Davis, 22 CRIM. JUST., Spring 2007, at 10, 12 ("I do not believe that Davis has totally displaced Crawford because Davis limited itself to police interrogation while Crawford's dicta reached statements made in judicial contexts, and both cases left unresolved whether statements to private individuals can be testimonial, and if so by what criteria.").

(112.) See, e.g., State v. Gilfillan, No. 08AP-317, 2009 WL 638264, at *16 (Ohio Ct. App. Mar. 12, 2009) ("The 'primary purpose' test applies to determining whether statements in police interrogations are testimonial. The 'objective witness' test applies to determining whether statements in non-police interrogations are testimonial.").

(113.) See Crawford v. Washington, 541 U.S. 36, 52 (2004).

(114.) See Davis, 547 U.S. at 822.

(115.) Tom Lininger, Reconceptualizing Confrontation After Davis, 85 TEX. L. REV. 271, 280 (2006) [hereinafter Lininger, Reconceptualizing] ; see also Tom Lininger, Davis and Hammon: A Step Forward, or a Step Back? 105 MICH. L. REV. FIRST IMPRESSIONS 28, 30 (2006) [hereinafter Lininger, Forward] ("The focus in Davis on police motives is theoretically inconsistent with Crawford...."); Ross, supra note 25, at 404-05 ("Davis v. Washington must be understood as a retreat from the principles laid out in Crawford.... It makes little sense that a person's opportunity to cross-examine a witness rests not on whether the out-of-court statement serves to accuse the defendant at trial, but on the police officer's reason for gathering the statement in the first place.").

(116.) See Davis, 547 U.S. at 826.

(117.) See Ross, supra note 25, at 411 ("By using the word 'core' to identify a perimeter, the Davis Court collapsed the broad possibilities of the term core in Crawford.... [T]he core became the circumference or perimeter of the scope of the clause."). Melendez-Diaz v. Massachusetts, 557 U.S. 305 (2009), collapses this distinction more explicitly. Whereas Crawford described its three definitions as formulations of the "core" class of testimonial statements covered by the Confrontation Clause, Melendez-Diaz refered to those same three definitions as simply "the class of testimonial statements covered by the Confrontation Clause." Id. at 309. Since the Confrontation Clause applies only to testimonial hearsay, the class of testimonial statements covered by the Confrontation Clause equals the entire class of statements covered by the Confrontation Clause.

(118.) See supra note 84.

(119.) See supra note 85.

(120.) See Crawford v. Commonwealth, 686 S.E.2d 557, 567-69 (Va. Ct. App. 2009) (holding that "[a]lthough the Supreme Court of the United States recently stated that affidavits fall within the 'core class of testimonial statements' subject to the Confrontation Clause," an affidavit used "to obtain a civil, preliminary protective order" was not testimonial because it was made in anticipation of a civil action, and not a criminal trial (citations omitted)). This case seems anomalous and may well have been decided differently in the wake of Bullcoming v. New Mexico, 131 S. Ct. 2705 (2011).

(121.) See, e.g., United States v. Yeley-Davis, 632 F.3d 673, 680-81 (10th Cir.), cert. denied, 131 S. Ct. 2172 (2011) (holding that certificates of authenticity and accompanying affidavits for cell phone records are nontestimonial).

(122.) See, e.g., Holz v. State, No. 06-09-00172-CR, 2010 WL 1041068, at *1-2 (Tex. App. Mar. 23, 2010) (holding that veterinary reports were nontestimonial under the Davis primary purpose test even though the authors of the reports "could have reasonably believed that [their veterinary reports] would be available at a later trial").

(123.) Cf. Richard D. Friedman, Crawford, Davis, and Way Beyond, 15 J.L. & POL'Y 553, 562 (2007) ("When we speak of the anticipation of a reasonable person in the declarant's position, we are referring to a hypothetical person who has all the information about the particular situation that the declarant does, and no more. Thus, if the declarant is speaking to an undercover police officer, the hypothetical person would not know that her audience is collecting information for use in prosecution.").

(124.) See supra note 89.

(125.) See Lininger, Reconceptualizing, supra note 115, at 274.

(126.) See, e.g., id. at 284-85 (arguing that Davis reduced the value of the excited utterance exception, encouraged batterers to murder their victims, provided perverse incentives for police to prolong ongoing emergencies, and incentivized police to lock up victims using material witness warrants).

(127.) See Ross, supra note 25, at 404-05 ("Davis v. Washington must be understood as a retreat from the principles laid out in Crawford.... [The "primary purpose" test in Davis] means that even if one person accuses another person of a crime, that initial accusation may serve as the basis to bring charges, hold the defendant before trial, and convict the defendant, if the police were engaged in resolving an emergency at the time the accusation was made.").

(128.) See Lininger, Forward, supra note 115, at 29 ("[P]olice will likely try to adapt their practices so that they can accomplish the same goals by simply incanting the right rationale for their actions. For example, police will probably be much more careful in their reports to list circumstances supporting an inference of 'ongoing emergency' at the time the police question hearsay declarants."). Professor Lininger concludes that this concern is overblown because the circumstances objectively must indicate an ongoing emergency. Id. at 29. Given the Court's expansive interpretation of ongoing emergencies in Bryant, see infra notes 164-166 and accompanying text, Professor Lininger's optimism might have been premature.

(129.) I will not discuss Giles v. California, 554 U.S. 353 (2008), or Bullcoming v. New Mexico, 131 S. Ct. 2705 (2011). The former considers the relationship between the Confrontation Clause and the "forfeiture by wrongdoing" doctrine, while the latter addresses the particular witness whom a defendant has a right to confront under Melendez-Diaz. Though interesting, these cases are not directly relevant to the central questions of this Note.

(130.) 557 U.S. 305 (2009).

(131.) 131 S. Ct. 1143 (2011).

(132.) 132 S. Ct. 2221 (2012).

(133.) 557 U.S. at 308-09.

(134.) Id. at 308.

(135.) Id. at 309.

(136.) Id.

(137.) Id. at 329.

(138.) Id. at 310 ("The documents at issue here, while denominated by Massachusetts law 'certificates,' are quite plainly affidavits.... ").

(139.) Id. at 329.

(140.) Id.

(141.) The dissent in Melendez-Diaz emphasized this point. Id. at 343-44 (Kennedy, J., dissenting) ("Laboratory analysts are not 'witnesses against' the defendant as those words would have been understood at the framing.... [T]he [Confrontation] Clause refers to a conventional 'witness'--meaning one who witnesses ... an event that gives him or her personal knowledge of some aspect of the defendant's guilt."). The majority argued that there is no constitutional basis for distinguishing between conventional and nonconventional witnesses. Id. at 315-16 (majority opinion). Even if the majority is ultimately correct that the Confrontation Clause ought to apply to laboratory analysts, I cannot agree that this conclusion is especially obvious. Crawford's reasoning is conspicuously originalist. See Laird C. Kirkpatrick, Nontestimonial Hearsay After Crawford, Davis and Bockting, 19 REGENT U. L. REV. 367, 382 (2007) (characterizing the Crawford opinion as a "model of originalist interpretation"). Thus, it makes good sense to carefully consider how the Framers would have regarded the hearsay testimony of forensic analysts.

(142.) See Triplett v. Hudson, No. 3:09-GV-01281, 2011 WL 976575, at *6 (N.D. Ohio Mar. 17, 2011) ("Indeed, the courts who addressed this question generally found that the Crawford rule did not apply to these laboratory reports.... Many other courts addressing the issue also held that Crawford was not violated by the admission of laboratory reports where the technician who prepared the report was not available for cross examination."); Commonwealth v. Williams, 69 Va. Cir. 277, 280 (2005) ("This Court follows the majority of state courts and the direction of the Court of Appeals of Virginia and finds that the certificate of analysis is not testimonial."). But see Garcia v. Roden, 672 F. Supp. 2d 198, 208 (D. Mass. 2009) (noting that Melendez-Diaz was "clearly foreshadowed" by Crawford).

(143.) See 557 U.S. at 310 (quoting in full the three Crawford definitions).

(144.) Id. at 310-11.

(145.) See supra notes 138-139 and accompanying text.

(146.) Melendez-Diaz, 557 U.S. at 316.

(147.) Id. at 312 n.2 ("[M]edical reports created for treatment purposes ... would not be testimonial under our decision today.").

(148.) Id. at 324 ("Business and public records are generally admissible absent confrontation not because they qualify under an exception to the hearsay rules, but because--having been created for the administration of an entity's affairs and not for the purpose of establishing or proving some fact at trial--they are not testimonial.").

(149.) Michigan v. Bryant, 131 S. Ct. 1143, 1150 (2011).

(150.) Id.

(151.) Id.

(152.) Id. at 1150-51.

(153.) Id. Covington's statements were not introduced as dying declarations because the prosecution failed to establish the required factual predicates. Id. at 1151 n.1.

(154.) Id. at 1151.

(155.) Id. at 1166-67.

(156.) Id. at 1163-66 (explaining that an ongoing emergency existed because the officers knew a potential shooter with a gun was wandering the area and neither the police nor Covington knew the shooter's location).

(157.) See Langley v. State, 28 A.3d 646, 652 n.3 (Md. 2011) ("[A]fter Bryant, the 'ongoing emergency' concept is no longer construed so narrowly.").

(158.) Bryant, 131 S. Ct. at 1154.

(159.) Id. at 1155.

(160.) Id. at 1168-69 (Scalia, J., dissenting) ("[B]ecause the Court picks a perspective so will I: The declarant's intent is what counts.... A declarant-focused inquiry is also the only inquiry that would work in every fact pattern implicating the Confrontation Clause.").

(161.) See supra notes 115, 125-127 and accompanying text.

(162.) See Bryant, 131 S. Ct. at 1170 (Scalia, J., dissenting) ("The only virtue of the Court's approach ... is that it leaves judges free to reach the 'fairest' result under the totality of the circumstances.... Unfortunately, under this malleable approach 'the guarantee of confrontation is no guarantee at all.'" (quoting Giles v. California, 541 U.S. 353, 375 (2008) (plurality opinion)).

(163.) See Cicchini, supra note 7; Kenneth W. Graham, Jr., Michigan v. Bryant: The Counter-Revolution Begins 1 (UCLA Sch. of Law Pub. Law & Legal Theory Research Paper Series, Paper No. 11-07, 2011), ("[T]he majority of the [Bryant] [C]ourt, over a bitter dissent from Justice Scalia, effectively over-ruled Crawford and pushed confrontation doctrine back in the direction of Roberts.").

(164.) Bryant, 131 S. Ct. at 1158 ("[The Michigan Supreme Court] employed an unduly narrow understanding of 'ongoing emergency'.... An assessment of whether an emergency that threatens the police and public is ongoing cannot narrowly focus on whether the threat solely to the first victim has been neutralized because the threat to the first responders and public may continue.").

(165.) Id. at 1158.

(166.) See Richard D. Friedman, Preliminary Thoughts on the Bryant Decision, CONFRONTATION BLOG (Mar. 2, 2011, 12:42 AM), /preliminary-thoughts-on-bryant-decision.html ("So one of my concerns is that police officers will quickly learn that they can get statements characterized as non-testimonial if they testify, in effect, 'I came up to the scene and didn't know what was happening. My principal concern was securing the public safety.'").

(167.) Bryant, 131 S. Ct. at 1155 ("[T]he most important instances in which the Clause restricts the introduction of out-of-court statements are those in which state actors are involved in a formal, out-of-court interrogation of a witness to obtain evidence for trial.").

(168.) Id. at 1160-62 ("In addition to the circumstances in which an encounter occurs, the statements and actions of both the declarant and interrogators provide objective evidence of the primary purpose of the interrogation.... Taking into account a victim's injuries does not transform this objective inquiry into a subjective one. The inquiry is still objective because it focuses on the understanding and purpose of a reasonable victim in the circumstances of the actual victim--circumstances that prominently include the victim's physical state.").

(169.) Id. at 1160.

(170.) Id.

(171.) 132 S. Ct. 2221 (2012).

(172.) Id. at 2227.

(173.) Id. at 2229-30.

(174.) Id. at 2231.

(175.) Id. at 2238-41.

(176.) Id. at 2242-44.

(177.) Id. at 2242.

(178.) Id. at 2243-44.

(179.) Id. at 2255 (Thomas, J., concurring in the judgment).

(180.) Id. at 2259-60.

(181.) Id. at 2260.

(182.) See Richard D. Friedman, Thoughts on Williams, Part h Reasons To Think the Impact May Be Limited, CONFRONTATION BLOG (June 19, 2012, 7:52 AM), http://confrontationright

(183.) See id.

(184.) See id. ("The case appears to stand for nothing more tha[n] the proposition that in the circumstances of this case there is no Confrontation Clause violation.").

(185.) Cf. Gregg v. Georgia, 428 U.S. 153, 169 n.15 (1976) ("[T]he holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds...."). Justice Thomas's opinion, which decided that the statements in Williams were nontestimonial because they were not formalized, is arguably controlling because it is narrower than the majority's version of the primary purpose test. See Jeffrey Fisher, The Holdings and Implications of Williams v. Illinois, SCOTUSBLOG (June 20, 2012, 2:20 PM),

(186.) See supra Section I.B (explaining the relative breadth of the "objective witness" formulation).

(187.) See supra notes 20-22.

(188.) Specifically, I keycited all state court, federal district court, and federal appellate court opinions citing Davis v. Washington through October 17, 2011. This yielded a population of 2,095 cases. I then randomly generated three hundred numbers between 1 and 2,095, arranged the cases in chronological order, and read the cases with those three hundred numbers.

(189.) See, e.g., Cicchini, supra note 7; Cicchini, supra note 12; Ross, supra note 25.

(190.) Kay L. Levine, The Law Is Not the Case: Incorporating Empirical Methods into the Culture of Case Analysis, 17 U. FLA. J.L. & PUB. POL'Y 283, 300 (2006).

(191.) Id.

(192.) Cf. Russell V. Lenth, Some Practical Guidelines for Effective Sample Size Determination, 55 AM. STATISTICIAN 187, 187 (2001) ("[T]he study must be of adequate size, relative to the goals of the study. It must be 'big enough' that an effect of such magnitude as to be of scientific significance will also be statistically significant.").

(193.) See Mark A. Hall & Ronald F. Wright, Systematic Content Analysis of Judicial Opinions, 96 CALIF. L. REV. 63, 79-80 (2008) ("At the most basic level, empirically minded legal scholars can be more systematic in selecting cases for analysis ... by specifying a reproducible selection of cases ... such as reading every tenth case ... or every federal appellate products liability case over three years.... Disciplined research ... guards against subliminal biases in selecting only cases that prove the author's point.").

(194.) This is a specific manifestation of the scientific requirement of a prior research design. See KIMBERLY A. NEUENDORF, THE CONTENT ANALYSIS GUIDEBOOK 11 (2002) ("Too often, a so-called content analysis report describes a study in which variables were chosen and 'measured' after the messages were observed. This wholly inductive approach violates the guidelines of scientific endeavor. All decisions on variables, their measurement, and coding rules must be made before the observations begin.").

(195.) See Hall & Wright, supra note 193, at 107-09 (explaining the coding process).

(196.) Id. at 80-81 (explaining how systematic empirical analysis disciplines the reader and hedges against confirmation bias in reading) (citing Charles M. Haar et al., Computer Power and Legal Reasoning: A Case Study of Judicial Decision Prediction in Zoning Amendment Cases, 2 AM. B. FOUND. RES. J. 651, 746 (1977)).

(197.) I chose Davis rather than Crawford because Davis arguably began to stabilize the Court's Confrontation Clause doctrine. Moreover, Davis introduced explicitly the importance of an ongoing emergency--a fact that plays a significant role in recent Confrontation Clause cases.

(198.) This figure refers to total statements. For the remaining descriptive statistics and regression analysis, I excluded statements concerning forensic analyses. Out of 255 total statements not concerning forensic analyses, 81 were testimonial, a difference of less than one percentage point compared to the sample of all statements.

(199.) See, e.g., Taylor v. State, 248 P.3d 362-, 375 (Okla. Crim. App. 2011) (holding that the trial court's error in admitting testimonial statements against the defendant was harmless). The flip side is that nearly all the cases evaluated were cases on appeal where the government had prevailed below. So the one-third figure, while it overestimates defendant success on appeal, may underestimate the overall success defendants have in mounting Confrontation Clause challenges at any stage of a criminal case.

(200.) Hartsfield v. Commonwealth, 277 S.W.3d 239, 244 (Ky. 2009).

(201.) Davis v. Washington, 547 U.S. 813, 822 (2006).

(202.) Michigan v. Bryant, 131 S. Ct. 1143, 1161-62 (2011).

(203.) Out of 60 statements made near a slightly injured party, 37 (or about 62%) were made to state actors. Out of 176 statements made with no injured party nearby, 84 (or about 48%) were made to state actors. This comparison excludes statements made to 911 operators.

(204.) See DAVID R. ANDERSON ET AL,, STATISTICS FOR BUSINESS AND ECONOMICS 562 (11th ed. 2010) ("[R]egression analysis can be used to develop an equation showing how [two or more] variables are related." (emphasis omitted)); Ewout W. Steyerberg et al., Stepwise Selection in Small Data Sets: A Simulation Study of Bias in Logistic Regression Analysis, 52 J. CLINICAL EPIDEMIOLOGY 935, 935 (1999) ("Multivariable regression analysis is a valuable technique to quantify the relation between two or more covariables and an outcome variable.").

(205.) For an explanation of how logistic regression differs from linear regression, see generally FRED C. PAMPEL, LOGISTIC REGRESSION: A PRIMER ch. 1 (2000).

(206.) See NEUENDORF, supra note 194, at 169, 170-71 box 8.1.

(207.) The general rule for logistic regression analysis is that one wants ten or more events per variable. See, e.g., Peter Peduzzi et al., A Simulation of the Number of Events per Variable in Logistic Regression Analysis, 49 J. CLINICAL EP1DEMIOLOGY 1373, 1373 (1996) (recommending at least ten events per variable); Chao-Ying Joanne Peng et al., An Introduction to Logistic Regression Analysis and Reporting, 96 J. EDUC. RES. 3, 10 (2002) (same). Events are defined as the less likely dependent variable category. In my sample, eighty-one statements were held to be testimonial--fewer than were held to be nontestimonial. With 10 variables, I had 8.1 events per variable--less than ideal, but not fatal or even severely damaging to the results. See Eric Vittinghoff & Charles E. McCulloch, Relaxing the Rule of Ten Events per Variable in Logistic and Cox Regression, 165 AM. J. EPIDEMIOLOGY 710, 717 (2007) ("[S]ystematic discounting of results, in particular statistically significant associations, from any model with 5-9 EPV [events per variable] does not appear to be justified."). To the extent that low sample size matters, it increases the variance of the estimated coefficients and makes a Type II error more probable. Id. The danger of low sample size is usually a false negative and not a false positive.

(208.) The nonsignificant variables are those for the following: statement made before the crime, opinion in a habeas petition, statement was a present sense impression, statement was a dying declaration, statement was made by the crime victim, and statement was made in a government building. I excluded all forensic analyses from the sample, so this variable could not be significant.

(209.) The right-hand column shows how a change in each independent variable changes the natural logarithm of the event that a statement is held to be testimonial (the dependent variable). This effect is constant, regardless of the values of other independent variables, whereas the percentage values in the middle column have a clearer practical meaning but can vary depending on the values of different independent variables.

(210.) See infra Section III.B.

(211.) Id.

(212.) Michigan v. Bryant, 131 S. Ct. 1143, 1155 (2011) ("[T]he most important instances in which the Clause restricts the introduction of out-of-court statements are those in which state actors are involved in a formal, out-of-court interrogation of a witness to obtain evidence for trial.").

(213.) No. E2008-02541-CCA-R3-CD, 2010 WL 3706090 (Tenn. Crim. App. Sept. 22, 2010), aff'd in relevant part, 350 S.W.3d 883 (Tenn. 2011).

(214.) Id. at *1-2.

(215.) Id.

(216.) Id. at *2, *16.

(217.) Id. at *5.

(218.) Id. at *15.

(219.) Id. at *15-18.

(220.) See Davis v. Washington, 547 U.S. 813, 823 n.2 (2006) (noting that 911 operators are agents of law enforcement when they are questioning 911 callers).

(221.) See, e.g., State v. Camarena, 176 P.3d 380, 387-88 (Or. 2008) (holding that the first half of a 911 call was nontestimonial but the second half of the call was testimonial because the questions and responses "were directed at establishing facts only relevant to a subsequent criminal action").

(222.) See, e.g., State v. Bentley, 739 N.W.2d 296 (Iowa 2007) (describing in detail how interviews by the Child Protection Center are testimonial because they are formal and intertwined with the evidence-gathering functions of the police).

(223.) See supra note 200 and accompanying text.

(224.) No. 91571, 2009 WL 2894472 (Ohio Ct. App. Sept. 10, 2009).

(225.) Id. at *1.

(226.) Id.

(227.) Id.

(228.) Id.

(229.) Id.

(230.) Id. at *2.

(231.) Id. at *4.

(232.) Id. at *5, *7.

(233.) Id. at *5-6.

(234.) See supra note 207.

(235.) See ARTHUR S. GOLDBERGER, A COURSE IN ECONOMETRICS 248-51 (1991) (noting that the consequences of a small sample size are exactly as serious and have the same effect--increased variance of the parameter estimates- as multicollinearity).

(236.) Id. at 248-52.

(237.) See Robert M. O'Brien, A Caution Regarding Rules of Thumb for Variance Inflation Factors, 41 QUALITY & QUANTITY 673, 674 & n.2 (2007) ("Not uncommonly a VIF of lo or even one as low as 4 ... have been used as rules of thumb to indicate excessive or serious multicollinearity."); see also JAMES P. STEVENS, INTERMEDIATE STATISTICS: A MODERN APPROACH 235 (3d ed. 2007) (discussing the rule often for VIF).

(238.) See James J. Heckman, Sample Selection Bias as a Specification Error, 47 ECONOMETKICA 153, 153 (1979) (modeling sample selection bias).

(239.) See supra Section I.C.

(240.) For a particularly frustrating case requiring that I determine how to code a veterinarian, see Holz v. State, No. 06-09-00172-CR, 2010 WL 1041068 (Tex. Crim. App. Mar. 23, 2010).

(241.) See NEUENDORF, supra note 194, at 143.

(242.) See sources cited supra notes 11-18, 25; see also Isley Markman, The Admission of Hearsay Testimony Under the Doctrine of Forfeiture-by-Wrongdoing in Domestic Violence Cases: Advice for Prosecutors and Courts, CRIM. L. BRIEF, Spring 2011, at 9 ("The Crawford decision failed to provide a clear definition of 'testimonial.'").

(243.) See United States v. Brito, 427 F.3d 53, 55 (1st Cir. 2005) ("Crawford ... produced a miasma of uncertainty.").

(244.) See Fowler v. State, 809 N.E.2d 960, 966 (Ind. Ct. App. 2004) (Crone, J., concurring in the result) (noting that "fallout from Justice Scalia's 'clarification' of the Confrontation Clause in Crawford will reverberate through the evidentiary landscape for some time to come and will create countless dilemmas for trial and appellate courts"), vacated, 829 N.E.2d 459 (Ind. 2005); Markman, supra note 242, at 12 ("[Lower] courts read the vague Davis standard to admit a significant amount of hearsay evidence as non-testimonial and therefore as unaffected by Crawford.").

(245.) See Cicchini, supra note 7, at 1320 (noting that "lower courts continued to abuse their discretion" in the wake of Davis); Cicchini, supra note 12, at 754-55 (arguing for a "bright-line" Confrontation Clause test as the "only viable solution to the problem of judicial manipulation" by judges who do not "honestly attempt to uphold the Constitution"); Ross, supra note 25, at 449-57 (describing how lower courts "strain[] the leash" by pushing the limits of ambiguous Confrontation Clause doctrine).

(246.) I say "effectively engage" because most courts are not following these steps explicitly. Rather, courts are resolving Confrontation Clause challenges as if they had followed this process. An analogy from welfare economics might be familiar to some readers. Economists generally recognize that consumers do not consciously think about maximizing utility when they make consumption choices, but they may nonetheless behave as though they were rational utility-maximizers. WALTER NICHOLSON & CHRISTOPHER SNYDER, MICROECONOMIC THEORY: BASIC PRINCIPLES AND EXTENSIONS 92 (8th ed. 2002) ("[T]he utility-maximization model predicts many aspects of behavior even though no one carries around a computer with his or her utility function programmed into it. To be precise, economists assume that people behave as if they made such calculations ....").

(247.) See Crawford v. Washington, 541 U.S. 36, 51 (2004) (contrasting testimonial, formal statements to government officials with nontestimonial, casual remarks to a friend).

(248.) See Michigan v. Bryant, 131 S. Ct. 1143, 1160-62 (2011) (describing the role of injury in determining whether a statement is testimonial).

(249.) See supra notes 242-245.

(250.) See, e.g., People v. Blacksher, 259 P.3d 370, 410-11 (Cal. 2011) (holding that a statement "made while [the declarant] was in a police car, in the presence of a city mental health worker, and with multiple officers nearby" was nontestimonial); Crawford v. Commonwealth, 686 S.E.2d 557, 567-69 (Va. Ct. App. 2009) (holding that a civil protective order affidavit was nontestimonial) ; Ross, supra note 25, at 443-49 (citing cases where courts have selectively emphasized statements in Crawford and Davis to minimize Confrontation Clause protections).

(251.) Levine, supra note 190, at 300 (arguing that although theories from a "small series of cases" can be "thought-provoking," they rarely provide a "robust explanation of how the law works").

(252.) See Michael Heise, The Importance of Being Empirical, 26 PEPP. L. REV. 807, 813-15 (1999) (arguing that empirical research allows scholars to "better maintain neutrality," leading to more accurate descriptions of the legal system and sounder normative theories); cf. Hall & Wright, supra note 193, at 79-80 (explaining how the approach of systematically reading a representative sample of cases differs from the conventional "leading case" approach).

(253.) Raeder, supra note 16, at 775-76.

(254.) Etter, supra note 13, at 1168.

(255.) See supra Section III.A.

(256.) See Berger, supra note 69; see also Akhil Reed Amar, Confrontation Clause First Principles: A Reply to Professor Friedman, 86 GEO. L.J. 1045, 1048 (1998) ("[T]he Constitution is mainly addressed to state action .... The Sixth Amendment is triggered when [a defendant] is 'accused' by the state ....").

(257.) Berger, supra note 69, at 586 ("Confrontation was part of an arsenal designed ... to restrain the government in criminal trials from acting in a covert, repugnant manner that would be concealed from the people.").

(258.) Id. at 568-81 (describing the English antecedents and roots of the Confrontation Clause and the way that history influenced the Framers).

(259.) Id. at 560-63.

(260.) U.S. CONST. amend. VI.

(261.) See Crawford v. Washington, 541 U.S. 36, 43-56 (2004).

(262.) United States v. Brito, 427 F.3d 53, 68 (1st Cir. 2005) (Howard, J., concurring in part and concurring in the judgment) (citing Crawford, 541 U.S. at 51-52).

(263.) Id.; see also People v. Hurtado, No. F047195, 2006 WL 1364999, at *4-5 (Cal. Ct. App. May 19, 2006) (stating that Crawford's three definitions should be read not as distinct but as sharing a common goal to exclude civil ex parte examinations).

(264.) See Miranda v. Arizona, 384 U.S. 436, 469 (1966) ("[A]nything said can and will be used against the individual in court.").

(265.) See, e.g., Richard D. Friedman, Grappling with the Meaning of "Testimonial," 71 BROOK. L. REV. 241, 273-74 (2005).

(266.) See, e.g., Josephine Ross, What's Reliability Got To Do with the Confrontation Clause After Crawford?, 14 WIDENER L. REV. 383, 384-85, 403-05 (2009).

(267.) See, e.g., Cicchini, supra note 7, at 1316-20.

(268.) Cf. Crawford v. Washington, 541 U.S. 36, 61, 67 (2004) ("[The Confrontation Clause] commands, not that evidence be reliable, but that reliability be assessed in a particular manner: by testing in the crucible of cross-examination.... The Constitution prescribes a procedure for determining the reliability of testimony in criminal trials, and we, no less than the state courts, lack authority to replace it with one of our own devising."). I quote Justice Scalia's point extensively because it is a fitting and decisively less pro-defense statement than scholars assume. When lower courts admit evidence that scholars perceive to be unreliable, they often argue that these courts are circumventing Crawford. But Justice Scalia's language shows the flaw in this reasoning. The Clause prevents the government from introducing testimonial statements without allowing cross-examination, but it also provides defendants with no guarantees that they will be able to cross-examine declarants whose statements were made in nontestimonial circumstances. This is true even when those statements are probably unreliable.

(269.) See Jacob Cohen, A Coefficient of Agreement for Nominal Scales, 20 EDUC. & PSYCHOL. MEASUREMENT 37, 40 (1960) (proposing K as a measure of "the proportion of agreement [between two coders of a nominal variable] after chance agreement is removed from consideration").

(270.) See NEUENDORF, supra note 194, at 150.

(271.) See id. at 143.

AUTHOR. Yale Law School, J.D. 2013; University of Michigan, B.A. 2008. Thanks to Professor Steven Duke for his suggestions and supervision during the drafting process, to The Yale Law Journal editorial staff (especially Chris Michel) for outstanding feedback and editing, and to Laura Pisarello for her constant support. Any errors are my own.
Table 1.

                           TESTIMONIAL UNDER      NONTESTIMONIAL UNDER
                           THE "OBJECTIVE         THE "OBJECTIVE
                           WITNESS" TEST          WITNESS" TEST

TESTIMONIAL UNDER THE      --Police               Videotaped
"FORMALITY" TEST           interrogations (84)    statements by child
                                                  witnesses (86)
                           --Prior grand jury
                           testimony (85)

NONTESTIMONIAL UNDER THE   --Suicide notes (87)   Statements to
"FORMALITY" TEST                                  friends prior to the
                           --Autopsy              crime (89)
                           reports (88)

Table 2.

                         "PRIMARY PURPOSE" TEST  THE "PRIMARY PURPOSE"

TESTIMONIAL UNDER THE    --Police                --Spouse's affidavit
"OBJECTIVE WITNESS"      interrogations (118)    for a preliminary
TEST                                             protective order
                         --Prior grand jury      (120)
                         testimony (119)
                                                 --Private business
                                                 certifications of
                                                 authenticity (121)

                                                 technician and
                                                 veterinarian report

NONTESTIMONIAL UNDER     Statements to           Statements to friends
THE "OBJECTIVE           undercover informants   prior to the crime
WITNESS" TEST            (123)                   (124)

Table 3.

                                                         NUMBER OF
                                                         STATEMENTS IN
VARIABLE      DESCRIPTION                                THE SAMPLE

Testimonial   Whether the statement was classified by    90
              the court as testimonial

Before        Whether the statement was made before      42
              the charged offenses

During        Whether the statement was made during      26
              the commission of the charged offenses

Cop           Whether the statement was made to a law    128
              enforcement officer other than a 911

Doc           Whether the statement was made to a        24
              medical health professional such as an
              EMT, doctor, or nurse

Gov           Whether the statement was made to a        14
              government official other than a law
              enforcement officer or 911 operator

Excited       Whether the statement was classified as    85
              an excited utterance

Present       Whether the statement was classified as    9
              a present sense impression

Dying         Whether the statement was classified as    8
              a dying declaration

Record        Whether the statement was classified as    35
              a business or government record

Medical       Whether the statement was made during      19
              medical treatment or contained in
              medical records

Abuse         Whether the statement was made during      9
              abuse counseling not performed by a
              medical professional (as with a social

Forensic      Whether the statement was contained in a   23
              forensic or laboratory analysis report

Call          Whether the statement was made during a    38
              911 call

Station       Whether the statement was made in a        16
              government facility such as a police
              station, courthouse, or city hall

Victim        Whether the declarant of the statement     120
Speaking      was the victim of the charged offenses

Injured       Whether the statement was made with a      74 (102 in
              significantly injured party nearby         alternative

Taped         Whether the statement was audio- or        67

Habeas        Whether the opinion was in a case on       36
              habeas review

Table 4.

VARIABLE                             CHANGE IN             LOGGED ODDS
                                     PROBABILITY THAT A    COEFFICIENT
                                     STATEMENT IS FOUND

Recipient is a nonpolice             +82.3%                5.258
government official ***              (0.051)               (1.315)

Recipient is a police officer ***    +75.1%                4.711
                                     (0.074)               (0.814)

Statement is made during abuse       +65.8%                3.158
counseling ***                       (0.175)               (1.233)

Statement is made during a 911       +40.7%                1.980
call **                              (0.191)               (0.874)

Recipient of the statement is a      +39.7%                1.881
health professional **               (0.214)               (0.938)

Statement is audio- or video-        +26.2%                -1.434
recorded                             (0.160)               (0.797)

Statement is made during the         -13.9%                -1.272
commission of the offense            (0.058)               (0.744)

Statement is made while a party      -18.2%                -1.437
is injured ***                       (0.051)               (0.443)

Statement is an excited              -21.1%                -1.628
utterance ***                        (0.055)               (433)

Statement is a business or           -21.2%                -3.874
government record ***                (0.043)               (1.390)

This table presents the results from a stepwise logistic regression.
Three variable selection processes-stepwise regression, forward
elimination, and backward elimination-all yielded the same set of
variables, which had a statistically significant correlation with a
statement being found testimonial at the .15 level or better. Standard
errors for each term are in parentheses below the term. All listed
variables are statistically significant at the .10 level. Variables
with ** are significant at the 05 level and variables with *** are
significant at the .01 level.

Figure 1.

                                         BE TESTIMONIAL

Police Officers                             53/106
Other Government Officials                   7/14
Health Professionals                         9/23
911 Operators                                8/38
Private Citizens                             4/74

Each of the listed categories is defined by the coded variables.
"Health Professionals" includes EMTs, nurses, and doctors but not
social workers. "Private citizens" is defined residually, as
recipients who are not police, other government officials, health
professionals, or 911 operators. They may include friends or
family members.

Note: Table made from bar graph.

Figure 2.


State Actors                                 60/120
Nonstate Actors                              13/97

"State Actors" includes statements made to police or other government
actors, not including statements to 911 operators. "Nonstate Actors"
includes any statement made to a recipient other than a police officer,
other government official, or 911 operator.

Note: Table made from bar graph.

Figure 3.


                       Injury Specification A    Injury Specification B

Injured Party            11/74                       27/102
Not Injured Party        70/181                      54/153

Specification A defined an injury as any shooting, stabbing, or
other injury requiring immediate medical assistance. This
necessarily involved subjective judgment, as explained above.
Specification B coded the injury variable as 1 if a party was
stabbed or shot, and 0.5 if a party was otherwise physically harmed
(for example, punched, kicked or strangled). In this Figure,
however, I included statements under "Injured Party" in
Specification B if the variable was coded either 1 or 0.5. To put
the matter simply, Specification B characterized more cases as
having injured parties, and many of those injuries were rather less

Note: Table made from bar graph.

Figure 4.


Severely Injured Party             4/42
Slightly Injured Party            23/60
No Injured Party                  54/153

"No Injured Party" refers to situations in which no party to the
events was injured in any way. "Slightly Injured Party" refers to
situations in which the injury was coded as 0.5 under Specification
B--in other words, where physical harm occurred but it was not a
shooting or stabbing. "Severely Injured Party" refers to situations
where the injury was coded as 1 under Specification B--a shooting
of stabbing.

Note: Table made from bar graph.
COPYRIGHT 2012 Yale University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:II. Empirical Evidence through Appendix B: Intercoder Reliability, with footnotes, p. 809-836
Author:Keenan, Dylan O.
Publication:Yale Law Journal
Date:Dec 1, 2012
Previous Article:Confronting Crawford v. Washington in the lower courts.
Next Article:Copyright protection in an opt-out world: implied license doctrine and news aggregators.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters