Here is one example: it is a case decided on June 11, 1999, by the United States District Court in Massachusetts.  The court explored the admissibility of handwriting identification testimony and the continuing debate of whether it qualifies as science, or whether it is something else. Also important were arguments concerning about the reliability and admissibility of expert testimony attacking the reliability of the Government’s eyewitness.

The case came about because Johanna Haines, who had been charged with robbery of a bank, sought to bar the admission of handwriting comparison testimony which the Government intended to produce, showing that he was the author of the handwriting on the robbery note. In the alternative, the defendant sought permission to have a critic of handwriting identification testimony testify to the weakness of handwriting analysis.  In addition, the defendant offered to have a psychologist testify on the unreliability of eyewitness identifications.

To give you a flavor of what the arguments were all about, here are significant excerpts--with most internal citations and all footnotes omitted--of the Memorandum and Order of Judge Gartner in United States v. Haines, 55 F.Supp. 2d 62 (D.C.Mass. 1999):

This case raises questions concerning the application of Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) and Kumho Tire Co. v. Carmichael, 526 U.S. 137, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999) to technical fields, that are not, strictly speaking, science. Two fields are involved:  The first is an old field, handwriting analysis, which has been the subject of expert testimony for countless years. The second is a comparatively new field, the psychology of eyewitness identification.  The government maintained that handwriting analysis is science, meeting the Daubert and Kumho tests, while the psychology of eyewitness identification is not.  Not surprisingly, the defendant insisted that the opposite is true.

Johannes Haines (Haines) is charged under 18 U.S.C. § 2113 for allegedly robbing the Broadway National Bank in Chelsea, Massachusetts on January 27, 1997. The government's principal evidence consisted of the eyewitness identification of the teller who was robbed, Ms. Jeanne Dunne, and the handwriting analysis of the robbery note. In connection with the latter, the government offered Diana Harrison (Harrison), a document examiner with the Federal Bureau of Investigations, to testify as to the authorship of a stick-up note found at the scene of the crime.

Haines sought to exclude the handwriting analysis.  This testimony, defense claims, notwithstanding its venerable history, does not meet the standards of Daubert and Kumho.  In the alternative, if the court permitted the jury to hear the handwriting testimony, Haines sought to have his expert--Professor Mark Denbeaux (Denbeaux)--testify as to the weaknesses of the methodology and the basis of Harrison's conclusions.  The government, on the other hand, argued for its handwriting expert under the applicable tests, and rejects Denbeaux.

Haines also offered the testimony of Dr. Saul Kassin (Kassin), a psychologist who studies perception and memory, and who has been qualified as an eyewitness identification expert in other cases.  The government opposed.  Should I allow the testimony of either defense's handwriting analysis critic or eyewitness expert, the government seeks counter experts, Dr. Ebbe B. Ebbesen (Ebbesen), on eyewitness identification, and Dr. Moshe Kam (Kam) on handwriting.

With respect to handwriting analysis, I held a Daubert/Kumho hearing at which Denbeaux testified, and the government cross examined.  In addition, the government offered several articles as well as the curriculum vitae of its expert, Kam. Harrison did not testify, as she had previously testified during the first trial, but I reviewed her prior testimony. . . .

This trial ended in a hung jury (as did the first). Since the issues noted in this memorandum will recur in the next trial, I outline my reasoning for the reasoning.

To some, Daubert/Kumho have considerably raised the bar for the admissibility of expert testimony.  In some ways that is true;  in some ways it is not.  It is true that Daubert/Kumho have focused a great deal of attention on the judge's role as a gatekeeper for expert testimony under Fed.R.Evid. 104(a).  To the extent that there is more pre-trial review of expert testimony, there are bound to be more exclusions.  In other ways, however, the impact of Daubert/Kumho has been the opposite--opening the door to testimony previously excluded.

Our evidentiary rules put a premium on firsthand observations.  Opinion testimony is disfavored except under certain circumstances;  hearsay is generally excluded. The jury is to draw reasonable inferences from the firsthand data.  When an expert witness is called upon to draw those inferences, several concerns are raised.  The rules give expert witnesses greater latitude than is afforded other witnesses to testify based on data not otherwise admissible before the jury.  In addition, a certain patina attaches to an expert's testimony unlike any other witness;  this is science, a professional's judgment, the jury may think, and give more credence to the testimony than it may deserve.

Accordingly, the trial court is supposed to review expert testimony carefully.  The court is to admit the testimony not only where it is relevant to the issues at bar, the usual standard under Fed.R.Evid. 401, but when certain additional requirements are met under Fed.R.Evid. 702.

The first requirement has to do with the necessity for the testimony:  expert testimony may be admitted where the inferences that are sought to be drawn are inferences that a jury could not draw on its own. .   .   .  Significantly, Daubert also emphasized the fact that expert testimony is admissible where it would assist the trier of fact.  In the latter case, even if the inferences may be drawn by the lay juror, expert testimony may be admissible as an aid in that enterprise.  For example, the subject looks like one the jury understands from every day life, but in fact, the inferences the jury may draw are erroneous. . . .

The second requirement concerns the nature of the inferences to be drawn.  In outlining the standards for admissibility, the Daubert Court noted the difference between information gleaned in a scientific setting and information presented in a courtroom. In the former, the decision makers are professionals;  there is no need to come to a definitive conclusion;  the decision making process comports with certain rules established by the professional scientific community.  In the courtroom, the decision makers are lay, a jury; there is a need to come to a definitive conclusion;  the decision making process has to satisfy norms of due process and fairness.  In our tradition, for example, the adversary system and party examination and cross examination are central.  The issue then, is not only how objectively reliable the evidence is, but also the legitimacy of the process by which it is generated.  It is not just how valid the data is, but how well the jury can understand it after direct and cross examination, and legal instructions.

It was not surprising, then, that the Court found the rules for admissibility of expert testimony, even with respect to scientific fields, were flexible, bearing in mind their unique setting and purposes. A trial judge could consider some or all of the following factors:  (1) whether the expert's technique can be or has been tested;  2) whether the method has been subjected to peer review and publication;  3) the known or potential rate of error of a technique or theory when applied;  and, 4) whether there is general acceptance within the relevant scientific community, to the extent that one exists.

Significantly, the general acceptance test of Frye v. United States, although now one of many standards, was still important.  Daubert's principal focus was necessarily on new science.  Its rules were intended to open the door to the admission of theories challenging existing orthodoxies.  The petitioners, children born with serious birth defects, allegedly caused by their mothers' ingestion of Bendectin during pregnancy, sought to introduce scientific studies, performed by well credentialed medical experts, that contradicted the existing generally accepted theory that Bendectin was not a teratogen (a substance capable of causing malformations in fetuses). Frye 's general acceptance test was inadequate to deal with new theories and approaches.  Having relinquished the Frye test, the Court was obliged to map other standards.

But the standards articulated by the majority in Daubert raised serious concerns. The Frye test was perfectly suited to the court's competence.  The judge did not have to be a scientist.  He or she only had, in effect, to count the scientific votes;  determine what the majority accepted.  Daubert/Kumho, however, required a judge, a notable generalist, to second guess a scientist.  Thus, by listing the scientific standards and encouraging review, the decision pointed in the direction of more rigor.  But, by underscoring the uniqueness of the trial setting, the assist the trier standard and flexibility, the decision does the opposite.

One way of reconciling the competing vectors is this:  Traditional science, generally accepted by judges for decades, may not need the same kind of rigorous analysis as new science which lacks the legitimacy of a chorus of sponsors.  As to the latter, the Daubert analysis is critical, and the court should be an especially vigilant gatekeeper.

Kumho extended Daubert to non-scientific fields.  In this category, for example, are the fields that are based on observations, not traditional science. . . .

At the same time, the Court plainly recognized the limitations of the  Daubert approach. It reaffirmed Daubert's holding that the court must be the gatekeeper, and must assure the reliability and relevance of testimony with respect to all expert testimony included in Fed.R.Evid. 702, not just scientific testing. But the nature of the test, notably, how much the Daubert standards are modified, was within the court's discretion.  And if the Daubert standards were flexible when applied to traditional scientific fields, surely they were flexible when applied to nonscientific, technical fields. . . .

Moreover, if the Daubert standard takes into account the unique trial setting--how well the lay trier will understand the testimony after examination and instructions--Kumho plainly does as well.  In fact, cross examination and limiting instructions may be more effective in technical fields because they are more accessible to the jury, than fields with the charisma of science.

Again, a mixed message:  Apply Daubert to technical fields, even though the scientific method may not really fit, but be flexible.  Moreover, in this setting, because few technical fields are as firmly established as traditional scientific ones, the new science/old science comparison is less clear.  The court is plainly inviting a reexamination even of generally accepted venerable, technical fields.

A. Handwriting

Handwriting analysis is one such field.  The Harrison testimony may be divided into two parts:  Part 1 is Harrison's testimony with respect to similarities between the known handwriting of Haines, and the robbery note. Part 2 is Harrison's testimony with respect to the author of the note, that the author of the robbery note was indeed Haines.  I concluded that Harrison could testify only as to the former.

One qualification:  This handwriting challenge was raised at the eleventh hour.  The hearing was necessarily constrained by the demands of the imminent trial and the schedules of the experts.  The Court was unwilling on this record to throw out decades of "generally accepted" testimony. Haines' challenge was far reaching.  To be sure, it would not be the first time that a "field" was invented and legitimized only by those within it, that then dissolved when tested by independent standards. Moreover, the compromise solution I accepted, allowing testimony about similarities and dissimilarities, not conclusions, derived largely from the case law that pre- dated Kumho.  Finally, the trial testimony suggested that the line I drew between testimony on similarities and conclusions about authorship, may well be a difficult one to enforce.  Haines, for example, objected to the government's closing on the ground that she implied that Harrison was a scientist, that her identification of similarities and dissimilarities between the Haines sample and the robbery note met that level of verification. . . .

Haines challenges Harrison's testimony under Daubert/Kumho.  If I were to give special emphasis to general acceptance or to treat Daubert/Kumho as calling for a rigorous analysis only of new technical fields, not traditional ones, then handwriting analysis would largely pass muster. Handwriting analysis is perhaps the prototype of a technical field regularly admitted into evidence. But, if I were to apply the Daubert/Kumho standards rigorously, looking for such things as empirical testing, rate of error, etc., the testimony would have serious problems.

According to Denbeaux, handwriting analysis by experts suffers in two respects.  It has never been subject to meaningful reliability or validity testing, comparing the results of the handwriting examiners' conclusions with actual outcomes. . . .   It has never been shown to be more reliable than the results obtained by lay people.  Some tests have been done, but all lacked a control or comparison group of lay persons. Thus, Denbeaux concludes, there is no need for expert testimony on handwriting analysis.  Lay people can do as well.  In this regard, I accept Denbeaux's testimony, and the article supporting it.

I do not believe that the government's expert, Kam, and the studies he has cited suggest otherwise.  While Kam has conducted several interesting and important tests, purporting to validate handwriting analysis, they are not without criticism.  They cannot be said to have "established" the validity of the field to any meaningful degree.

There is no question that lay witnesses are permitted to draw inferences of authorship from handwriting.  Where the lay witness is familiar with the handwriting, he or she can testify about it.  No expert testimony is necessary.  But just because lay witnesses can evaluate handwriting in some circumstances, does not necessarily mean that they can do it as well as an expert can in all.

Handwriting analysis typically involves reviewing two samples, a known sample and an unknown one, to determine if they are similar.  Both defense and government experts agree that unlike DNA or even fingerprints, one's handwriting is not at all unique in the sense that it remains the same over time, or uniquely separates one individual from another.  Everyone's handwriting changes from minute to minute, day to day.  At the same time, our handwriting is sufficiently similar to one another so that people can read each other's writing.  Given that variability, the expert is obliged to make judgments--these squiggles look more like these, these lines are shaped more like these, etc.  And those judgments are, as Harrison conceded, subjective.

When a lay witness, the girlfriend of the defendant for example, says "this is my boyfriend's writing," her conclusion is based on having been exposed to her paramour's handwriting countless times.  Without a lay witness with that kind of expertise, the government is obliged to offer the testimony of experts who have looked at, and studied handwriting for years.  These are, essentially, observational experts, taxonomists--arguably qualified because they have seen so many examples over so long.  It is not traditional, experimental science, to be sure, but Kumho 's gloss on Daubert suggests this is not necessary.  I conclude that Harrison can testify to the ways in which she has found Haines' known handwriting similar to or dissimilar from the handwriting of the robbery note;  part 1 of her testimony.

Part 2 of the Harrison testimony is, however, problematic.  There is no data that suggests that handwriting analysts can say, like DNA experts, that this person is the author of the document.  There are no meaningful, and accepted validity studies in the field. No one has shown me Harrison's error rate, the times she has been right, and the times she has been wrong.  There is no academic field known as handwriting analysis.  This is a field that has little efficacy outside of a courtroom.  There are no peer reviews of it.  Nor can one compare the opinion reached by an examiner with a standard protocol subject to validity testing, since there are no recognized standards.  There is no agreement as to how many similarities it takes to declare a match, or how many differences it takes to rule it out.

Moreover, the issue here is not only the validity and reliability of the expert testimony, but its validity and reliability in the context of this lay proceeding.  Harrison's account of what is similar or not similar in the handwriting of Haines and the robber can be understood and evaluated by the jury.  The witness can be cross examined, as she was, about why this difference was not considered consequential, while this difference was, and the jury can draw their own conclusions.  This is not rocket science, or higher math.

Her conclusion of authorship, however, has a difference resonance:  "Out of all of my experience, and training, I am saying that he is the one, the very author."  That leap may not at all be justified by the underlying data;  and in the context of this case, is extraordinarily prejudicial.

I find Harrison's testimony meets Fed.R.Evid. 702's requirements to the extent that she restricts her testimony to similarities or dissimilarities between the known exemplars and the robbery note.  However, she may not render an ultimate conclusion on who penned the unknown writing.  See Paul, 175 F.3d 906, 910 (letting writing expert identify points of comparison that he recognized between the two writings and allowing jury to conduct its own comparison and reach its own conclusion regarding authorship of the two writings); . . .

B. Eyewitness Identification

The government offered the testimony of Jeanne Dunne (Dunne), the teller.  Dunne, a white woman, gave the following identification moments after the robbery occurred:  She identified the man as black with dark skin, a wide nose, and a medium build. Her description was as close to a generic identification of an African American man as one can imagine.  Dunne was unable to identify Haines from a book of photographs of African American men shortly after the robbery. She picked out a few photographs, but none of them were as "dark black" as the robber.  Working with a police artist, she helped construct a sketch of the robber.  Immediately following, she was shown eight photographs, including one of Haines, and indicated that the Haines photograph "resembled" the robber, that it "looked like him," but she was still not sure.  (Since the robber was wearing a hat, she tried to envision the man in the photograph with a baseball cap.)  Months later Dunne picked Haines out of a lineup.

Haines offered the testimony of Kassin, a psychologist studying human perception at Williams College with substantial credentials, and trial experience.  The government offered a similarly credentialed expert, Ebbesen. I allowed the testimony of both.

The Kassin testimony was offered to show, inter alia, the following:  the decreased accuracy of cross-racial identification relative to same-race identification, the effect of stress on identification, the effect of time on memory as it relates to identification, the "confidence-accuracy" phenomenon which suggests the absence of any correlation between the amount of confidence expressed by an eyewitness in his or her memory and the accuracy of that witness' identification, the suggestiveness of subtle aspects of the identification process, such as the darkness of a particular photo as compared to others in the array, the fact that the eyewitness knows there is a suspect in the mix, the transference phenomenon by which a witness may believe that a face looks familiar but is unable to say whether her familiarity comes from seeing a previous mug shot, or from the robbery, etc.

On direct examination, Kassin identified those factors in the Dunne identification that were implicated in the studies with which he was familiar, and could undermine accuracy--the cross racial issues, the differences between the photographs of the other men and Haines' in the photo array,  the differences between Haines and the other men in the lineup.  He noted problems with what he called relative, comparative judgments:  The witness would like to resolve the case and so compares the photographs of one man to another in the array, rather than attempting to compare the photographs to the man she saw.  On cross, the government brought up the factors that enhanced accuracy, the nature of the lighting, the distance from the robber, the instructions that were given to the witness, especially at the lineup.  The government questioned Kassin about the instructions given to Dunne at the lineup.  Kassin agreed that those instructions were "ideal."

Unlike handwriting analysis, there is no question as to the scientific underpinnings of Kassin's testimony.  They are based on experimental psychological studies, testing the acquisition of memory, retention, and retrieval of memory under different conditions.  Indeed, the central debate before the jury, eloquently articulated by Ebbesen, the government's expert, is the polar opposite of the debate in the handwriting field--whether conclusions obtained in an experimental, academic, setting with college students should be applied to a real life setting. Kassin and others believe that these conclusions are appropriately applied to eyewitness identifications in court.  Ebbesen disagreed.

The government claimed that the jury did not need this testimony at all, that it was not necessary under Fed.R.Evid. 702 to assist the trier of fact.  I disagree.  While jurors may well be confident that they can draw the appropriate inferences about eyewitness identification directly from their life experiences, their confidence may be misplaced, especially where cross-racial identification is concerned.  Indeed, in this respect the rationale for the testimony tracks that for battered women syndrome experts.  The jury, for example, may fault the victim for not leaving an abusive spouse, believing that they are fully capable of putting themselves in the shoes of the defendant.  In fact, psychological evidence suggests that the "ordinary" response of an "ordinary" woman are not in play in situations of domestic violence where the victim suffers from "battered women syndrome."  Common sense inferences thus may well be way off the mark.

Nor do I agree that this testimony somehow usurps the function of the jury.  The function of the expert here is not to say to the jury--"you should believe or not believe the eyewitness."  (Indeed, it has far fewer pretensions to conclusions than does handwriting analysis, with far more science attached to it.)  All that the expert does is provide the jury with more information with which the jury can then make a more informed decision.  And only the expert can do so.  In the absence of an expert, a defense lawyer, for example, may try to argue that cross racial identifications are more problematic than identifications between members of the same race, or that stress may undermine accuracy, but his voice necessarily lacks the authority of the scientific studies Kassin cited.

Finally, the fact that the expert has not interviewed the particular eyewitness makes it less likely that the jury will merely accept the expert testimony and more likely that the testimony will be appropriately cabined. The witness can only be providing the jury with the tools to analyze the eyewitness;  he has no more specific information.  The science makes no pretensions that it can predict whether a particular witness is accurate or mistaken.

In my judgment, the accuracy of these proceedings was enormously enhanced by treating the jury to all sides of the eyewitness debate, rather than assuming there was no controversy, that the issue, notwithstanding this literature is clear.

II. CONCLUSION

Accordingly, the testimony of the government's handwriting expert was limited as described above, to identifying the similarities and dissimilarities of the known writing and the unknown.  The defense expert on eyewitness identification was *74 permitted along with the testimony of the government witness.

SO ORDERED.