Opinion
No. CR. H-02-0040
June 9, 2003.
Michael Wayne Ramsey, Attorney at Law, Houston, for Robert Nicholas Angleton, defendant.
Edward F Gallagher, Terry Clark, Melissa Annis, U.S. Attorneys Office, Houston, TX, James L Turner, U.S. Attorneys Office, Houston, TX, for U.S. Attorneys.
MEMORANDUM AND OPINION
Defendant Robert Angleton seeks to introduce expert testimony of Steven Smith, a professor of psychology at Texas A M University, relating to the weaknesses of "earwitness" identification of voices. The government has moved to strike Smith's expert testimony on the ground that it does not meet Federal Rule of Evidence 702.
This court has carefully considered the record, including Smith's testimony presented at the hearing held on April 28, 2003; the motions and responses; the parties' submissions; and the applicable law. Based on this review, this court concludes that certain portions of Smith's testimony meet the standard necessary for admission under Rule 702. The government's motion to exclude the testimony of Steven Smith is GRANTED in part and DENIED in part, as explained below.
I. Analysis
A. The Record as to the Opinions
Smith holds a bachelor's degree in psychology from the University of Michigan and a master's degree and Ph.D. in experimental psychology from the University of Wisconsin. (Defendant's Exhibit 7). He has been a professor at Texas A M since 1980. His research has focused on the factors that affect human memory and the phenomena of false memory and recovered memory. (Docket Entry No. 168, p. 140, 1.20-1.22). He has published or edited four books on the cognitive basis of creative thinking. (Defendant's Exhibit 7). Smith is presented as an expert on factors that affect the reliability of "earwitness" identification.
A false memory is "the experience of remembering something that doesn't correspond with physical facts." (Docket Entry No. 168, p. 142, 1.12-1.14).
Smith testified that "earwitness identification is just a subset of eyewitness identification." The term "earwitness identification" is used to mean identification of a speaker by listening to the voice similar to identification of a person by seeing them. (Docket Entry No. 168, p. 148, 1.22-p. 149, 1.4). Smith stated that research studies have found that earwitness identification is generally less accurate than eyewitness identification, which can be unreliable. (Docket Entry No. 168, p. 149, 1.21-1.23).
At the Daubert hearing, Smith testified that the length of a speech sample an individual hears affects that individual's ability to identify the speaker. The longer the speech sample a person hears, the more accurate that person's identification of the speaker will be, up to a point. Smith stated that "[i]f you hear a voice that says a single word, then your identification won't be as good as if you hear a single sentence. If you hear a single sentence, that won't be as good as if you hear a number of sentences." ( Id. at p. 155, 1.3-1.7). Smith testified that once the speech sample exceeds a certain length, however, the likelihood of a false identification increases. ( Id. at p. 155, 1.12-1.13, p. 155, 1.24-p. 156, 1.1). A false identification is more likely if the length of the recorded speech sample of the unknown speaker exceeds thirty minutes, as compared to a six-minute-long speech sample. ( Id. at p. 155, 1.24-p. 156, 1.1).
Smith testified that when a listener has an opinion that the recorded voice of an unknown speaker is the voice of a particular individual, the identification process is biased and misidentification is more likely. ( Id. at p. 158, 1.12-1.20). Discussing the possible identity of the speaker with others who are also trying to identify the speaker by listening to the recording increases the possibility of misidentification. ( Id. at p. 159, 1.1-1.3). Smith testified that listening to only one voice recording, as opposed to a "lineup" of different but similar-sounding recordings, may also cause the listener to identify the speaker as the person who has been accused of the crime at issue. ( Id. at p. 159, 1.8-1.14).
Familiarity with a speaker's voice also affects the accuracy of earwitness identification, according to Smith. Smith testified that where a person has a high degree of familiarity with the speaker, for example, if the speaker is a member of the person's immediate family, the misidentification rate for voice identification is 10 percent or less. ( Id. at p. 160, 1.10-1.16). If the listener is a friend or coworker of the speaker and has moderate familiarity with that speaker, the misidentification rate is between 10 and 20 percent. ( Id. at p. 160, 1.17-1.20). The rate of misidentification when the listener has less familiarity with the speaker is between 30 and 50 percent. ( Id. at p. 160, 1.21-1.24).
Smith stated that the quality of the recording containing the unknown speaker affects the reliability of earwitness identification. He testified that a distorted tape recording increases the likelihood of a misidentification. ( Id. at p. 166, 1.6-1.15). He also testified that a distorted tape exacerbates other biases present in the identification process, such as prior communication among several listeners trying to identify the speaker. ( Id.).
Based on these general factors that affect the reliability of earwitness identification, Smith criticized the process used to obtain several of the earwitness identifications made during the earlier state court trial. Smith noted that Mary Hill had testified that the tape allegedly containing the voices of Roger and Robert Angleton was brought to her home by Detectives Novak and Ferguson of the Houston Police Department. She and four other people gathered at her home to listen to the tape and try to identify the speakers. ( Id. at p. 162, 1.16-1.19; Defendant's Exhibit 9, p. 141, 1.11-1.23). Hill testified that the group at her home discussed the identification before the officers arrived with the tape. (Defendant's Exhibit 9, p. 141, 1.24-1.25). Hill testified in the state court trial that before she listened to the tape, Novak and Ferguson had told her that they believed the voices were those of Roger and Robert Angleton. ( Id. at p. 142, 1.8-1.9, 1.15-1.24). Smith testified that these circumstances created a bias, increasing the likelihood that Hill and the others would misidentify the voices on the recording as Roger and Robert Angleton. (Docket Entry No. 168, p. 163, 1.8-1.13).
Mary Hill described herself as a "very good friend of Doris [Angleton]." She stated that she also socialized with Robert Angleton. (Defendant's Exhibit 8, p. 131, 1.1-1.3).
Smith also testified to potential bias in Kevin Templeton's identification of the unknown speaker on the recording. Templeton, a Houston Police Department officer, testified in the state court trial that he was under internal investigation by the Houston Police Department for his relationship with Robert Angleton. (Defendant's Exhibit 10, p. 7, 1.11-1.21). Templeton testified that he thought the unknown speaker on the tape was Robert Apngleton. ( Id. at p. 8, 1.4-1.5). Templeton also testified that he felt there was a possibility that, as a result of the police department investigation, he would lose his job and could be subject to criminal prosecution if Angleton was acquitted in the state court trial. ( Id. at p. 11, 1.21-p. 12, 1.5). Smith testified that Templeton's fear of the consequences of Angleton's acquittal could bias his identification of the unknown speaker heard on the tape. (Docket Entry No. 168, p. 164, 1.20-1.24).
Smith summarized his criticisms of the voice identifications made in the state court trial by identifying the following "more obvious" problems with the voice identification performed by the Houston Police Department:
(1) the use of a single voice recording, rather than a lineup of several speakers;
(2) the bias created by witnesses interacting before they made the identification;
(3) the bias created by Novak's and Ferguson's statement that they thought the speakers on the tape were Roger and Robert Angleton; and
(4) the failure to warn the witnesses that the person they were supposed to be identifying might not be the person speaking on the recording.
( Id. at p. 165, 1.13-1.23).
Angleton seeks to introduce Smith's testimony as to the factors that may affect the reliability of the earwitness identifications made in this case.
B. The Case Law on Eyewitness Identification
Smith repeatedly stated that earwitness identification is a form of eyewitness identification. The case law on the admissibility of expert psychological testimony on the validity and reliability of eyewitness identification is helpful. Courts were initially skeptical of such testimony. See United States v. Thevis, 665 F.2d 616, 641 (5th Cir. 1982); United States v. Amaral, 488 F.2d 1148 (9th Cir. 1973). The trend is, however, toward admitting such testimony after ensuring that it meets Rule 702. See United States v. Smith, 156 F.3d 1046, 1053 (10th Cir. 1998) (stating that expert testimony on eyewitness identification is admissible under Daubert in certain circumstances); United States v. Brien, 59 F.3d 274, 277 (1st Cir. 1995) (refusing to adopt a "blanket rule that qualified expert testimony on eyewitness identification must routinely be admitted or excluded"); United States v. Moore, 786 F.2d 1308, 1312 (5th Cir. 1986) (stating that "the scientific validity of the studies confirming the many weaknesses of eyewitness identification cannot be seriously questioned at this point." (citation omitted)). A district court has broad discretion to determine the admissibility of expert eyewitness identificatior testimony. United States v. Langan, 263 F.3d 613, 621-22 (6th Cir. 2001); Moore, 786 F.2d at 1312-13. "There is no federal authority for the proposition that [eyewitness identification] testimony must be admitted." Moore, 786 F.2d at 131213.
Several courts have held that expert testimony as to the validity of eyewitness identification is admissible when focused on problems that lie outside the common experience of jurors, such as cross-racial identification, identification after a long delay, identification after observation under stress, and such psychological phenomena as the feedback factor and unconscious transference. See Langan, 263 F.3d at 621; Smith, 156 F.3d at 1053; Harris, 995 F.2d at 535. "The trend is to admit expert testimony on eyewitness identification under certain circumstances, which should be examined on a case-by-case basis." Smith, 156 F.3d at 1053.
The "feedback factor" involves the effect of post-event information on the memory of the event, including discussions among witnesses which may unconsciously reinforce mistaken identifications. Smith, 156 F.3d at 1052; Moore, 786 F.2d at 1311. "Unconscious transference" allows a person to remember a face but not the circumstances under which the person saw the face. Smith, 156 F.3d at 1052.
Some courts exclude expert testimony on eyewitness identification under Rule 702 if the testimony is on matters within the common knowledge of jurors and therefore not helpful to their deliberations. Langan, 263 F.3d at 621; United States v. Hall, 165 F.3d 1095, 1104 (7th Cir. 1999); Harris, 995 F.2d at 534. The court has discretion to exclude expert testimony that "will not aid the jury because it addresses an issue of which the jury already generally is aware, and it will not contribute to their understanding of the particular dispute." Hall, 165 F.3d at 1104; United States v. Mathis, 264 F.3d 321, 340-41 (3d Cir. 2001) (omitting expert testimony because it was "susceptible of elucidation without specialized knowledge"). "Jurors using common sense and their faculties of observation can judge the credibility of an eyewitness identification, especially since deficiencies or inconsistencies in an eyewitness' testimony can be brought out with skillful cross-examination." Harris, 995 F.2d at 535. The court may also exclude expert testimony as to factors affecting the reliability of eyewitness identification if those factors can be adequately explained in jury instructions. See United States v. Rincon, 28 F.3d 921, 925 (9th Cir. 1994).
Courts considering the admissibility of such expert testimony recognize that such testimony has the potential to be substantially prejudicial under Federal Rule of Evidence 403. The "aura" of expertise as to the effect of a particular factor on the reliability of eyewitness identification can mislead a jury to believe that the presence of that factor in an eyewitness identification makes that identification less reliable than it truly is. See United States v. Lester, 254 F. Supp.2d 602, 608-09 (E.D.Va. 2003); Rincon, 28 F.3d at 925.
C. Applying the Case Law to Smith's Proffered Testimony
Under Daubert, this court must determine whether Smith's testimony is reliable and whether it will assist the trier of fact. As to reliability, a number of courts have admitted testimony of psychologists who are experts in factors affecting eyewitness identification. See Smith, 156 F.3d at 1053 (expert testimony on eyewitness identification may properly be admitted under Daubert in certain circumstances). The Fifth Circuit has held that "the scientific validity of the studies confirming the many weaknesses of eyewitness identification cannot be seriously questioned at this point." Moore, 786 F.2d at 1312. The Sixth Circuit has found that "the science of eyewitness perception has achieved the level of `exactness, methodology, and reliability of any psychological research.'" Langan, 263 F.3d at 622 (citation omitted). Several other circuits have acknowledged that expert testimony on the factors affecting eyewitness identification is admissible under certain circumstances. Stevens, 935 F.2d 1380; Smith, 156 F.3d at 1053; Harris, 995 F.2d at 535.
The record shows that Smith is well qualified, having conducted research in this field for over twenty years and published numerous books and articles. In this case, the government made a limited challenge to the reliability of the research and theories Smith relied on in his direct testimony. Smith credibly responded that the studies he relied upon replicated their underlying experiments as part of the study. (Docket Entry No. 168, p. 178, 1.9-1.19). Smith also stated that those studies applied statistical analyses to ensure the significance of their conclusions. ( Id. at p. 178, 1.20-1.25). Smith could not name scientists who disputed the validity of the studies upon which he relied. The government did not produce studies disagreeing with any of the studies on which Smith relied. ( Id. at p. 179, 1.1-1.8). See United States v. Mathis, 264 F.3d 321, 339 (3d Cir. 2001) (where well-qualified eyewitness identification expert's testimony was close to the heart of his field of expertise, and the reliability of the studies he relied upon was not impeached, testimony met the Daubert reliability standard).
Nor did the government challenge the relevance of Smith's analysis. Smith testified as to the effect of specific factors that the state court record revealed as part of the voice identification at issue in this case. The facts on which Smith based his opinions "fit" this case. The government's primary argument is that Smith's testimony will not be helpful to a jury because it covers subjects within jurors' common knowledge.
Smith testified to several factors that cast doubt on the reliability of the earwitness identifications made by witnesses in the state court trial. Some of these factors are within the common knowledge of the jurors, but not all.
Smith generally testified that the longer the speech sample heard by an earwitness, the more accurate the identification of the unknown speaker. This observation is unremarkable and well within the common knowledge of the jurors. Expert testimony on this point unnecessary. However, Smith also testified that if the speech sample heard by the earwitness exceeds a certain length, the likelihood of false identification begins to rise. The proposition that increasing the length of the recorded speech increases the accuracy of voice identification seems intuitive; the proposition that a speech sample beyond a certain length leads to an increase in false identifications is not. This court cannot say that this principle is within the common knowledge of the juror. See Lester, 254 F. Supp.2d at 612 (admitting expert testimony where it could not be confidently said that a particular factor was within the knowledge of the average juror). It is not a point that can be easily raised on cross-examination. Jury instructions instructing the jury to consider the length of the recorded speech sample used by the earwitness to identify the speaker could be confusing and could not convey this nonintuitive point that a voice recording should be neither too short nor too long, to facilitate reliable speaker identification.
The recording at issue is forty-seven minutes long. Smith testified that recordings, of thirty minutes or longer result in increased misidentifications. Smith's testimony that the possibility of misidentification increases when a tape recording exceeds a certain length is relevant and could be helpful to the jury. Smith's testimony is admissible for the purpose of identifying possible weaknesses in earwitness identification associated with the length of the tape recording of the unknown speaker's voice.
Smith stated that misidentification is more likely when the listener already believes that the voice heard on a recording is that of a particular individual. The proposition that an earwitness's preexisting belief as to the identity of the speaker affects the reliability of the earwitness identification is intuitive. it falls within the common sense of an average juror. It is a point that can readily be elicited on cross-examination to challenge the credibility of the earwitness identification. The court can clearly instruct jurors to consider the possibility that the identification may be affected by such an outside influence. See Hall, 165 F.3d at 1107 (jury instructions instructing the jury to consider whether "the influences and circumstances under which the witness has made the identification"); Rincon, 28 F.3d at 925-26 (jury instructions which instructed the jury to consider whether "the identification was the product of the eyewitness's own recollection or was the result of subsequent influence or suggestiveness" were adequate, making expert testimony unnecessary). Expert testimony on this factor is unnecessary and unhelpful to the jury.
Smith also identified the earwitness's familiarity with the unknown speaker's voice as affecting the reliability of speaker identifications. Smith testified that the greater the familiarity of the earwitness with a person's voice, the more accurate the voice identification. This factor is intuitive and within common experience. The depth of an earwitness's familiarity with the voice of the person they claim to have identified can be examined in detail on cross-examination. Jury instructions can clearly inform jurors of the need to consider the familiarity of earwitnesses with the voice of the person they identify as the speaker on a recording. Expert testimony is unnecessary and unhelpful as to this factor.
Smith also testified that the quality of a voice recording affects the reliability of an earwitness identification. Smith stated that distortions on a recording can lead to less accurate earwitness identifications. This principle is also intuitive and well within a juror's common knowledge. The clarity of the recording from which an earwitness made an identification can readily be explored on cross-examination. Indeed, Angleton's counsel effectively explored the quality of the tape in cross-examining Templeton in the state trial. (Defendant's Exhibit 10, p. 8, 1.1-1.21).
Smith testified that the fact that Novak and Ferguson told Hill that they thought Robert Angleton's voice was on the tape would increase the likelihood that Hill would misidentify the speaker on the tape as Robert Angleton. The fact that misidentifications are more likely where the person making the identification is told in advance who is believed to be the unknown speaker on a tape is a proposition that is within the common knowledge of the jury. The possibility of bias can be explored with cross-examination. Expert testimony on this point is not necessary. In United States v. Stevens, 935 F.2d 1380 (3d Cir. 1991), the defendant proffered an expert to testify that a board of photographs of possible suspects shown to the eyewitness was overly suggestive, where the actual suspect was one of the few suspects displayed in a photograph rather than a sketch, was the only person whose photograph was displayed twice, and was the only person whose photograph was in color. The court stated that the fact that the wanted board was overly suggestive was "a rather intuitive proposition . . . [that defendant] could have argued . . . to the jury without adducing expert testimony." 935 F.2d at 1399. Similarly, the proposition that Novak's and Ferguson's statement that they believed Robert Angleton to be speaking on the tape was overly suggestive is within the common knowledge of jurors. Expert testimony is unnecessary and unhelpful.
Smith also stated that the possibility of a misidentification increases when the earwitness is told in advance that the speaker is a "good guy" or a "bad guy." (Docket Entry No. 168, p. 161, 1.3-1.9). Smith did not cite a particular study for this observation, and only recalled in passing that he had seen one. This observation is not relevant and inadmissible because it does not fit the facts of this case. Hill testified that Novak and Ferguson told her that they believed Robert Angleton's voice was on the recording, not that they believed that an unspecified "bad guy" was on the recording.
Smith also testified that the interaction among Mary Hill and the four other witnesses gathered at her home before hearing the tape supplied by Novak and Ferguson possibly biased the identification those witnesses made of the speaker on that tape. The Fifth Circuit in Moore recognized the admissibility of expert testimony as to the effects of the "feedback factor," where witnesses who discuss the case unconsciously may reinforce mistaken identifications. Moore, 786 F.2d at 1312. Other courts have also recognized the admissibility of expert testimony as to the feedback factor. See Harris, 995 F.2d at 535 (stating that the feedback factor is one of the recognized "narrow circumstances" in which expert testimony is admissible); State v. Chapple, 135 Ariz. 281, 660 P.2d 1208, 1221 (1983) (finding that under the facts of the case, trial court abused its discretion in excluding testimony as to the effect of discussions between witnesses on their identifications); United States v. Downing, 753 F.2d 1224, 1230-31 (3d Cir. 1985) (citing Chapple with approval).
The facts of this case support the admission of the testimony on the effect of the group interaction before the tape was played. The identification by Hill and the others gathered at her home occurred after they had discussed the case. This does not appear to be an exact application of the feedback factor, because the definitions of the feedback factor in the case law indicate that the feedback factor reinforces misidentifications already made. However, the possibility that earwitness discussions before identifying a voice could lead to unconscious bias during identification is analogous to the effects of the feedback factor. The possibility that such discussions could bias voice identification, even when the discussions did not necessarily cause the listener to form a definite opinion as to who might be speaking on the recording, is not within the common knowledge of the jury and cannot be easily elucidated on cross-examination. The effect of conversations among the witnesses gathered at the Hill home on the earwitness identification of the speaker on the tape as a result of the feedback factor is relevant and helpful and is a proper subject for expert testimony.
Smith raised the possibility that Templeton's identification of Angleton as the unknown speaker on the tape was influenced by his concerns about the Houston Police Department's ongoing investigation into Templeton's relationship to Angleton. The point that Templeton may have had a personal motive or interest in a particular outcome of the state court trial generally and of the voice identification issue in particular is not a point on which expert testimony is helpful. Jurors understand that testimony can be affected by self-interest and bias. Cross-examination is the time-honored tool for exposing such bias, and jury instructions on the credibility of witnesses routinely address bias factors the jury should consider.
Smith also testified that the fact that Novak and Ferguson presented a single tape recording to the earwitnesses at the Hill home, without presenting other voice recordings for comparison, and the fact that Novak and Ferguson did not warn the earwitnesses that the person whom they were supposed to be identifying might not be the speaker on the tape, may have increased the risk of misidentification. Smith testified as to the effect of presenting earwitnesses a single tape with an unknown speaker to identify, without other voices for comparison, analogous to a lineup for visual eyewitness identification: "if you have to make some inference as to why they are having me listen to this one tape or why they are having me look at this one photograph, the clear inference for most people is that they think this is the person who did it. Just the possibility of that inference is what increases the likelihood of misidentification." (Docket Entry No. 168, p. 159, 1.9-1.14).
The proposition that the use of a single recording of a voice, rather than a "lineup" of similar voices as is done for eyewitness identification, increases the reliability of the voice identification, is not intuitive. Smith emphasized that "earwitness" identification is a form of eyewitness identification. Eyewitness identification of a person suspected of a crime from a lineup of people is within a juror's common knowledge. However, a juror may not think of earwitness identification as requiring the same type of forensic approach as visual eyewitness identification. The common experience for the average juror likely is that voice identification consists of playing a single recording and asking the listener to identify the speaker. Smith testified that this is not the most reliable method of earwitness identification. This proposition cannot readily be elucidated solely on cross-examination and jury instructions alone will not be as helpful as expert testimony. This testimony is relevant to the facts of this case because the earwitnesses only listened to the one recording allegedly containing Angleton's voice. Expert testimony on this factor's effect on the reliability of the voice identification is admissible.
II. Conclusion
Smith's testimony involves some factors relating to the reliability of earwitness identification that are intuitive and are well within the common knowledge of the jury or are "susceptible of elucidation without specialized scientific knowledge." These factors may adequately be explored on cross-examination and be addressed in jury instructions. Smith's testimony as to these factors is inadmissible under Daubert because it does not assist the trier of fact.
Some of Smith's testimony involves factors relating to the reliability of earwitness identification that are unintuitive and are admissible. These factors are: the effect of the length of the sample of the unknown speaker heard by the person making the identification up to a certain point; the effect of discussions among earwitnesses prior to identification; and the effect of listening to only one sample recording rather than a lineup of similar sample recordings. These factors, relevant to this case, are not intuitive, are not within the common knowledge of the jury, and cannot easily be explored on cross-examination or covered in jury instructions. Smith's testimony as to these specific factors would be helpful to the jury and is admissible.