People v. McDonald

690 P.2d 709, 37 Cal. 3d 351, 208 Cal. Rptr. 236, 46 A.L.R. 4th 1011, 1984 Cal. LEXIS 129

• Court: California Supreme Court
• Decided: November 21, 1984
• Docket: Crim. 21770
• Status: Published

Opinion

Opinion

MOSK, J.

We address here a contention that is increasingly heard in the courts of California and our sister jurisdictions, i.e., that it may be an abuse of discretion to exclude the testimony of a psychologist who is a qualified expert witness on psychological factors shown by the evidence that may affect the accuracy of an eyewitness identification of the defendant. As will appear, we hold that on a proper showing such testimony is admissible, and that it should have been admitted in the case at bar.

Defendant was charged in count I with the murder of Jose Esparza. (Pen. Code, § 187.) As a special circumstance it was alleged that defendant committed the murder while robbing or attempting to rob Esparza. (Id., § 190.2, subd. (a)(17)(i).) In count II defendant was charged with the same robbery of Esparza as a substantive offense. (Id., § 211.) He pleaded not guilty on both counts. The jury convicted him of the murder and found the robbery special circumstance allegation to be true; nevertheless, the jury found defendant not guilty of the substantive charge of robbery in count II. The jury thereafter fixed the penalty at death. This appeal is automatic. (Id., § 1239, subd. (b).)

At trial it was established without dispute that August 20, 1979, was payday for Esparza, a restaurant worker. At 4 p.m. he took a break from his job to cash his paycheck. Shortly after 5 p.m. he was shot and killed by a black man at the intersection of Pine and Seventh Streets in downtown Long Beach. The principal issue was the identity of the perpetrator. The prosecution presented seven eyewitnesses who identified defendant as that person with varying degrees of certainty, and one eyewitness who categorically testified that defendant was not the gunman; the defense presented six witnesses who testified that defendant was in another state on the day of the crime. Because of these discrepancies and their bearing on the principal issue, we will set forth the relevant testimony in some detail.

Four prosecution witnesses positively identified defendant in the courtroom as the perpetrator; in the testimony of each, however, there were factors that could have raised reasonable doubts in the minds of jurors as to the accuracy of the identification. Thus Patricia Molinar testified that she was driving home after work on August 20 when she stopped for a red light at the intersection of Pine and Seventh Streets. While waiting at the light she noticed two men standing on the sidewalk diagonally across the intersection; one was Hispanic (Esparza) and the other was black. She testified that they “seemed to be arguing.” The black man then grabbed the other “around his hands,” and they began to struggle. After a few moments the witness heard a shot and saw the black man pointing a gun at the victim. The latter fell back against a wall, a second shot was heard, and he slumped to the sidewalk. The gunman leaned over him for a few seconds, walked quickly across the street, and ran from the scene. The entire episode took no more than a few minutes. In open court Ms. Molinar identified defendant as the black man in question.

The witness conceded, however, that her view had been partly blocked by cars parked in front of the spot where the confrontation took place. A number of cars also drove through the intersection while she was waiting for the light to change, passing between her and the two men. She testified that during part of the struggle the gunman had his back towards her, and that when he looked in her direction as he left the scene she was “frightened.” She acknowledged that after the event she failed to select defendant’s picture out of a group of some 10 photographs shown to her by the police. Yet she claimed that a few days later she picked defendant’s photograph out of a set of six, saying, “I’m not sure, but from the photograph I can’t be completely positive, but I’m more than sure it’s him.” On cross-examination; however, she admitted that at the preliminary hearing she testified that she had been unable to identify defendant in either photographic display.

David Iglesias testified he was a passenger in the front seat of the car driven by Ms. Molinar, who was his fianceé. His version of the events was essentially the same as hers. He also identified defendant in the courtroom, but conceded that when he picked defendant out of a pretrial photographic display he told the police that although he was “pretty sure” defendant was the gunman, he was “not positive.” On cross-examination Mr. Iglesias admitted that in fact he selected photographs of not one but two people (out of six) in the display—the defendant and another— “because they kind of looked alike.”

Chad Wise testified he was sitting in his parked car at the intersection in question when he heard a gunshot. He got out of his car and went to the corner, where he saw a black man and a Hispanic who seemed to be fighting; the black man pointed a gun at the Hispanic and fired a second time, reached into the victim’s back pocket, then ran down the street. The witness identified defendant in the courtroom. There was no evidence, however, that he had identified defendant before the trial; and defense counsel brought out certain discrepancies between the version that the witness gave on the stand and the statement he furnished to the police on the day of the events. Finally, his estimate of the time that it took the gunman to fire the second shot, bend over the victim, and run from the scene differed drastically from that of the other witnesses: they all testified that no more than two or three minutes elapsed, but he insisted on both direct and cross-examination that it was “almost 25 minutes.”

La Wahna Eldred testified she was walking towards the intersection in question when she heard two gunshots in close succession. Looking in the direction of the sounds, she saw a Hispanic struggling with a black man. The latter then stood up and walked across the street towards her, holding his hand at his waistband. She looked away because she was frightened; as he passed by she took a “sideward glance” at him. On this basis she identified defendant in the courtroom as the gunman.

Ms. Eldred conceded that her view of the encounter had been partly blocked by both parked and passing cars. Several days later she was shown a set of six photographs by the police; after studying them for ten minutes, she selected the photograph of defendant as having the most “similarities” to the face she remembered seeing. She told the police, however, that his hairline looked different and she “wasn’t totally positive” that defendant was the gunman.

None of the other prosecution witnesses were positive in their courtroom identifications of defendant. Thus the testimony of Erik Soderholm was very similar to that of Ms. Eldred. 1 When asked if he could identify the black man in the courtroom, however, Mr. Soderholm replied only that “I think this is the man,” pointing to defendant. He explained that “The feeling I get, looking at his face, are [sz'c] similar to the feelings I had when I saw the man go.” He conceded there was fairly heavy rush-hour traffic in the intersection at the time of the shooting, and that a number of cars passed between him and the scene as he watched. He also conceded that he had selected not one but two faces out of the set of six photographs subsequently shown to him by the police, and that he told the officer at the time, “I’m not sure. I only saw the side view of him and his back.” When defense counsel squarely asked him on cross-examination whether “There is some doubt in your mind” about his courtroom identification of defendant, Mr. Soderholm said, “Yes.”

Two other witnesses did not see the actual shooting and were even less certain that the black man at the scene was defendant. Harold Malone testified he was in a hamburger stand on the corner of Pine and Seventh Streets when he heard shots and went outside to investigate. He saw a black man standing over a fallen figure, holding a gun; the man put the gun in his waistband and crossed the street at a fast pace. The witness followed until the man got into a parked car and drove away.

There was no evidence that the witness identified defendant before the trial. When asked on direct examination, “Would you recognize that man if you saw him again?” Mr. Malone replied, “That I’m not sure of, sir.” He explained that his attention had been focused on the man’s weapon and on the need to follow him, and that at all times he was across the street and behind the man. The prosecutor persisted, asking the witness to “look around the courtroom and tell us if you see that man in court today.” Mr. Malone did so, and answered, “I couldn’t be positive.” The prosecutor then asked if he could see anyone “similar” to the black man in question, and the witness finally pointed to defendant. 2 When defense counsel asked him on cross-examination, “you’re not certain it’s [defendant], are you?” the witness admitted, “No, sir. I couldn’t be positive, no.”

Similarly, Richard Kaley testified he was standing near the intersection in question when he heard a shot and saw a man running “lickety-split” towards him. As the man passed him, Mr. Kaley saw that he was holding a gun in his waistband. The entire episode took about 90 seconds. The witness conceded that the police never showed him a photographic display and that he did not testify at the preliminary hearing. At trial he testified the man was black, but added that he (Kaley) was colorblind and was wearing sunglasses at the time. When the prosecutor asked, “Would you recognize this person if you saw him again?” Mr. Kaley answered, “That would be debatable.” In response to the prosecutor’s directive to look around the courtroom and say if the man was present, the witness said that defendant greatly “resembled” him. On cross-examination defense counsel asked him whether his identification of defendant was “positive,” and Mr. Kaley replied, “I said that it was not.”

The prosecution witnesses were in general agreement in their description of the clothing worn by the gunman, but two (Molinar and Iglesias) claimed the man had a large, round, gold earring in his left ear, about the size of a quarter, while none of the other witnesses testified that he wore any such distinctive jewelry. In addition, two witnesses (Eldred and Kaley) described the man as having “pockmarks” or “acne-like” scars on the lower part of his face, while none of the other witnesses so testified.

Finally, one of the prosecution’s own witnesses unequivocally testified that the black man at the scene was not defendant. Helen Waller was driving slowly through the intersection of Pine and Seventh Streets when her attention was caught by two men “arguing and making loud noises” on the sidewalk just outside her right-hand window. She took her foot off her accelerator pedal and watched while one of the men, a black, shot the other with a large handgun and sought to wrest a bag or purse from his victim’s grasp. Mrs. Waller testified that the gunman “looked directly in my eyes, and I looked at him.” Her car continued to coast, and the struggle of the two men remained, she said, “all very much in my view.” At that point there were no parked cars between her and the scene, and the traffic was light. After about a minute the struggle ended and the witness watched the gunman as he passed directly behind her car and across the street.

Mrs. Waller reported to the police that evening; several days later an officer showed her a set of six photographs, the same set that was shown to witnesses Molinar, Iglesias, Eldred, and Soderholm. She told the officer that “none of the photographs actually looked like the man,” although she noted that in one of them the hair, eyes, and general shape of the face looked similar to those of the assailant. 3 Mrs. Waller was troubled by the complexion of the person in the photograph, but the officer told her that a photograph may not accurately reproduce the subject’s skin color.

It was then brought out that Mrs. Waller testified in defendant’s presence at the preliminary hearing; the prosecutor there asked her whether the black man who shot Esparza was in the courtroom, and she said he was not. 4 At trial she acknowledged there were certain similarities between that man and defendant, pointing to their eyes, the shape of their faces, and their height. But she explained that defendant also exhibited a feature that was “very different. He has a different complexion. This man here [i.e., defendant] has a complexion that has a yellowish hue to it. And the man that I saw had more brown, much deeper brown and much deeper color in his skin coloring than this man. And it’s something that doesn’t change with being inside or outside. It’s not like a tan. It’s a coloring.” 5 For this reason Mrs. Waller was certain they were “Two different men altogether.” On cross-examination defense counsel squarely asked her whether defendant was the black man in question, and she replied, “No, he’s not.”

The prosecution offered no other evidence to connect defendant with the crime in this case. The defense, however, called six witnesses to establish that on the date of the shooting (Aug. 20, 1979) defendant was visiting his grandfather in Saraland, Alabama, near Mobile. Lovie Banks, defendant’s fianceé, testified that on August 10, 1979, she drove defendant to San Diego, where he took a bus to Saraland, via Phoenix. She subsequently received two cards from defendant, both postmarked Phoenix, Arizona, August 11, 1979; the cards were introduced in evidence. She testified that she thereafter received a letter from defendant from Saraland, and that he called her twice from his grandfather’s house.

Jessie Mae Pruitt, defendant’s aunt, testified that on August 13, 1979, she received a collect call from defendant from the bus station in Mobile asking to be met; a copy of a long-distance telephone bill showing that a collect call was made on August 13 from Mobile to her number was introduced into evidence. The witness testified she then telephoned to Robert Pruitt, defendant’s grandfather, and asked that defendant be met at the bus station. Her telephone bill showed that several calls were made that same day from her number to Alabama.

Robert Pruitt testified that his wife took the telephone call concerning defendant’s arrival, and that he met defendant at the Mobile bus station between 4:30 and 5 p.m. on August 13 and took him home to Saraland. He further testified that defendant lived with him and his wife until the police took him back to California in September.

Three other witnesses placed defendant in Saraland from mid-August to September 1979, and specifically testified that they saw and/or spoke with him there on August 20, the date of the shooting in Long Beach. Two of these witnesses were unrelated to defendant. In addition, Ms. Banks testified that on August 7, before going to Saraland, defendant shaved his head; three witnesses testified that he was bald during his stay in Alabama. 6 Lastly, five witnesses testified that defendant had never worn earrings of any kind.

I. Expert Testimony on Eyewitness Identification

Defendant contends the court abused its discretion in excluding the testimony of an expert witness on the psychological factors that may affect the accuracy of eyewitness identification. Prior to trial the defense moved for an order admitting the testimony of Dr. Robert Shomer. (Evid. Code, § 402.) Dr. Shomer is a practicing psychologist and professor of psychology of almost 20 years’ experience. He has taught numerous courses on the psychology of perception, memory, and recall, and has spoken and written frequently on such topics in both medical and legal settings. He is conversant with the scientific literature on the psychology of eyewitness identification, has done experimental research on the subject himself, and has published articles on that research. He has qualified as an expert psychological witness in more than two dozen state and federal trials. The People do not question the witness’ qualifications.

At the hearing on the motion Dr. Shomer explained that he proposed to inform the jury of various psychological factors that may affect the reliability of eyewitness identification, and to “help to counter some common misconceptions” about the process. He noted first that all eyewitness identification begins with the observer’s initial perception of the event. The physical circumstances affecting that observation are generally known to laymen, such as lighting, distance, and duration. But psychological factors may also influence the accuracy of the perception: Dr. Shomer intended to review for the jury the results of certain experimental studies showing that perception may be affected by such factors as the observer’s state of mind, his expectations, his focus of attention at the time, the suddenness of the incident, the stressfulness of the situation, and differences in the race and/ or age of the observer and the observed. On the latter point he would have testified, for example, that there are substantial decreases in accuracy when the two persons are of different races or ages.

The next phase of the process is memory. Dr. Shomer intended to discuss with the jury the evidence showing that memory is not merely a passive recording event, producing an imperishable reproduction of the scene perceived; rather, it is both a selective and a constructive process, in which old elements fade and are lost while new elements—subsequent information or suggestions—are unconsciously interwoven into the overall recollection until the subject cannot distinguish one from the other. 7

The last step is retrieval. Dr. Shomer proposed to review the studies establishing that recall may be affected by such factors as the subject’s expectations, his suggestibility, the phrasing of the questions asked of him, and even the size and type of the photographs he is shown. For example, Dr. Shomer would have explained to the jury that witnesses who are asked to identify criminals in lineups or photo displays tend to find the experience psychologically unpleasant and wish to terminate it. Because of this self-induced pressure, such witnesses may subconsciously take a simple request to point out the offender if he is in the lineup and convert it into a demand that they find the face in the lineup that is the “most similar” to the offender; that alternative appears more legitimate to them than admitting they cannot identify anyone at all.

Turning to the case at bar, Dr. Shomer made it clear that he did not propose to offer an opinion that any particular witness at this trial was or was not mistaken in his or her identification of defendant. But he did intend to point out various psychological factors that could have affected that identification in the present case. Thus he emphasized that from the viewpoint of the witnesses the shooting of Esparza on a busy streetcorner was a sudden and unexpected event, occurring some distance away, and that because of parked and passing cars their observations were largely discontinuous. He also referred to the youth of certain of the witnesses, the words used in making the pretrial photographic identifications of defendant, and the ambiguity of those identifications. Dr. Shomer particularly noted the effect of the “cross-racial factor” in this case, emphasizing that the one witness who was certain that defendant, a black, was not the black man at the scene was herself a black (Waller); by contrast, two of the witnesses who positively identified defendant at trial as the assailant (Molinar and Iglesias) were of the same ethnic origin (Hispanic) as the victim.

Finally, Dr. Shomer intended to explain to the jury that empirical research has undermined a number of widespread lay beliefs about the psychology of eyewitness identification, e.g., that the accuracy of a witness’s recollection increases with his certainty, that accuracy is also improved by stress, that cross-racial factors are not significant, and that the reliability of an identification is unaffected by the presence of a weapon or violence at the scene.

On this showing, defendant offered the testimony of Dr. Shomer as an aid to the jurors in weighing the eyewitness identifications in this case. The People objected on the sole ground that to admit the testimony would “usurp the jury’s function,” citing People v. Johnson (1974) 38 Cal.App.3d 1, 6-7 [112 Cal.Rptr. 834], and similar decisions. 8 The trial court, conceding that it was the first time it had encountered this type of evidence, ruled the testimony inadmissible. The court declared that it “fully agreed” with the reasoning of Johnson-, observing that none of the prosecution witnesses have “psychological defects,” the court concluded that to allow Dr. Shomer to testify “would be invading the province of the jury.” Defendant protested that Dr. Shomer would not give an opinion on the credibility of any particular witness, but would simply provide the jurors with information to help them determine the accuracy of the various identifications put before them. The court stood by its ruling, adding the further grounds that it intended to give a standard instruction on discrepancies in testimony (CALJIC No. 2.21 (4th ed. 1979)), that expert testimony on eyewitness identification “maybe would have a tendency” to “maybe cause confusion in the jurors’ minds,” and that such testimony “is really not what I consider scientific enough at this point in time” to be admissible.

A.

The United States Supreme Court has recognized that “The vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification.” (United States v. Wade (1967) 388 U.S. 218, 228 [18 L.Ed.2d 1149, 1158, 87 S.Ct. 1926].) The court noted “the high incidence of miscarriage of justice” caused by such mistaken identifications, and warned that “the dangers for the suspect are particularly grave when the witness’ opportunity for observation was insubstantial, and thus his susceptibility to suggestion the greatest.” (Id. at pp. 228-229 [18 L.Ed.2d at pp. 1158-1159].)

Distinguished federal judges have echoed and amplified these warnings. Thus in Jackson v. Fogg (2d Cir. 1978) 589 F.2d 108, the court upheld an order vacating a robbery-murder conviction on habeas corpus because prelineup procedures were unduly suggestive and because the four eyewitnesses had only a brief opportunity to observe the gunman under stressful conditions and showed varying degrees of certainty in their identifications of the defendant. There was no other evidence connecting the defendant with the crime. Writing for a unanimous court, Judge Lumbard observed that “Centuries of experience in the administration of criminal justice have shown that convictions based solely on testimony that identifies a defendant previously unknown to the witness is highly suspect. Of all the various kinds of evidence it is the least reliable, especially where unsupported by corroborating evidence.” (Id. at p. 112.)

Some of the reasons for that unreliability were discussed by Judge (later Solicitor General) McCree in United States v. Russell (6th Cir. 1976) 532 F.2d 1063, 1066: “There is a great potential for misidentification when a witness identifies a stranger based solely upon a single brief observation, and this risk is increased when the observation was made at a time of stress or excitement. . . . [TJhis danger is inherent in every identification of this kind, . As Judge McCree noted, “This problem is important because of all the evidence that may be presented to a jury, a witness’ in-court statement that ‘he is the one’ is probably the most dramatic and persuasive. ” (Id. at p. 1067.)

The rule that the testimony of a single witness is sufficient to prove identity (see Evid. Code, § 411) is premised in part on the assumption that an eyewitness identification is generally reliable. Yet Judge Hufstedler has declared that premise to be “at best, highly dubious, given the extensive empirical evidence that eyewitness identifications are not reliable.” (United States v. Smith (9th Cir. 1977) 563 F.2d 1361, 1365 (conc. opn.).) And with his characteristic vigor, Chief Judge Bazelon has called on the courts to face up to the reliability problems of eyewitness identification, to inform themselves of the results of scientific studies of those problems, and to allow juries access to that information in aid of their factfinding tasks. (United States v. Brown (D.C. Cir. 1972) 461 F.2d 134, 145-146, fn. 1 (conc. & dis. opn.).) 9

In the dozen years since Judge Bazelon’s appeal, empirical studies of the psychological factors affecting eyewitness identification have proliferated, and reports of their results have appeared at an ever-accelerating pace in the professional literature of the behavioral and social sciences. No less than five treatises on the topic have recently been published, citing and discussing literally scores of studies on the pitfalls of such identification. (Eyewitness Testimony: Psychological Perspectives (Wells & Loftus edits. 1984) [hereinafter Eyewitness Testimony: Psychological Perspectives]; Evaluating Witness Evidence: Recent Psychological Research and New Perspectives (Lloyd-Bostock & Clifford edits. 1983); Sobel, Eyewitness Identification: Legal and Practical Problems (2d ed. 1983); Loftus, Eyewitness Testimony (1979); Yarmey, The Psychology of Eyewitness Testimony (1979); see also Johnson, Cross-Racial Identification Errors in Criminal Cases (1984) 69 Cornell L.Rev. 934 [hereinafter Cross-Racial Identification Errors]; Note, Did Your Eyes Deceive You? Expert Psychological Testimony on the Unreliability of Eyewitness Identification (1977) 29 Stan. L.Rev. 969 [hereinafter Expert Psychological Testimony].) Indeed, in 1984 two leading researchers estimated that on this topic “over 85% of the entire published literature has surfaced since 1978.” (Wells & Loftus, Eyewitness Research: Then and Now, in Eyewitness Testimony: Psychological Perspectives, p. 3.) The consistency of the results of these studies is impressive, and the courts can no longer remain oblivious to their implications for the administration of justice.

B.

A traditional way of bringing scientific information to the attention of the judicial system, of course, is by the testimony of expert witnesses. But when that testimony relates to psychological factors affecting the accuracy of eyewitness identification, the courts have shown reluctance to admit it: appellate decisions almost unanimously hold that rulings excluding such evidence do not constitute an abuse of discretion. (See, e.g., Com. v. Francis (1983) 390 Mass. 89 [453 N.E.2d 1204, 1207-1208], and cases cited.) We inquire whether that reluctance remains justified. 10

This court has not previously addressed the admissibility of expert testimony on eyewitness identification. Several opinions of the Court of Appeal discuss the issue, but on close analysis none appears satisfactory. The leading case in this state—followed by the trial court in the case at bar—is People v. Johnson, supra, 38 Cal.App.3d 1. There the sole evidence connecting the defendants with a robbery-murder at a liquor store was the eyewitness identification of the surviving robbery victims. The defense sought to discredit that identification on various grounds, including the offer of expert testimony by a psychologist as to the ability of eyewitnesses to accurately perceive, remember and relate, and the distorting effects of excitement and fear on those functions. The trial court excluded the testimony, and the Court of Appeal upheld the ruling on four grounds. None, however, is immune from criticism.

First, the opinion reasons that although Evidence Code section 780, subdivision (c), permits a witness to be impeached by discrediting his capacity to perceive, recollect or communicate, “it does not follow that a party has a right to impeach a witness by calling another witness to testify as to the former’s capacity.” (38 Cal.App.3d at p. 6.) The argument misses the point. The expert witness in Johnson—just as Dr. Shomer here—would not have testified that any particular prosecution witness lacked the capacity to perceive, remember and relate; rather, he would simply have informed the jury of certain psychological factors that may impair the accuracy of a typical eyewitness identification, including the emotions of excitement or fear, with supporting references to experimental studies of such phenomena. Such evidence falls well within the broad statutory description of “any matter that has any tendency in reason” to bear on the credibility of a witness. (Evid. Code, § 780.)

Second, the Johnson opinion states that Evidence Code section 801, subdivision (a), “limits expert testimony to subjects beyond the range of common experience” (38 Cal.App.3d at pp. 6-7). This paraphrase of the statutory scheme, however, is both incomplete and misleading. To begin with, by its terms section 801 applies only to expert testimony “in the form of an opinion.” If an expert testifies not as to his opinion but as to facts within his special knowledge, section 801 is inapplicable. Factual testimony by an expert is admissible if it complies with the general statutory requirements that the witness be “qualified” by his special knowledge (Evid. Code, § 720) and that his evidence be relevant to the issues (id., § 351). 11 Much of the proposed testimony of the psychologist in Johnson and of Dr. Shomer here would have related primarily to matters of fact: the contents of eyewitness identification studies reported in the professional literature—their methodology, their data, and their findings—are facts, verifiable by anyone who can read and understand the studies in question. 12

In any event, to the extent these cases may involve opinion testimony by the psychologist witness, the Johnson paraphrasing of section 801, subdivision (a), errs by omission. The statute does not flatly limit expert opinion testimony to subjects “beyond common experience”; rather, it limits such testimony to such subjects “sufficiently beyond common experience that the opinion of an expert would assist the trier of fact” (italics added). The emphasized words, omitted by the Johnson court, make it clear that the admissibility of expert opinion is a question of degree. The jury need not be wholly ignorant of the subject matter of the opinion in order to justify its admission; if that were the test, little expert opinion testimony would ever be heard. Instead, the statute declares that even if the jury has some knowledge of the matter, expert opinion may be admitted whenever it would “assist” the jury. It will be excluded only when it would add nothing at all to the jury’s common fund of information, i.e., when “the subject of inquiry is one of such common knowledge that men of ordinary education could reach a conclusion as intelligently as the witness” (People v. Cole (1956) 47 Cal.2d 99, 103 [301 P.2d 854, 56 A.L.R.2d 1435]).

We apply this test to expert testimony on eyewitness identification. It is doubtless true that from personal experience and intuition all jurors know that an eyewitness identification can be mistaken, and also know the more obvious factors that can affect its accuracy, such as lighting, distance, and duration. 13 It appears from the professional literature, however, that other factors bearing on eyewitness identification may be known only to some jurors, or may be imperfectly understood by many, or may be contrary to the intuitive beliefs of most. For example, in the case at bar Dr. Shomer would have testified to the results of studies of relevant factors that appear to be either not widely known to laypersons or not fully appreciated by them, such as the effects on perception of an eyewitness’ personal or cultural expectations or beliefs (see Loftus, Eyewitness Testimony (1979) pp. 36-48), the effects on memory of the witness’ exposure to subsequent information or suggestions (id. at pp. 54-87), and the effects on recall of bias or cues in identification procedures or methods of questioning (id. at pp. 89-99; see generally Hall et al., Postevent Information and Changes in Recollection for a Natural Event, in Eyewitness Testimony: Psychological Perspectives, pp. 124-141).

Dr. Shomer would also have explained to the jury the pitfalls of cross-racial identification, evidently an important factor on the record in this case. To be sure, many jurors are likely to have some awareness of the fact that an eyewitness is more accurate in identifying a person of his own race than one of another race. (See, e.g., People v. Dixon (1980) 87 Ill.App.3d 814 [410 N.E.2d 252, 256].) But it appears that few jurors realize the pervasive and even paradoxical nature of this “own-race effect,” information that has emerged from numerous empirical studies of the question. These studies establish that the effect is strongest when white witnesses attempt to recognize black subjects; in such circumstances “The impairment in ability to recognize black faces is substantial.” (Cross-Racial Identification Errors, pp. 938-939.) 14 In laboratory experiments, for example, it is common for own-race/other-race recognition rates to differ by as much as 30 percent. (Id. at pp. 942-943.) The studies also reveal two aspects of the matter that will probably be contrary to most jurors’ intuitions: first, that white witnesses who are not racially prejudiced are just as likely to be mistaken in making a cross-racial identification as those who are prejudiced; and second, that white witnesses who have had considerable social contact with blacks may be no better at identifying them than those who have not. (Id. at pp. 943-944.) Finally, some jurors may deny the existence of the own-race effect in the misguided belief that it is merely a racist myth exemplified by the derogatory remark, “they all look alike to me,” while others may believe in the reality of this effect but be reluctant to discuss it in deliberations for fear of being seen as bigots. (Id. at p. 969; see also Wells, A Reanalysis of the Expert Testimony Issue, in Eyewitness Testimony: Psychological Perspectives, p. 309.)

In addition to the foregoing counterintuitive aspects of the own-race effect, other psychological factors have been examined in the literature that appear to contradict the expectations of the average juror. Perhaps the foremost among these is the lack of correlation between the degree of confidence an eyewitness expresses in his identification and the accuracy of that identification. Numerous investigations of this phenomenon have been conducted: the majority of recent studies have found no statistically significant correlation between confidence and accuracy, and in a number of instances the correlation is negative—i.e., the more certain the witness, the more likely he is mistaken. (Wells & Murray, Eyewitness Confidence, in Eyewitness Testimony: Psychological Perspectives, pp. 159-162.) Indeed, the closer a study comes to reproducing the circumstances of an actual criminal investigation, the lower is that correlation (id. at pp. 162-165), leading the cited authors to conclude that “the eyewitness accuracy-confidence relationship is weak under good laboratory conditions and functionally useless in forensically representative settings.” (Id. at p. 165; see also Deffenbacher, Eyewitness Accuracy and Confidence: Can We Infer Anything About Their Relationship? (1980) 4 Law & Human Behav. 243.) The average juror, however, remains unaware of these findings: “A number of researchers using a variety of methods have found that people intuitively believe that eyewitness confidence is a valid predictor of eyewitness accuracy.” (Wells & Murray, supra, at p. 159, citing five recent studies.)

We conclude that although jurors may not be totally unaware of the foregoing psychological factors bearing on eyewitness identification, the body of information now available on these matters is “sufficiently beyond common experience” that in appropriate cases expert opinion thereon could at least “assist the trier of fact” (Evid. Code, § 801, subd. (a)). 15

The third ground of the Johnson opinion is premised on its observation (38 Cal.App.3d at p. 7) that “In cases not involving sex offenses California courts usually reject attempts to impeach a witness by means of psychiatric testimony,” citing Ballard v. Superior Court (1966) 64 Cal.2d 159, 172 [49 Cal.Rptr. 302, 410 P.2d 838, 18 A.L.R.3d 1416]. Ballard held (at pp. 173-175) that in a sex offense case expert medical testimony of a psychiatrist may be admitted to impeach the credibility of the complaining witness by showing that she suffers from a particular physical or mental illness or disorder that impairs her ability to tell the truth. (See also People v. Russel (1968) 69 Cal.2d 187, 193, 195-196 [70 Cal.Rptr. 210, 443 P.2d 794].) The Johnson opinion then reasons (38 Cal.App.3d at p. 7) that “The present occurrence was frightening but hardly deranging. There is no evidence or claim of emotional disturbance or psychological ‘abnormality’ of any of the prosecution witnesses.”

Again the argument misses the point. In neither Johnson nor the case at bar did the defense offer expert medical testimony by a psychiatrist to attack the truth-telling ability of any witness. The Ballard rule therefore has nothing to do with the case, and it is totally irrelevant that these are “cases not involving sex offenses” or that no witness was shown to be “abnormal.”

Fourth, the Johnson opinion (ibid.) also upholds the trial court’s ruling on the ground that to admit expert psychological evidence on eyewitness identification would “take over the jury’s task of determining the weight and credibility of the witness’ testimony”—or, to put it in the more colorful language of legal cliché used by many courts, would “invade the province” or “usurp the function” of the jury, because such evidence “embraces the ultimate issue.” As Dean Wigmore has said, however, such language “is so misleading, as well as so unsound, that it should be entirely repudiated. It is a mere bit of empty rhetoric,” and “remains simply one of those impracticable and misconceived utterances which lack any justification in principle.” (Fns. omitted.) (7 Wigmore on Evidence (Chadbourn rev. ed. 1978), § 1920, p. 18, § 1921, p. 22.) Specifically referring to expert psychological evidence on eyewitness identification, the author of a leading treatise on the topic asserts that “the objection based upon the ‘province of the jury’ is no more than a shibboleth which, if accepted, would deprive the jury of important information useful and perhaps necessary for a proper decision on a difficult issue.” (Wall, Eye-Witness Identification in Criminal Cases (1965) p. 213.)

The reasons for these criticisms are several. The expert testimony in question does not seek to take over the jury’s task of judging credibility: as explained above, it does not tell the jury that any particular witness is or is not truthful or accurate in his identification of the defendant. Rather, it informs the jury of certain factors that may affect such an identification in a typical case; and to the extent that it may refer to the particular circum stances of the identification before the jury, such testimony is limited to explaining the potential effects of those circumstances on the powers of observation and recollection of a typical eyewitness. The jurors retain both the power and the duty to judge the credibility and weight of all testimony in the case, as they are told by a standard instruction. 16

Nor could such testimony in fact usurp the jury’s function. As is true of all expert testimony, the jury remains free to reject it entirely after considering the expert’s opinion, reasons, qualifications, and credibility. Indeed, the Penal Code commands (§ 1127b) that an instruction so informing the jury be given in any criminal trial in which expert opinion evidence is received. 17

Finally, California has abandoned the “ultimate issue” rule in any event: “in this state we have followed the modern tendency and have refused to hold that expert opinion is inadmissible merely because it coincides with an ultimate issue of fact.” (People v. Cole, (1956) supra, 47 Cal.2d 99, 105, and cases cited.) Evidence Code section 805 codifies this case law, declaring that “Testimony in the form of an opinion that is otherwise admissible is not objectionable because it embraces the ultimate issue to be decided by the trier of fact.” 18

C.

For the reasons stated, the challenged ruling excluding the expert testimony of Dr. Shomer is not supported by the trial court’s professed “agreement” with the reasoning of Johnson, or by its observation that the prosecution witnesses have no “psychological defects” (i.e., the Ballard rule), or by its flat assertion that such expert testimony would “invade the province” of the jury.

Nor is the ruling supported by the court’s announced intent to give the standard instruction on discrepancies in testimony. (CALJIC No. 2.21 (4th ed. 1979).) The instruction contains only a few general remarks on the topic; 19 it does not even begin to convey to the jury the specific data on the eyewitness identification process that Dr. Shomer’s testimony would have provided, a task that in any event is beyond the function of instructions. (Pen. Code, § 1127 [“Either party may present to the court any written charge on the law, but not with respect to matters of fact”].) Nor, again, is the ruling supported by the court’s speculation that this testimony might tend to confuse the jurors. Evidence that is relevant to the prime theory of the defense cannot be excluded in wholesale fashion merely because the trial would be simpler without it. Rather, it should be accompanied by instructions clearly explaining to the jury the purpose for which it is introduced. As noted above (fn. 11, ante), any excess in the quantity or complexity of such testimony can be controlled by the court’s power to limit the presentation of evidence.

Lastly, the ruling is not supported by the court’s opinion that expert testimony on eyewitness identification is not yet “scientific enough” to be admissible. The court illustrated its point by referring to the general rule excluding lie-detector evidence. (See, e.g., People v. Wochnick (1950) 98 Cal.App.2d 124, 127-128 [219 P.2d 170].) The choice of words and example makes it clear the court was implicitly invoking the Kelly-Frye rule, i.e., the rule that evidence based on a new scientific method of proof is admissible only on a showing that the procedure has been generally accepted as reliable in the scientific community in which it developed. (People v. Kelly (1976) 17 Cal.3d 24 [130 Cal.Rptr. 144, 549 P.2d 1240]; Frye v. United States (D.C. Cir. 1923) 293 F. 1013 [34 A.L.R. 145].)

We are not persuaded, however, that the Kelly-Frye rule applies to expert testimony on eyewitness identification. (See Cross-Racial Identification Errors, p. 971; Expert Psychological Testimony, pp. 1021-1023.) It is important to distinguish in this regard between expert testimony and scientific evidence. When a witness gives his personal opinion on the stand—even if he qualifies as an expert—the jurors may temper their acceptance of his testimony with a healthy skepticism born of their knowledge that all human beings are fallible. But the opposite may be true when the evidence is produced by a machine: like many laypersons, jurors tend to ascribe an inordinately high degree of certainty to proof derived from an apparently “scientific” mechanism, instrument, or procedure. Yet the aura of infallibility that often surrounds such evidence may well conceal the fact that it remains experimental and tentative. (People v. Kelly, supra, at p. 32, and cases cited.) For this reason, courts have invoked the Kelly-Frye rule primarily in cases involving novel devices or processes such as lie detectors, “truth serum,” Nalline testing, experimental systems of blood typing, “voice-prints,” identification by human bite marks, microscopic analysis of gunshot residue, and hypnosis (People v. Shirley (1982) 31 Cal.3d 18, 51-54 [181 Cal.Rptr. 243, 641 P.2d 775], and cases cited), and, most recently, proof of guilt by “rape trauma syndrome” (People v. Bledsoe (1984) 36 Cal.3d 236, 246-251 [203 Cal.Rptr. 450, 681 P.2d 291]). In some instances the evidence passed the Kelly-Frye test, in others it failed; but in all such cases “the rule serves its salutary purpose of preventing the jury from being misled by unproven and ultimately unsound scientific methods.” (Shirley, supra, at p. 53.)

Here, by contrast, no such methods are in issue. We have never applied the Kelly-Frye rule to expert medical testimony, even when the witness is a psychiatrist and the subject matter is as esoteric as the reconstitution of a past state of mind or the prediction of future dangerousness, or even the diagnosis of an unusual form of mental illness not listed in the diagnostic manual of the American Psychiatric Association (People v. Phillips (1981) 122 Cal.App.3d 69, 86-87 [175 Cal.Rptr. 703] (“Munchausen’s syndrome by proxy”)). We see no reason to require a greater foundation when the witness is a qualified psychologist who will simply explain to the jury how certain aspects of everyday experience shown by the record can affect human perception and memory, and through them, the accuracy of eyewitness identification testimony. Indeed, it would be ironic to exclude such testimony on Kelly-Frye grounds on the theory that jurors tend to be unduly impressed by it, when jurors are far more likely to be unduly impressed by the eyewitness testimony itself.

D.

It remains true, of course, that “Where expert opinion evidence is offered, much must be left to the discretion of the trial court” (People v. Cole, supra, 47 Cal.2d 99, 105). Yet that discretion is not absolute: in various contexts it has been held that trial courts committed reversible error in excluding expert testimony. (E.g., Brown v. Colm (1974) 11 Cal.3d 639, 647 [114 Cal.Rptr. 128, 522 P.2d 688]; People ex rel. Dept. Pub. Wks. v. Douglas (1971) 15 Cal.App.3d 814, 820-822 [93 Cal.Rptr. 644]; Varas v. Barco Mfg. Co. (1962) 205 Cal.App.2d 246, 259-261 [22 Cal.Rptr. 737]; Reynolds v. Natural Gas Equipment, Inc. (1960) 184 Cal.App.2d 724, 739-740 [7 Cal.Rptr. 879]; Burch v. Valley Motor Lines, Inc. (1947) 78 Cal.App.2d 834, 840-844 [179 P.2d 47].) A recent decision of the Arizona Supreme Court illustrates the kind of case in which the exclusion of expert testimony on psychological factors affecting eyewitness identification is a prejudicial abuse of discretion. (State v. Chapple (1983) 135 Ariz. 281 [660 P.2d 1208].) There the defendant was convicted of three counts of first degree murder. No direct or circumstantial evidence of any kind connected him with the crime other than the testimony of two eyewitnesses, and neither eyewitness had ever seen him before the events in question. The witnesses observed the crimes being committed, inter alia, by a man who had been introduced to them as “Dee.” A few days later they were shown groups of photographs of acquaintances of a known participant. One of the witnesses tentatively identified a certain James Logan as “Dee”; although the defendant’s photograph was in another of the displays, the witness did not identify it. More than a year later the witness was shown a further display that included a photograph of the defendant but none of Logan, and he identified the defendant as “Dee”; he was then shown the photograph of the defendant that he had previously failed to identify, but said he had no recollection of having seen it before. The other eyewitness then identified the defendant as “Dee” from the later display, and both witnesses positively identified the defendant at trial.

The defendant presented seven witnesses who testified he was in another state on the date of the crimes. He also called psychologist Elizabeth Loftus, a nationally known expert on perception and memory, to testify to various psychological factors shown by the record that can affect the accuracy of eyewitness identification. 20 The trial court excluded this testimony on the grounds that the information it sought to convey was within the common knowledge of the jurors and would be covered in cross-examination and closing arguments.

The Arizona Supreme Court recognized that the admissibility of such testimony was in the first instance a matter of trial court discretion, but held that such discretion was abused on the record before it. The court began by rejecting claims that the testimony of Dr. Loftus would have “invaded the province of the jury,” and pointed to an Arizona rule of evidence identical to our Evidence Code section 805. (Id. at p. 1219.) It further rejected claims that Dr. Loftus’s testimony would have been prejudicial because the jury might have given it undue weight, and that such testimony lacked probative value because it would not have included an opinion on whether the particular eyewitnesses in the case at hand were in fact mistaken in their identification. {Ibid.)

Turning to the key question of whether Dr. Loftus’s testimony was a proper subject for expert opinion, the court summarized that testimony (see fn. 20, ante) and declined to assume that the average juror would be aware of the information it contained. {Id. at p. 1221.) After reviewing the record, the court found that the testimony would have assisted the jury in deciding specific factual issues in the case. {Id. at pp. 1222-1223.) Finally, the court emphasized that the facts were close, the accuracy of the eyewitness identifications was a crucial issue, and the exclusion of Dr. Loftus’s testimony “undercut the entire evidentiary basis” of the defendant’s arguments on that issue. {Id. at p. 1222.) 21 The court concluded (at p. 1224), “there were a number of substantive issues of ultimate fact on which the expert’s testimony would have been of significant assistance. Accordingly, we hold that the order precluding the testimony was legally incorrect and was unsupported by the record. It was, therefore, an abuse of discretion.” Holding the error to be prejudicial on the facts, the court reversed the judgment.

In the case at bar a similar analysis leads to a like result. Here, too, the expert witness was undoubtedly qualified to testify on the particular matters he proposed to address. 22 Like the Arizona Supreme Court, we decline to assume that the subject matter of Dr. Shomer’s testimony would have been fully known to the jurors; rather, the professional literature persuades us to the contrary. Also as in Chappie, the record establishes that Dr. Shomer’s testimony would have been of significant assistance to the jury. Because no other evidence connected defendant with the crime, the crucial factor in the case was the accuracy of the eyewitness identifications. Yet on that issue the evidence was far from clear. As we noted at the outset, in the testimony of each of the witnesses who identified defendant in the courtroom there were elements that could have raised reasonable doubts as to the accuracy of the identification. These elements included the suddenness and unexpectedness of the event, discontinuity and other difficulty of observation, fear and other stress at the time of perception, overestimation of the duration of the event, “feedback” factors following the event, failure or uncertainty of several witnesses in selecting defendant’s photograph from police displays, and, particularly important, apparent cross-racial identification discrepancies. Further doubts could have arisen from the dramatic declaration in open court by a prosecution eyewitness that defendant was not the perpetrator, and from the testimony of six witnesses that defendant was not in the state on the day the crime was committed.

In these circumstances the exclusion of Dr. Shomer’s testimony undercut the evidentiary basis of defendant’s main line of defense—his attack on the accuracy of the eyewitness identifications—and deprived the jurors of information that could have assisted them in resolving that crucial issue. The ruling excluding such testimony was therefore unsupported by the record. We have previously shown (Part I C, ante) that it was unsupported by the law. It follows that the ruling constituted an abuse of discretion.

An error in excluding expert testimony may be found harmless. (E.g., Majetich v. Westin (1969) 276 Cal.App.2d 216, 218-219 [80 Cal.Rptr. 787].) In the case at bar, however, the record compels us to conclude that the error was prejudicial. As we have seen, the issue affected by the ruling was crucial, given the absence of any other evidence connecting defendant with the crime; and the evidence on that issue was close, given the potential weaknesses in the prosecution’s testimony and the presence of both eyewitness and alibi testimony favorable to the defense. 23 An error that impairs the jury’s determination of an issue that is both critical and closely balanced will rarely be harmless. Rather, after an examination of the whole record we find it reasonably probable that a result more favorable to defendant would have been reached in the absence of this error. (People v. Watson (1956) 46 Cal.2d 818, 836 [299 P.2d 243].) There has therefore been a miscarriage of justice, and the judgment must be reversed. (Cal. Const., art. VI, § 13.)

We reiterate that the decision to admit or exclude expert testimony on psychological factors affecting eyewitness identification remains primarily a matter within the trial court’s discretion; like the court in Chapple, “we do not intend to ‘open the gates’ to a flood of expert evidence on the subject.” (660 P.2d at p. 1224.) We expect that such evidence will not often be needed, and in the usual case the appellate court will continue to defer to the trial court’s discretion in this matter. 24 Yet deference is not abdication. When an eyewitness identification of the defendant is a key element of the prosecution’s case but is not substantially corroborated by evidence giving it independent reliability, and the defendant offers qualified expert testimony on specific psychological factors shown by the record that could have affected the accuracy of the identification but are not likely to be fully known to or understood by the jury, it will ordinarily be error to exclude that testimony.

II. Other Issues

Defendant makes other contentions relating to the guilt and penalty phases that are not likely to arise on retrial. Because of the unusual posture of the case after the jury’s verdict, however, we address certain contentions dealing with the crimes for which defendant may be prosecuted on such retrial.

A. Robbery

The jury acquitted defendant of the robbery charged in count II of the information. Penal Code section 1023 unequivocally provides that when a defendant is acquitted of an offense, that acquittal “is a bar to another prosecution for the offense charged . . ., or for an attempt to commit the same, or for an offense necessarily included therein, of which he might have been convicted under that accusatory pleading.” (See also Stone v. Superior Court (1982) 31 Cal.3d 503, 510 [183 Cal.Rptr. 647, 646 P.2d 809].) Thus, on retrial, double jeopardy considerations will bar any further prosecution of defendant for robbery or attempted robbery. 25

B. Special Circumstance

The determination that double jeopardy considerations prohibit retrial of defendant for robbery or attempted robbery also precludes the prosecution from retrying him on the robbery special circumstance allegation. Penal Code section 190.4, subdivision (a), provides in relevant part that “Whenever a special circumstance requires proof of the commission or attempted commission of a crime, such crime shall be charged and proved pursuant to the general law applying to the trial and conviction of the crime.” Future prosecution of defendant on the underlying robbery is barred by the double jeopardy clause because defendant has been acquitted of that crime and impliedly of any lesser included offenses; accordingly, the statutory prerequisite to a finding of robbery special circumstance, i.e., charging and proving a robbery or attempted robbery, cannot be met and defendant cannot be retried on that special circumstance.

We recognize that in People v. Robertson (1982) 33 Cal.3d 21, 47 [188 Cal.Rptr. 77, 655 P.2d 279], and People v. Velasquez (1980) 26 Cal.3d 425, 434, footnote 6 [162 Cal.Rptr. 306, 606 P.2d 341], we held that failure to separately charge the underlying felony was not prejudicial error. In both decisions, however, the focus was on prejudice, not on the error itself; both noted that the statutory requirement was not met and that the omission was in fact error. They held that the error was not prejudicial because the defendant was put on notice by the special circumstance allegation that he was required to defend against the underlying crime. The situation here is completely different. In Velasquez and Robertson there was no legal impediment to charging the underlying crime; by contrast, in the present case the prosecution may not recharge defendant with robbery or any lesser included offense without violating double jeopardy protections, and hence may not retry defendant on the special circumstance allegation predicated on that crime.

C. Murder

Defendant contends that the jury’s failure to specify the degree of murder in its verdict renders his conviction second degree murder by operation of law. (Pen. Code, § 1157.) Although the issue is not likely to arise on retrial in this precise factual form, the effect of the jury’s action may well be to bar the prosecution of defendant for a crime greater than second degree murder because of double jeopardy considerations. For this reason we inquire into the matter in some detail.

Defendant was charged with the crime of murder in the usual manner, i.e., without specification of degree. (Pen. Code, § 187; see also § 951.) The jury was instructed that “Before you may return a verdict in this case, you must agree unanimously not only as to whether the defendant is guilty or not guilty, but also, if you should find him guilty of an unlawful killing, you must agree unanimously as to whether he is guilty of murder of the first degree. ” (Italics added; CALJIC No. 8.74 (1976 rev.).) The jury was also instructed with respect to the special circumstance that “If you find the defendant guilty of murder in the first degree, you must then determine if the murder was committed under the following special circumstance, while engaged in the commission or the attempted commission of a robbery.” (Italics added; CALJIC No. 8.80 (4th ed. 1979).) The jury returned the following verdict: “We, the jury in the above-entitled action, find the Defendant Eddie Bobby McDonald, guilty of MURDER, in Violation of Section 187 Penal Code, a felony, as charged in Count I of the information. ” (Italics added.) The jury was polled and the verdict was recorded and filed.

Three and a half weeks later, the jury was reconvened for the penalty phase. 26 At that time, the court submitted a new guilty verdict form to the jury and explained that because of inadvertence or mistake there had been an omission in the original verdict form. The new form added the phrase, “and we further find it to be murder of the first degree, to be true/not true.” The jury deliberated briefly and returned a finding of first degree murder on this form.

Defendant contends that because the jury failed to specify the degree of murder in its original verdict, the degree of the crime was fixed at second degree murder by operation of law. Penal Code section 1157 provides that “Whenever a defendant is convicted of a crime or attempt to commit a crime which is distinguished into degrees, the jury . . . must find the degree of the crime or attempted crime of which he is guilty. Upon the failure of the jury ... to so determine, the degree of the crime or attempted crime of which the defendant is guilty, shall be deemed to be of the lesser degree. ” (Italics added.) 27

Respondent insists that because the statute requires only that the jury “find” the degree of the crime and does not mandate any specific procedure for so doing, a determination of degree can be inferred from the jury’s separate finding on the special circumstance allegation. Respondent stresses that the jury was instructed to determine whether or not the special circumstance was true only if it found defendant guilty of first degree murder; because the jury found the special circumstance true, respondent reasons, it must have found first degree murder as well.

This precise contention has been rejected in a long line of decisions which require that the degree be explicitly specified by the verdict. (People v. Dixon (1979) 24 Cal.3d 43, 51-52 [154 Cal.Rptr. 236, 592 P.2d 752]; People v. Flores (1974) 12 Cal.3d 85, 94-95 [115 Cal.Rptr. 225, 524, P.2d 353]; People v. Beamon (1973) 8 Cal.3d 625, 629, fn. 2 [105 Cal.Rptr. 681, 504 P.2d 905]; People v. Thomas (1978) 84 Cal.App.3d 281, 285 [148 Cal.Rptr. 532]; People v. Baeske (1976) 58 Cal.App.3d 775, 778, fn. 1 [130 Cal.Rptr. 35]; People v. Doran (1974) 36 Cal.App.3d 592 [111 Cal.Rptr. 793]; People v. Cox (1973) 33 Cal.App.3d 378, 381-382 [109 Cal.Rptr. 43]; People v. Fernandez (1963) 222 Cal.App.2d 760, 769 [35 Cal.Rptr. 370]; People v. Hughes (1959) 171 Cal.App.2d 362, 369-370 [340 P.2d 679]; see also In re Candelario (1970) 3 Cal.3d 702, 706, fn. 2 [91 Cal.Rptr. 497, 477 P.2d 729].)

In Beamon, for example, the defendant was convicted of robbery; an allegation that the defendant was armed with a deadly weapon at the time of the commission of the crime was found to be true. The jury, however, failed to fix the degree of the crime. We held that despite the jury’s finding on the arming allegation, its failure to specify the degree of the crime required that the conviction be deemed to be of the second degree. (8 Cal.3d at p. 629, fn. 2.) “We cannot assume, contrary to the clear legislative direction, that because a factual finding was made which would have war ranted a determination of first degree robbery, the jury unmistakably intended [citation] to make that determination when it refrained from expressly fixing the degree. ” {Ibid.', italics added.)

Further support for defendant’s position on this point is found in our opinion in Dixon, in which we comprehensively reviewed the historical development of section 1157. As we noted in Dixon, throughout its history the statute has been applied in cases in which the “failure” of the jury to determine the degree of crime “consisted in an omission to perform that function because of mistake or inadvertence or in circumstances suggesting an intended act of leniency.” (24 Cal.3d at pp. 51-52.) The rule is thus firmly established that the statute applies whenever the jury neglects to explicitly specify the degree of the crime. 28

People v. Hughes, supra, 171 Cal.App.2d 362, presents a factual situation almost identical to that before us. In Hughes, the defendant was charged with first degree murder in a capital case. The verdict form returned by the jury found him guilty of murder “as charged in the information”; this verdict was received and entered in the minutes. The jury was then released and told to return the next morning to consider the issue of penalty. At the outset of the penalty phase, the court inquired of the jury whether it understood that “guilty as charged in the information” referred to a charge of murder in the first degree; the foreman responded affirmatively. After evidence was taken at the penalty phase, the court informed the jurors that because of technical legal requirements the verdict fixing the degree had to be in writing. It therefore submitted a supplemental verdict form as to the degree of the crime, and the jury fixed the degree at first degree murder.

The appellate court held that the second verdict was invalid because it essentially constituted a resubmission of the issue of degree to the jury. (171 Cal.App.2d at p. 369.) 29 The court concluded that after the original verdict “had been received and the jury had been released that verdict under the provisions of the code section was a verdict of second degree murder. Of that crime and of that crime only has appellant been convicted and by that verdict he had been acquitted of first degree murder. ... So far as the jury was concerned it ended the trial on the issue of guilt. That being so, it results that all proceedings thereafter were nullities. There was no issue of penalty for the jury to determine. . . . Therefore the trial of the appellant was complete and the court had no jurisdiction to recall the jury for further proceedings.” {Id. at p. 370.)

These decisions illustrate the rule that the statute applies to reduce the degree even in situations in which the jury’s intent to convict of the greater degree is demonstrated by its other actions, i.e., by signing a subsequent verdict form (Hughes) or making a finding on an enhancement {Beamon). Contrary to respondent’s assertion, the key is not whether the “true intent” of the jury can be gleaned from circumstances outside the verdict form itself; instead, application of the statute turns only on whether the jury specified the degree in the verdict form. In the present case the verdict form failed to specify the degree; in the absence of such specification, the jury’s finding on the special circumstance allegation is irrelevant and the conviction must be deemed second degree murder as a matter of law pursuant to the unambiguous language of section 1157.

Respondent contends further that because the jury was instructed solely on first degree murder, any verdict of guilt on the murder charge could only be in the first degree. The jury was instructed that before it could return a verdict of guilt on the murder charge, it must unanimously agree on whether defendant was guilty of murder of the first degree. Thus, respondent submits, the jury’s verdict of guilty of murder “as charged” constituted an implied finding of first degree murder.

While respondent is correct that the jury was not instructed on the lesser included offense of second degree murder, 30 we see no reason why this variation in the facts should lead to a different result. First, the terms of the statute are unambiguous. No special exception is created for the situation presented by this case; had the Legislature chosen to make section 1157 inapplicable to cases in which the jury was instructed on only one degree of a crime, it could easily have so provided. The statute requires that “if the jury shall find the defendant guilty, the verdict shall specify the degree of murder .... It establishes a rule to which there is to be no exception, and the Courts have no authority to create an exception when the statute makes none.” (People v. Campbell (1870) 40 Cal. 129, 138.)

Furthermore, prior applications of the statute suggest no rationale for excepting this case from the plain language of section 1157. As we have noted, this is not the first case in which the statute compels the court to deem the crime to be of the lesser degree despite indications that the jury’s failure to specify degree was not intentional but resulted from mistake or inadvertence. (See, e.g., People v. Hughes, supra, 171 Cal.App.2d at pp. 369-370.)

In fact, this court in Campbell was faced with a dilemma similar to that which respondent asserts exists in the present case. In Campbell, the People claimed that because the facts alleged in the indictment would support only a conviction of first degree and not of second degree murder, the failure of the jury to specify the degree did not require reversal. The court rejected this contention, stating that “We have no right to disregard a positive requirement of the statute, as it is not our province to make laws, but to expound them.” (40 Cal. at p. 138.) In interpreting the statutory provision which then required that the jury “designate” (rather than the equivalent current term “find”) the degree of the crime, the court stated: “The word ‘designate,’ as here employed, does not imply that it will be sufficient for the jury to intimate or give some vague hint as to the degree of murder of which the defendant is found guilty; but it is equivalent to the words ‘express’ or ‘declare,’ and it was evidently intended that the jury should expressly state the degree of murder in the verdict so that nothing should be left to implication on that point. . . . [T]he very letter of the statute . . . requires the jury to ‘designate,’ or in other words, to express or declare by their verdict the degree of the crime. However absurd it may, at the first blush, appear to be to require the jury to designate the degree of the crime, when it appears on the face of the indictment that the offense charged has but one degree, there are plausible and, perhaps, very sound reasons for this requirement. . . . But whatever may have been the reasons for this enactment, it is sufficient for the Courts to know that the law is so written and it is their duty to enforce it.” (Id. at pp. 139-140.)

Respondent’s attempt to distinguish the present case on this basis must therefore fail, and it must be deemed as a matter of law that defendant was convicted of second degree murder. (See also People v. Johns (1983) 145 Cal.App.3d 281, 294-295 [193 Cal.Rptr. 182].) 31

The judgment is reversed.

Bird, C. J., Kaus, J., Broussard, J., Reynoso, J., and Grodin, J., concurred.

On December 20, 1984, the opinion was modified to read as printed above.

1 He added that during the encounter the gunman was holding a “long black object,” and that after reaching underneath the victim he put something in his pocket. Mr. Soderholm also followed the gunman to the next corner and saw him drive away, apparently alone, in a car that was parked there.

2 Asked what it was about the defendant that appeared “similar” to the gunman, Mr. Malone emphasized, “I know his race. He was Negro.” Pressed further, he also mentioned generally the defendant’s complexion, build, and age.

3 The photograph referred to was the same one (No. 3) that the other witnesses had selected.

4 The relevant preliminary hearing testimony was as follows:

“Q. [by the prosecutor]. When you first looked over and saw these two men, do you think that you would recognize either of them if you saw them again? A. Yes, I would.

“Q. Would you look around the courtroom and tell us if you see either of those two men in court today? A. No, I don’t.

“Q. Could you describe either of those two men for us? A. Okay. The man that was shot, I really didn’t get too good a look at him because he was on the ground at the time that I looked.

“Q. All right. A. But the man that was doing the shooting looked directly into my face, and I got a very good view of him. And the man is not here today.”

The witness reiterated that testimony a few minutes later.

5 At the preliminary hearing Mrs. Waller further explained that “His complexion was darker than mine . ...” As will appear, in a pretrial hearing on the admissibility of expert testimony on eyewitness identification, it was brought out that Mrs. Waller is also black.

6 According to the prosecution witnesses, the black man who shot Esparza had a moderate Afro hairdo.

7 We took note of studies demonstrating this phenomenon in People v. Shirley (1982) 31 Cal.3d 18, 57-62 [181 Cal.Rptr. 243, 641 P.2d 775],

8 People v. Guzman (1975) 47 Cal.App.3d 380, 385-386 [124 Cal.Rptr. 492]; People v. Brooks (1975) 51 Cal.App.3d 602, 608-609 [124 Cal.Rptr. 492]; see also People v. Bradley (1981) 115 Cal.App.3d 744, 751-752 [171 Cal.Rptr. 487],

9 Thus Judge Bazelon pointed out that “One critical problem [of eyewitness identifications] concerns their reliability, yet courts regularly protest their lack of interest in the reliability of identifications, as opposed to the suggestivity that may have prompted them, arguing that reliability is simply a question of fact for the jury. [Citation.] There already exists, however, great doubts—if not firm evidence—about the adequacy and accuracy of the process. Unquestionably, identifications are often unreliable—perhaps consistently less reliable than lie detector tests, which we have in the past excluded for unreliability.” (Id. at p. 145, fn. 1.) Judge Bazelon continued, “we need more information about the reliability of the identification process and about the jury’s ability to cope with its responsibility. For it should be obvious that we cannot strike a reasonable and intelligent balance if we take pains to remain in ignorance of the pitfalls of the identification process. The empirical data now available indicates that the problem is far from fanciful. [Citations.] But for a variety of reasons we have been unwilling to face up to the doubts to which this data gives rise. And despite repeated charges and counter-charges concerning the accuracy of inter-racial identifications, we have developed a reluctance that is almost a taboo [citation] against even acknowledging the question, much less providing the jury with all of the available information.” (Id. at pp. 145-146, fn. 1.) And Judge Bazelon concluded that “More information is needed to assist the jury’s resolution of identification issues,” and “our doubts will not disappear merely because we run away from the problem.” (Id. at p. 146, fn. 1.) (See also Bazelon, Eyewitness News (Mar. 1980) Psychology Today, p. 101.)

10 Of course, the virtual unanimity of appellate decisions on the topic may well be misleading. Expert testimony on eyewitness identification is usually offered by the defendant. In cases in which the testimony is admitted, the issue will not arise on appeal: if the defendant is convicted, he cannot complain of the admission of his own evidence; and if he is acquitted, no appeal is possible in any event. It follows that appellate courts ordinarily confront the issue only when the testimony has been excluded-, and in all such cases appellate courts tend to affirm, because of the deference traditionally accorded to discretionary rulings of trial courts. Nevertheless, in a number of published opinions it emerges in various contexts that expert testimony on eyewitness identification was in fact admitted at the trial. (E.g., United States v. Booth (9th Cir. 1981) 669 F.2d 1231, 1240; People v. Brown (1982) 110 Ill.App.3d 1125 [443 N.E.2d 665, 668]; State v. Chapman (La. 1981) 410 So.2d 689, 702; State v. Sellars (1981) 52 N.C.App. 380 [278 S.E.2d 907, 921-922]; Hampton v. State (1979) 92 Wis.2d 450 [285 N.W.2d 868, 870-871].) In State v. Warren (1981) 230 Kan. 385 [635 P.2d 1236, 1243, 23 A.L.R.4th 1070], the court noted that Dr. Elizabeth Loftus presented an affidavit stating that she had been allowed to testify as an eyewitness identification expert in more than 34 cases in various jurisdictions, and that another nationally recognized expert, Dr. Robert Buckhout, had so testified in more than 20 cases.

11 The testimony also remains subject to the court’s general discretionary power to exclude evidence that is unduly time-consuming or confusing (Evid. Code, § 352), and to its special discretionary power to limit the? number of expert witnesses called by any party (id., § 723).

12 To illustrate, assume the witness testifies that journal A published an article B in which researcher C reported that he conducted an empirical study of the effect of factor D on the accuracy of eyewitness identification, that he designed the experiment in manner E, that the experiment produced raw data F, that he analyzed those data by statistical method G, and that such analysis yielded finding H; in that event, A, B, C, D, E, F, G, and H are facts, not opinions, and in relating them to the jury the expert witness is not testifying “in the form of an opinion.” (Evid. Code, § 801, subd. (a).) By contrast, if the same expert goes on to assert, on the basis of these facts, that a particular eyewitness in the case before him was or was not mistaken in his identification of the defendant, that assertion would be opinion testimony. As noted above, however, no such testimony was offered in Johnson or the case at bar.

13 Even with respect to these factors, however, expert psychological testimony may be helpful in appropriate cases. For example, the length of time that an eyewitness observes the person he later identifies is often given significant weight by the jury. But in virtually every such case the only evidence of that duration is the witness’s own estimate. Studies show that witnesses consistently overestimate the length of brief periods of time, especially in the presence of stressful stimuli: “during sudden, action-packed events such as crimes, people almost always overestimate the length of time involved because the flurry of activity leads them to conclude that a significant amount of time has passed.” (Expert Psychological Testimony, p. 977; see also Schiffman & Bobko, Effects of Stimulus Complexity on the Perception of Brief Temporal Durations (1974) 103 J. Experimental Psychology 156.) In the case at bar, for example, the eyewitnesses’ estimates of the duration of the crime ranged widely from 2 minutes to 25 minutes.

14 A recent overview of the literature cites no less than 10 studies documenting this impairment. (I d. at pp. 938-939, fn. 18.)

15 We realize there is a minority view on this question: two psychologists are on record as opposing the use of expert testimony on the factors affecting eyewitness identification. They argue that for most of these factors, the claimed effect on witness accuracy either is not proved or is probably obvious to jurors. (Egeth & McCloskey, Expert Testimony About Eyewitness Behavior: Is It Safe and Effective? in Expert Testimony: Psychological Perspectives, p. 283; see also Egeth & McCloskey, Eyewitness Identification: What Can a Psychologist Tell a Jury? (1983) 38 Am.Psychologist 550.) Their reasoning, however, has been vigorously disputed by their peers. (E.g., Wells, A Reanalysis of the Expert Testimony Issue, in Eyewitness Testimony: Psychological Perspectives, p. 304; Loftus, Silence is Not Golden (1983) 38 Am.Psychologist 564.) And on close examination it appears the principal complaint of Egeth and McCloskey is not so much that expert testimony on eyewitness identification should never be admissible, as that it is too soon to admit it: additional research is needed. (See, e.g., their cited article in 38 Am.Psychologist at p. 558 & fn. 6.) But this is a frequent conclusion of academic authors. As the present case makes plain, appellate judges do not have the luxury of waiting until their colleagues in the sciences unanimously agree that on a particular issue no more research is necessary. Given the nature of the scientific endeavor, that day may never come.

16 CALJIC No. 2.20 (4th ed. 1979) provides in relevant part: “You are the sole judges of the believability of a witness and the weight to be given to his testimony.”

17 Implementing Penal Code section 1127b, CALJIC No. 2.80 (4th ed. 1979) provides in relevant part: “Duly qualified experts may give their opinions on questions in controversy at a trial. To assist you in deciding such questions, you may consider the opinion with the reasons given for it, if any, by the expert who gives the opinion. You may also consider the qualifications and credibility of the expert.

“You are not bound to accept an expert opinion as conclusive, but should give to it the weight to which you find it to be entitled. You may disregard any such opinion if you find it to be unreasonable."

18 The progeny of Johnson are no more persuasive than Johnson itself. Thus People v. Guzman, (1975) supra, 47 Cal.App.3d 380, 385-386, merely quotes portions of Johnson and Ballard; in turn, People v. Brooks, (1975) supra, 51 Cal.App.3d 602, 608-609, quotes Johnson, and People v. Bradley, (1981) supra, 115 Cal.App.3d 744, 751-752, quotes Guzman. Each opinion briefly disposes of the issue by deferring to the trial court’s discretion to reject such expert testimony in order to prevent it from “invading the province” of the jury.

19 E.g., “Failure of recollection is a common experience; and innocent misrecollection is not uncommon. It is a fact, also, that two persons witnessing an incident or a transaction often will see or hear it differently.”

20 In addition to several of the matters that Dr. Shomer would have discussed in the case at bar, Dr. Loftus offered to testify on the phenomenon of the “forgetting curve,” i.e., the tendency of memories to fade very rapidly at first, and then more slowly. She would also have explained “unconscious transfer,” whereby a witness confuses a person seen in one situation with a different person seen in another, and the “feedback factor,” whereby witnesses who discuss the case with each other after the event can unconsciously reinforce their confidence in their identifications. The eyewitnesses in Chappie were brother and sister, and had discussed the identity of “Dee.”

21 The court also rejected any claim that admission of this expert testimony would have consumed an undue amount of time, “since time spent on the crucial issue of the case can not be considered as ‘undue’ loss of time.’’ {Ibid.)

22 Not all psychologists, of course, have the special knowledge, experience, or training to qualify as experts on psychological factors affecting eyewitness identification—any more than all psychiatrists are qualified as experts, for example, on the “Munchausen syndrome by proxy.” (See People v. Phillips (1981) supra, 122 Cal.App.3d 69, 85.) In the case at bar, Dr. Shomer's relevant expertise was both demonstrated by defendant and impliedly conceded by the prosecution.

23 Indeed, the closeness of the matter was dramatically illustrated by the course of the trial itself. The People’s case was relatively uncomplicated: there was only one victim, one perpetrator, and essentially one crime, a robbery-murder. The defense was equally simple: defendant was in Alabama when the crime occurred. Apparently for these reasons, the trial was unusually brief for a capital case: the prosecution put on its entire case-in-chief in only four and a quarter hours, the defense took merely ninety-six minutes, and no rebuttal was offered. The main question to be answered by the jury was also relatively uncomplicated: were defendant and the black gunman the same person? Nevertheless, the jurors proceeded to deliberate for a total of 19V4 hours over a period of 6 days, before reaching their verdict of guilty. When jurors deliberate in these circumstances for more than three and a half times longer than it took to put on the entire prosecution and defense case, we may fairly infer they found the issue difficult to decide. (Cf. People v. Rucker (1980) 26 Cal.3d 368, 391 [162 Cal.Rptr. 13, 605 P.2d 843]; People v. Woodard (1979) 23 Cal.3d 329, 341 [152 Cal.Rptr. 536, 590 P.2d 391].)

24 Even when the trial court correctly excludes such testimony, the defendant may be entitled to a special instruction specifically directing the jury’s attention to other evidence in the record—e.g., facts developed on cross-examination of the eyewitnesses—that supports his defense of mistaken identification and could give rise to a reasonable doubt of his guilt. (See People v. Hall (1980) 28 Cal.3d 143, 158-160 [167 Cal.Rptr. 844, 616 P.2d 826]; People v. Palmer (1984) 154 Cal.App.3d 79, 85-89 [203 Cal.Rptr. 474]; People v. Aho (1984) 152 Cal.App.3d 658, 661-663 [199 Cal.Rptr. 671]; People v. West (1983) 139 Cal.App.3d 606, 608-610 [189 Cal.Rptr. 36].) The proper wording of such an instruction remains unsettled, however, and we express no view on the question here.

25 Although the jury was instructed on attempts in general (CALJIC No. 6.00 (4th ed. 1979)), there was no specific instruction concerning attempted robbery and informing the jury that if it was not satisfied beyond a reasonable doubt that defendant was guilty of the robbery, he could be found guilty of any lesser included offense shown by the evidence, and that attempted robbery is a lesser included offense of robbery. (See CALJIC No. 17.10.) The evidence here would have supported a sua sponte instruction on attempted robbery as a lesser included offense of robbery; the failure to so instruct was yet another complication among the many that plague this case.

As a result of this omission, the jury may not have realized that if it acquitted defendant of robbery it could consider his guilt of the lesser included offense of attempted robbery. Nevertheless, the acquittal of defendant for the robbery operates as an acquittal of all lesser included offenses including attempted robbery, and therefore bars retrial of defendant on these charges.

26 The delay between the guilt and penalty phases was caused by defendant’s effort to obtain a writ of prohibition from the Court of Appeal. He sought to prohibit the court from going forward with the penalty phase because of the inconsistency in the verdict between the acquittal on the robbery charge and the affirmative finding on the robbery special circumstance allegation.

27 A parallel provision appears in Penal Code section 1192: “Upon a plea of guilty, or upon conviction by the court without a jury, of a crime or attempted crime distinguished or divided into degrees, the court must, before passing sentence, determine the degree. Upon the failure of the court to so determine, the degree of the crime or attempted crime of which the defendant is guilty, shall be deemed to be of the lesser degree.”

28 It does not, however, extend to situations in which the jury expressly disagrees on the matter of degree. {Id. at p. 52.)

29 Respondent appears to concede that the trial court’s attempt to- “correct” the error in the verdict in the present case by resubmitting a modified form at the outset of the penalty phase similarly failed to cure the original defect.

30 Defendant challenges as error the court’s failure to instruct sua sponte on the lesser included offenses of second degree murder and voluntary manslaughter. Our disposition of the case renders it unnecessary to discuss the merits of this claim.

31 We do not decide at this time the question whether double jeopardy principles will bar retrial of defendant on a charge greater than second degree murder. First, the question has not been raised by the parties, and its answer is not immediately obvious. In the usual case a defendant is convicted by the trier of fact of a lesser degree of the crime charged and the judgment is reversed on appeal; in that event it has long been held that the defendant cannot be retried on the greater degree because of the double jeopardy clause. (Green v. United States (1957) 355 U.S. 184 [2 L.Ed.2d 199, 78 S.Ct. 221, 61 A.L.R.2d 1119]; Gomez v. Superior Court (1958) 50 Cal.2d 640 [328 P.2d 976].) Here defendant’s conviction of the lesser degree follows not from a finding of the trier of fact but by operation of law. Whether the same prohibition against retrial on the greater degree applies in such circumstances may require weighing a number of policy considerations that have not been briefed and argued on this appeal.

Second, the issue will not be presented on retrial unless the prosecution seeks a first degree murder conviction. But the prosecution’s sole theory of first degree murder at trial was felony murder; given the jury’s acquittal of defendant on the robbery charge and thus its implied acquittal on attempted robbery, the prosecution may be hard put to prove an underlying felony. If the prosecution limits itself to a maximum charge of second degree murder on retrial, the double jeopardy issue will manifestly not arise. Finally, as a general rule, the burden is on the defendant to enter a plea of double jeopardy at the appropriate time and to present a basis for the plea.