(dissenting, with whom Marshall, C.J., and Ireland, J., join). Eleven years ago, in suppressing a suggestive one-on-*292one identification remarkably similar to the one at issue here, we observed that “the danger of mistaken identification by a victim or a witness poses a real threat to the truth-finding process of criminal trials. Indeed, mistaken identification is believed widely to be the primary cause of erroneous convictions. . . . Compounding this problem is the tendency of juries to be unduly receptive to eyewitness evidence.” Commonwealth v. Johnson, 420 Mass. 458, 465 (1995) (Johnson). The court went on to note that the “admission of an identification obtained through unnecessarily suggestive procedures can only serve to exacerbate this problem,” id. at 467, and that these dangers “require the utmost protection against mistaken identifications.” Id. at 465. Evidence accumulated during the decade since we decided the Johnson case confirms that those observations were well founded.
Prior to the 1990’s, psychological literature and studies were the principal sources of skepticism about the reliability of eyewitness identification. For at least one-quarter of a century, leading psychologists in the field had warned the criminal justice system regarding the problems of such identifications and of the suggestive procedures often used to secure them. While the experimental findings of psychologists seemed compelling, it was not until the 1990’s that the criminal justice system began to take such findings seriously. It is not coincidental that this deepening interest paralleled the increasing use, refinement, and acceptance of DNA (deoxyribonucleic acid) evidence. As powerful a tool as DNA evidence was to become in securing convictions of those who were guilty, it was an equally powerful tool for reexamining the convictions of persons who were in fact innocent. The ability authoritatively to exonerate such persons (even in the relatively small number of criminal cases where DNA was collected, preserved, and proved dispositive) provided an extraordinary opportunity to identify and study the mistakes that had led to the conviction of the innocent, and the corresponding escape of the guilty.
This opportunity was embraced and pursued by researchers and the defense bar alike,1 and led to an increase in eyewitness identification case studies and recommendations for reform in *293the criminal justice system.* 2 Even the United States Department of Justice, through its National Institute of Justice, became fully engaged after 1996 in studying the scope of the problem of misidentification through the appointment of an extraordinary working group of prominent law enforcement experts, scientists, prosecutors, and defense attorneys.3
Much has been learned from these efforts. Of the first 163 exonerations secured through the use of DNA evidence, for example, we know that seventy-seven per cent of the convictions were the product of mistaken eyewitness identifications.4 There is no longer any doubt that mistaken eyewitness identification is the primary cause of erroneous convictions, outstripping all other causes combined, and that suggestive identification procedures are the primary cause of mistaken identifications.
Armed with the irrefutable data accumulated since the Johnson decision, the court now, inexplicably, takes a significant step backward in our jurisprudence regarding one-on-one identifications arranged by the police (showups), opening the door to certain error in the future.5 Rather than ensuring “the *294utmost protection^]” against misidentification and the suggestive procedures that too often produce them, the court substantially weakens the protections we have already put in place. Because I conclude that the identification procedure used initially to identify the defendant in this case was unnecessarily suggestive and should have been excluded at the trial, and that its admission in evidence was prejudicial in the context of what occurred at that trial, I respectfully dissent.
Discussion. Our court has consistently maintained that the due process clause of art. 12 of the Massachusetts Declaration of Rights requires suppression of identifications that are the product of unnecessarily suggestive procedures. Commonwealth v. Odware, 429 Mass. 231, 235 (1999) (“If a defendant establishes that [an identification procedure] was unnecessarily suggestive, then the identifications are excluded . . .”); Commonwealth v. Botelho, 369 Mass. 860, 866 (1976) (where defendant shows witness subjected to identification procedure “that was unnecessarily suggestive, and thus offensive to due process,” prosecution barred from introducing identification because “there is a ‘per se’ exclusion”). We have also consistently ruled that one-on-one identifications employed by the police are “inherently suggestive” and thus “generally disfavored.” Commonwealth v. Thompson, 427 Mass. 729, 735, cert. denied, 525 U.S. 1008 (1998). See Stovall v. Denno, 388 U.S. 293, 302 (1967) (practice has been widely condemned), overruled in part on other grounds, Griffith v. Kentucky, 479 U.S. 314 (1987).
In Johnson, we declined to trade our demanding test for the admission of inherently suggestive one-on-one identifications for the more relaxed “reliability test” adopted by the United States Supreme Court in Manson v. Brathwaite, 432 U.S. 98 *295(1977) (Brathwaite). See Johnson, supra at 465. We rejected the Brathwaite test, which permitted an unnecessarily suggestive identification to be admitted in evidence if it was nevertheless found to be “reliable,” because we concluded that such a test “provide[d] little or no protection from unnecessarily suggestive identification procedures, from mistaken identifications and, ultimately, from wrongful convictions.” Id. at 466. Today’s decision erodes the test we insisted on in Johnson, suffers from the defects we pointed out in rejecting the Brathwaite decision, and risks similar results.
To the extent our jurisprudence on the subject needs further clarification, we should affirm that a showup is “unnecessarily suggestive” and violates due process unless a “good reason exists for the police” to use it as an identification procedure. Commonwealth v. Austin, 421 Mass. 357, 361 (1995) (Austin). The court’s opinion, however, succeeds only in muddying jurisprudential waters that should remain clear. It does so by re-characterizing what was clearly a showup (and therefore “inherently suggestive”) as the equivalent of a “continuous nonsuggestive line up,” ante at 279, 281; essentially eviscerating the holding in Austin that a “good reason” must exist to conduct an inherently suggestive one-on-one showup; implying that a showup conducted five days after the assault may have been in the immediate aftermath of the crime; placing no burden on the police even to consider the use of less suggestive procedures, if a showup happens to be convenient; and suggesting that the reliability of the witness is indeed a factor to be weighed in determining whether a showup identification is “unnecessarily” suggestive, a proposition we specifically rejected in Johnson.6
1. The showup. The court, remarkably and without citation, reasons that the showup in this case was, in essence, not a showup at all. The court maintains that “the confrontation . . . was one in a continuum of events, part of a ‘daily ritual of viewing,’ ” and thus “the equivalent of a continuous nonsuggestive lineup.” Ante at 281. This conclusion, it appears, is the *296principal basis for the court’s validation of the challenged showup. See ante at 282 (“We conclude that, in these circumstances of a ‘daily ritual’ of attempted identifications similar in nature to the final one, the identification was not unnecessarily suggestive”); ante at 284 (police had good reason to take victim to view suspects because they could not bring all suspects to the police station or take their photographs).
To support its continuous lineup theory, the court relies on the motion judge’s following findings: the police and the victim “cruised the streets and beach areas” for the four days between the incident and the showup; they “stopped at least six times to view individuals fitting the general description [of the assailant]; “[o]n two of these occasions, the person stopped was in the company of uniformed police officers” and “the locus and procedure [of the contested showup] were not significantly different from what had become a daily ritual of viewing on the part of the victim.”7
Before today, this court has on no occasion concluded that a showup was not inherently suggestive because it had been preceded by other showups.8 Such a conclusion turns our jurisprudence on its head. It will, quite naturally, encourage (rather than discourage) the routine holding of multiple showups, in order to increase the probability that a successful one will pass constitutional muster. Indeed, it is unclear how, *297after this decision, we can continue to declare that showups are generally disfavored in the Commonwealth. The number of persons the victim viewed and rejected, whether as part of a photographic array, random drive-by viewings, or other showups, does not alter the essential, inherent suggestiveness of the one-on-one confrontation between the victim and the defendant at which the identification was made, and certainly should not be considered to provide the police with “good reason” to justify an even more suggestive showup.9
Moreover, the facts in this case simply do not support the conclusion that the challenged showup was not significantly different from what had become a daily ritual of viewing on the part of the defendant. On that morning, the victim arrived at the police station to embark on another drive-by viewing trip. Instead, she was told “there was a suspect on the beach,” and was driven to that location by detectives. In the beach parking lot, the unmarked cruiser pulled up to and stopped by the defendant, who was talking with two uniformed police officers and near whom stood the victim’s father. The victim was aware that her father had been looking for the perpetrator, and when she saw him with the police and the defendant, assumed that he had been the person who had picked the defendant out and notified the police. The detectives asked the victim to look at the defendant. After the defendant was moved closer for a better view, she recognized him from a distinctive mark he had on his face, a mark that she had not previously mentioned to the investigating officers or the police sketch artist. The showup lasted approximately seven minutes.
Clearly, this confrontation was different from most of the viewings the victim had participated in during the previous four days, where she looked for her attacker out the window of a moving police cruiser among persons in the general population. While the judge found that during this four-day period, “the detectives and [the victim] stopped at least six times to view *298individuals fitting the [attacker’s] general description,” there are no further findings regarding the majority of these stops: whether they occurred at the victim’s request (because she saw someone at whom she wanted to take a closer look), or at the suggestion of the detectives with whom she was riding (for similar reasons), or in response to other police officers’ calling in the location of a person fitting the general description of the attacker.10
On two of these six occasions, the judge found that the “person stopped was in the company of uniformed police officers.” Assuming that this finding supports the conclusion that on these occasions the detectives and the victim responded to sightings by other officers,11 and arrived and proceeded in a manner similar to the beach showup, this only occurred twice in four days. Two showups (of undetermined nature and length) in four days does not constitute a “daily ritual” of viewings similar in type and procedure to the showup at issue here.
Most important, the showup with the defendant was the only time that the fifteen year old victim viewed a suspect whom she believed her father had previously identified. It was also the only time she had viewed a suspect in his presence. Although the judge eventually concluded that the presence of the victim’s father was not fatal to the showup, she never suggests, nor is it *299reasonable to assert, that the father’s presence did not make this showup different from the viewings that had taken place on previous days.
The showup at issue here should be treated no differently from every other showup that this court has considered. It was inherently suggestive, and we should thus turn to the more relevant question whether it was unnecessarily so.
2. The good reason requirement. In Austin, supra at 361-362, we outlined the test governing the constitutionality of one-on-one identifications:
“Such showup identifications are disfavored because they are viewed as inherently suggestive. See Commonwealth v. Johnson, supra at 461; Commonwealth v. Santos, 402 Mass. 775, 781 (1988). See also Stovall v. Denno, 388 U.S. 293, 301-302 (1967), overruled in part on other grounds, Griffith v. Kentucky, 479 U.S. 314 (1987). Nonetheless, a one-on-one pretrial identification raises no due process concerns unless it is determined to be unnecessarily suggestive. Whether an identification procedure is ‘unnecessarily’ or ‘impermissibly’ suggestive, see Commonwealth v. Thornley, 406 Mass. 96, 98-99 (1989), involves inquiry whether good reason exists for the police to use a one-on-one identification procedure, see 1 W.R. LaFave & J.H. Israel, Criminal Procedure § 7.4 (b), at 581 (1984), bearing in mind that this court has said that ‘[ejxigent or special circumstances are not a prerequisite to such confrontations.’ Commonwealth v. Harris, 395 Mass. 296, 299 (1985). See Commonwealth v. Barnett, 371 Mass. 87, 92 (1976), cert. denied, 429 U.S. 1049 (1977); Commonwealth v. Coy, 10 Mass. App. Ct. 367, 371 (1980).
“Relevant to the good reason examination are the nature of the crime involved and corresponding concerns for public safety; the need for efficient police investigation in the immediate aftermath of a crime; and the usefulness of prompt confirmation of the accuracy of investigatory information, which, if in error, will release the police quickly to follow another track. Commonwealth v. Barnett, supra. The analysis cannot be generalized. Each case must be resolved on its own peculiar facts, considering as well *300that the existence of ‘good reason’ presents a question of law for the appellate court to resolve on the facts found by the motion judge. See Commonwealth v. Colon-Cruz, 408 Mass. 533, 538-539 (1990); Commonwealth v. Barnett, supra at 91.”
The court cites, in large part, these same passages in its opinion. Unfortunately, it then essentially ignores Austin’s requirement that the police have a “good reason” to hold a showup.12 Even when the court finally discusses “good reason,” it so broadens the concept as to only exclude completely arbitrary showups performed for no reason at all. The court does so by shifting the focus of the good reason requirement from the particular showup at issue to the over-all investigatory procedures the police used that led to the showup. As a result, the court appears to transform “good reason” from a significant precondition for an inherently suggestive and disfavored showup to simply one less than rigorous factor to be considered in determining its constitutionality.13 This novel,14 and ultimately unpersuasive reading of Austin, would allow (and, again, certainly not discourage) the use of showups in the absence of any “good reason” to do so.15
*301As the court concedes, Austin, supra at 361, stated that “[w]hether an identification procedure is ‘unnecessarily’ or ‘impermissibly’ suggestive . . . involves inquiry whether good reason exists for the police to use a one-on-one identification procedure . . . .” Ante at 283. This sentence follows Austin’s acknowledgment that showups are “inherently suggestive,” coupled with a reminder that the procedure only violates due process if it is “unnecessarily suggestive” (emphasis in original). Quite clearly, the sentence cited by the court explains that the difference between a suggestive showup and an unnecessarily suggestive showup is “whether good reason exists for the police to use” the procedure. This is only logical. If there is no good reason to hold a showup, such a procedure is unnecessary. If an inherently suggestive procedure is unnecessary, it is unnecessarily suggestive. Certainly, our landmark case in this area supports this view. See Johnson, supra at 471, quoting Brathwaite, supra at 128 (Marshall, J., dissenting) (rejecting Federal reliability test because “[b]y relying on the probable accuracy of a challenged identification, instead of the necessity for its use, the Court seems to be ascertaining whether the defendant was probably guilty” [emphasis added]).
Moreover, the support for the good cause requirement adopted in Austin comes from 1 W.R. LaFave & J.H. Israel, Criminal Procedure § 7.4(b), at 581 (1984). In § 7.4(b), the authors explain how to determine whether any identification procedure is unnecessarily suggestive. The “inquiry can ... be broken down into two constituent parts: that concerning the suggestiveness of the identification, and that concerning whether there was some good reason for the failure to resort to less suggestive procedures.” They continue: “Assuming suggestive circumstances” (the court here concedes, as it must, that one-on-one identifications are inherently suggestive), “the question then is whether they were impermissible or unnecessary.” This inquiry turns on whether there was some articulable need to hold a showup. See id. at 581-582 (describing this part of the test as *302the “Stovall necessity” test).16 Citation to this source confirms that Austin’s language was meant to conform to the prior jurisprudence requiring police to have a specific good reason before holding a particular showup.17
The remainder of the Austin decision also stands in stark contrast to the court’s diminishment of the good reason requirement. Immediately after introducing the good reason test, as explained above, Austin lists the factors “relevant to the good reason examination!)] the nature of the crime involved and the corresponding concerns for public safety; the need for efficient police investigation in the aftermath of a crime; and the usefulness of prompt confirmation of the accuracy of investigatory information, which, if in error, will release the police quickly to follow another track.” Austin, supra at 362. *303Not surprisingly, these factors are precisely those which our courts have traditionally taken into account in passing on the constitutionality of a showup. See, e.g, Commonwealth v. Thompson, 427 Mass. 729, 735 (1998); Commonwealth v. Barnett, 371 Mass. 87, 92 (1976); Commonwealth v. Bumpus, 354 Mass. 494, 501 (1968), cert. denied, 393 U.S. 1034 (1969).18
Although past decisions have referred to the analyses of these factors as a “totality of the circumstances” test, Austin insight-fully explains that existence of one or more of the factors validates a one-on-one identification precisely because it signifies that the police had “good reason” for what they did, and, therefore, the showup was not unnecessarily suggestive. It is for this reason that Austin, supra at 362, citing Commonwealth v. Colon-Cruz, 408 Mass. 553, 538-539 (1998), notes that “the existence of ‘good reason’ presents a question of law for the appellate court to resolve on the facts found by the motion judge.” As we explained in Commonwealth v. Colon-Cruz, supra, it is the requirement of appellate courts to “review the judge’s application of constitutional principles to the facts.” The unmistakable implication is that the question whether “good reason” exists is the test to determine whether a showup was unnecessarily suggestive, and, consequently, unconstitutional. Cf. State v. Dubose, 285 Wis. 2d 143, 155-156 (2005) (showups inadmissible unless, based on “totality of the circumstances,” they are “necessary”).19
Perhaps most notably, Austin never examines — nor even *304mentions the existence of — any factor other than “good reason” that would bear on the question whether a showup was unnecessarily suggestive. It certainly includes no hint that the “true” test of a showup’s constitutionality is whether the police procedure which led to the showup was sensible. Instead, Austin’s entire inquiry focused on whether the police had good reason for the showup in that case, and concluded “that, in the circumstances, which included a serious risk to public safety posed by a series of crimes having a similar modus operand! [including the use of a gun] and eyewitnesses who appeared exceptionally nonsuggestible, the police had good reason for what they did, and the [procedure] was not, as a result, unnecessarily suggestive.” Austin, supra at 364. Put more aptly, the procedure was held not to be unnecessarily suggestive “as a result” of the good reason police had to use it, not as a result of some multi-factorial test that might or might not take into account whether good reason for the particular showup existed.
In my view, Austin is clear, and quite correct, that in order to afford the “utmost protection against mistaken identifications,” see Johnson, supra at 465, this court requires the police to have “good reason” to hold an identification confrontation that is disfavored because of its inherent suggestivity.20 In this case, however, there does not appear to be any “good reason” that would justify the challenged showup. Although the court alludes to three possible “good reasons” for the showup, none is adequate.21
The court notes, in passing, that showups arranged “promptly *305after the criminal event” are an exception to the general presumption against showups, and “regularly held [to be] permissible.” Ante at 280, quoting Commonwealth v. Bowden, 379 Mass. 472, 479 (1980). See, e.g., Commonwealth v. Barnett, 371 Mass. 87, 92 (1976).22 That exception, however, is confined to showups held in the immediate aftermath of a crime and thus has no applicability to this case (where the showup took place on the fifth day after the incident). The remaining “good reasons” suggested by the court run afoul of our warning in Austin that the analysis of good reason “cannot be generalized,” but “must be resolved on [each case’s] peculiar facts.” Austin, supra at 362. To accept any of the court’s justifications for the showup would stretch the concept of good reason so far as to make it a toothless standard.
*306The court suggests that the showup was permissible because the police did not need to consider less suggestive alternatives where they were dealing with a man walking in a beach area on a summer day, where the “police clearly did not have enough evidence to compel the defendant to accompany them to the police station,” ante at 280, and where there was “no evidence . . . that the defendant would have agreed to an alternate procedure or that he would not have fled during the time necessary to arrange one.” Ante at 280. The court’s position, in essence, is that police may permissibly hold a showup when they are investigating a suspect whom they do not have probable cause to arrest. Such a rule would lead to an odd result: the less evidence there is to suspect a person of a crime, the greater the ability of the police to subject that person to an inherently suggestive identification procedure, which we have acknowledged often leads to misidentification. If the lack of probable cause alone creates a sufficiently good reason to hold a showup, we can expect a significant increase in the use of this generally disfavored procedure, as good reason would exist (and showups would be permissible) in an unimaginable number of investigations.
In any event, the police were not without available identification alternatives simply because no probable cause existed. They could have attempted to obtain the defendant’s cooperation to appear in a lineup, could have attempted to get his permission to take his photograph, could have photographed him without permission, or could have placed him under surveillance. The court tries to explain away the police conduct by noting that there was no evidence “that the defendant would have agreed to an alternate procedure.” Ante at 280. However, there was no evidence precisely because the police do not seem to have even considered whether they could have obtained this identification through less suggestive means.23 Had the suspect refused to agree to an alternate procedure, the police may have *307had good reason to hold the showup. But that did not occur in this case.
The court misses the point when it argues that “[f] allure of the police to pursue alternate identification procedures does not in itself render an identification unduly suggestive.” Ante at 280. Showups constitute a suggestive identification procedure. The issue is whether the showup was unnecessary. This question is quite appropriately linked to whether the police had, attempted to pursue, or even considered pursuing, readily available alternatives.24 Indeed, it seems quite unnecessary to hold a showup in these circumstances before finding out if the suspect will agree to a less flawed form of identification or whether a suspect can be properly subjected to a more covert involuntary form of identification (i.e., photographing without his consent).25 Cf. Austin, supra at 363 (validating identification procedure akin to showup in part because “defendant does not suggest an alternate identification procedure was available and should have been employed” and because court concluded no other option existed).
The court finally suggests that good reason for the showup existed because the process of “having the victim, whom the police viewed as ‘a very good witness [who] . . . felt she could make an identification,’ drive with them to view all men in a geographically reasonable area was a sensible approach to a dif*308ficult investigative problem.” Ante at 284. The court explains that “the police could not bring into the station or photograph each man in the entire area who met the victim’s description of her assailant” and that “[t]he police did the equivalent in a more efficient manner by canvassing the area in the company of the victim.” Ante at 284. This explanation is irrelevant. The challenged procedure here is not the general investigatory policy of drive-by viewings. Rather, the challenge is to one specific showup. The “good reason” required of the police must be correspondingly specific. See Austin, supra at 362 (“the analysis cannot be generalized”). The question is not whether the police could have taken photographs of all men matching the suspect’s description, but, for example, whether it could have readily done so in the circumstances of a single individual identified by the victim’s father.
The court is, of course, correct that the police procedure used here is more efficient than many other identification procedures. But this would be true in a large number of investigations. It cannot be seriously contended that something inherently and generally true of almost all showups can provide the constitutionally required good reason to justify the use of such a disfavored identification procedure in the present case.
The facts of this case are a far cry from what we have found to be good reason for a showup not conducted in the immediate aftermath of a crime. For example, in Austin, supra at 362-363, “[tjhree armed bank robberies (in New Hampshire, Massachusetts, and Rhode Island) had occurred within a few days of each other”; they “bore a high degree of similarity”; and the robber had fired his weapon at police and bystanders. Id. at 362, 363 & n.4. The robber’s appearance on the surveillance video tape of the third (Rhode Island) bank bore a resemblance to the composite drawing created from the description given by bank employee witnesses to the second (Massachusetts) bank robbery. Id. at 362. The witnesses, each of whom told police that they had a very good look at the robber in the bank, and were certain they could identify him if they saw him again, were asked to view the surveillance video tape and each identified the robber of the third bank as the robber of the second bank. Id. at 359-360, 362. We equated the identification *309procedure with a showup, see id. at 361, but explained that “a prompt viewing by the witnesses of the surveillance tape . . . was needed. Specifically, the police needed to determine . . . whether the robberies were committed by a single individual” in order to “focus their investigation accordingly.” Id. at 363. We noted that “[fjacing robberies in three States, the police departments involved could not make effective determinations on how to proceed in attempting to locate what might have been several perpetrators.” Id. We concluded that “in the circumstances, which included a serious risk to public safety posed by a series of crimes having a similar modus operand! and eyewitnesses who appeared exceptionally nonsuggestible, the police had good reason for what they did, and the . . . videotape identification was not, as a result, unnecessarily suggestive.” Id. at 364. Nothing in the case today resembles the facts that underlay the finding of good reason in Austin26 or any other case in this area of law.
3. Reliability. Equally troubling about the court’s opinion is its backtracking on the principles we adopted in Johnson. The court first finds that the drive-by viewings engaged in by the victim are relevant because the victim rejected all of the subjects before she made a positive identification of the defendant. It then credits the judge’s assessment of the victim’s credibility, and considers these adequate grounds on which the judge could “permissibly determine that the procedure was not unduly suggestive.” Ante at 281. But the proper question for the court is whether there was a good reason to use an inherently suggestive procedure in the circumstances, not whether the witness was reliable or credible in making her identification. See Johnson, supra at 471, quoting Brathwaite, supra at 128 (Marshall J., dissenting) (“The reliability approach has the ... effect of expanding a reviewing court’s powers because it allows the court to examine subjectively the circumstances surrounding the original crime on a case-by-case basis instead of objectively examining the identification procedures employed by the police. . . . ‘By relying on the probable accuracy of the chal*310lenged identification, instead of the necessity for its use, the Court seems to be ascertaining whether the defendant was probably guilty’ ” [emphasis added]). Reliability is the Brathwaite standard that we rejected in Johnson. Such findings are only relevant to the admissibility of subsequent identifications, and whether the victim had an adequate basis independent of the suggestive identification to make such an identification; they are not relevant to the necessity of the procedure in the first place. See Commonwealth v. Botelho, 369 Mass. 860, 867 (1976), quoting Stovall v. Denno, 388 U.S. 293, 302 (1967), overruled in part on other grounds, Griffith v. Kentucky, 479 U.S. 314 (1987) (“In deciding whether a particular confrontation was unnecessarily suggestive, the judge is to consider ‘the totality of the circumstances surrounding it.’ . . . This has been understood to refer to the episode itself; it does not extend to a consideration of the witness’s entire connection with the case to determine whether the confrontation, although set up in such a way as to be unnecessarily suggestive, was nevertheless reliable, and therefore usable — for example, because the witness had a clear perception of the offender and would not be misled by a one-on-one confrontation or the like.”). Id. at 867 n.5 (“Such evidence . . . rather enters necessarily into the second stage if that is reached [concerning the potential taint to a subsequent in-court identification] — consideration of the question whether the witness’s ability to identify had an independent source”).
Because there was no “good reason” for the showup in this case, particularly in light of the suggestive manner in which the showup was conducted, I agree with the Appeals Court that the identification made at that showup should have been barred at the trial as evidence from an “unnecessarily suggestive” identification procedure. Commonwealth v. Martin, 63 Mass. App. Ct. 587, 593, 594 (2005).
4. Harmless error analysis. We review admissions of evidence of unconstitutional identifications for harmless error. See Commonwealth v. Jones, 423 Mass. 99, 106 (1996). Erroneous admissions of such evidence will not amount to reversible error only “if the reviewing court may confidently say, on the whole record, that the constitutional error was harmless beyond a reason*311able doubt.” Commonwealth v. Howard, 446 Mass. 563, 570 (2006), quoting Commonwealth v. Miles, 420 Mass. 67, 73 (1995) , quoting Delaware v. Van Arsdall, 475 U.S. 673, 681 (1996).
In this case, the victim made three identifications of the defendant. The first was the showup, the second was a photographic array arranged seven hours after the showup, and the third was an in-court identification. Assuming there was clear and convincing evidence of an adequate independent source for the in-court identification, which, from the record, appears plausible, I cannot say confidently that the admission of evidence of the victim’s identification of the defendant at the showup was harmless beyond a reasonable doubt.
Identification was the key issue at trial. Shortly after the jury began deliberating, they returned with a question regarding the mark the victim recognized on the defendant’s face at the showup, specifically, whether the victim had “testified] to the mark ... in her initial description which she gave to the person who drew the composite.” She had not. The first time she had mentioned this mark to anyone was at the showup. The judge, however, informed the jury that they had to “determine [the answer] from [their] collective memory of the facts.”
The jurors’ question demonstrates that they were struggling over whether the victim’s initial identification of her assailant was reliable, and whether the victim’s primary basis for the showup identification and future identifications — the mark — was credible. The jurors then sent a second note to the judge that reinforces this conclusion: “With regard to the identity of the assailant, we cannot come to agreement that he in fact is the right person and therefore cannot move forward to deliberate on the indictments.” Two days later, after the judge gave a Rodriquez charge (see Commonwealth v. Rodriquez, 364 Mass. 87, 101 [1973] [Appendix A]) in response to the foreperson’s reporting that they were deadlocked, the jury returned guilty verdicts.
The ultimate determination of the credibility of a particular identification or set of identifications rests, as with all other credibility determinations, with the jury. Similarly, the weight to be given to the various types of identifications in this case is also uniquely the province of the jury. See Commonwealth v. *312Melvin, 399 Mass. 201, 207-208 (1987), and cases cited (“degree of suggestiveness present in the identification” if not constitutionally impermissible “went to the weight, not the admissibility, of this evidence,” which is “for the jury to assess”). In the circumstances of this trial, however, I cannot say that the judgment of the jury was not swayed by the erroneous admission in evidence of the unnecessarily suggestive showup identification. I would reverse the convictions and order that a new trial be held.
The Innocence Project, for example, was established in 1992 for the *293purposes of both assisting in securing freedom for the wrongly convicted and serving as a clearinghouse for information about such cases — tracking, documenting, and analyzing them from a systemic perspective.
See, e.g., Lee, No Exigency, No Consent: Protecting Innocent Suspects from the Consequences of Non-Exigent Show-ups, 36 Colum. Hum. Rts. L. Rev. 755 (2005); Steblay, Eyewitness Accuracy Rates in Police Showup and Lineup Presentations: A Meta-Analytic Comparison, 27 L. & Hum. Behav. 523 (2005); Wells, Eyewitness Testimony, 54 Ann. Rev. Psychol. 277 (2003); Collins, Improving Eyewitness Evidence Collection Procedures in Wisconsin, 2003 Wis. L. Rev. 529; Warden, How Mistaken and Perjured Eyewitness Identification Testimony Put 46 Innocent Americans on Death Row, Center on Wrongful Convictions, Nw. L. Sch. (2001); Eyewitness Identification Procedures, 22 L. & Hum. Behav. 603, 605 (1998); Wells, “Good, You Identified the Suspect”: Feedback to Eyewitnesses Distorts their Reports of the Witnessing Experience, 83 J. Applied Psychol. 360 (1998); Lindsay, Simultaneous Lineups, Sequential Lineups, and Showups, 21 Law & Hum. Behav. 391 (1997). See also State v. Dubose, 285 Wis. 2d 143 (2005).
United States Department of Justice, Eyewitness Evidence: A Guide for Law Enforcement (1999); United States Department of Justice, Convicted by Juries, Exonerated by Science: Case Studies in the Use of DNA Evidence to Establish Innocence After Trial (1996).
Mistaken Eyewitness Identifications, The Innocence Project (2006).
We need not look beyond our own borders for evidence of wrongful convictions based on suggestive one-on-one showups. In Commonwealth v. *294Leaster, 395 Mass. 96, 102-104 (1985), for example, we held that a showup identification made by the wife of a robbery and murder victim, in the parking lot outside a hospital, ninety minutes after the shooting, was not unnecessarily suggestive, and was therefore properly admitted in evidence. The witness was certain that Leaster was the shooter (whom she observed for three minutes during the robbery), because of the “mark under his eye” that she pointed out at the showup. After serving fifteen years in prison, Bobby Joe Leaster was exonerated of the crime. Charles Kenney, Justice for Bobby Joe, The Boston Globe Magazine (Feb. 28, 1988), at 15.
The reliability of the witness, of course, may still he relevant to the question whether a subsequent identification has a source independent of the suggestive one and therefore may he admitted. See Commonwealth v. Johnson, 420 Mass. 458, 476 & n.5 (1995) (Nolan, J., dissenting) (Johnson).
The Superior Court judge had not equated the showup with a nonsuggestive lineup.
The only case that would tangentially support this conclusion is Commonwealth v. Venios, 378 Mass. 24 (1979). In that case, a witness was shown two photographic arrays (one of two dozen persons and one of 150 persons) one day, and one additional photograph the next morning. He identified the man in the last photograph as the perpetrator. The court concluded that the “showing of one more photograph the next morning ‘was to some extent a continuation of an ongoing process of looking through police photos.’ ” Id. at 29, quoting Nassar v. Vinzant, 519 F.2d 798, 801 (1st Cir.), cert. denied, 423 U.S. 898 (1975). This case, involving photographs, not live confrontations, provides scant support for the court’s proposition. See Commonwealth v. Martin, 64 Mass. App. Ct. 1108 (2005) (unpublished memorandum and order) (where photographs shown to witness one at a time in continuous session, one-on-one confrontation cases inapplicable because “presentation of a photographic array does not implicate the kind of concerns that have led our courts to characterize such in person . . . show-ups as ‘generally disfavored’ ”).
Contrary to the court’s suggestion, ante at 280-281, drive-by viewings of persons in the general population of a community by a victim accompanied by police are not the same as showups, and are not generally viewed as inherently suggestive. They become so when the police drive a victim to view a specific person who has been identified as a suspect, and proceed to stage a one-on-one confrontation.
The court deems the lack of information regarding the drive-by viewings to be “irrelevant.” Ante at 281-282. Instead, it suggests that “what is relevant is that the victim rejected all the suspects until she made a positive identification.” Ante at 281. This entirely misses the point. That the victim rejected all previous suspects tells us next to nothing without specific evidence as to the circumstances surrounding those prior nonidentifications and their similarity to or difference from the showup in which she identified the defendant. Certainly (or at least hopefully), if the victim had spent the four days prior to the showup combing through police photograph books, the court would not suggest that her rejection of suspects she “viewed” bears on whether the showup was unnecessarily suggestive. That the drive-by viewings here may be more similar to the contested showup than viewing photographs at a police station, in no way makes “irrelevant” the question whether the drive-by viewings were similar enough to the showup to convert such a procedure into one that is not disfavored in this Commonwealth. In any event, the victim’s rejection of prior suspects is a completely irrelevant consideration in the determination whether a showup was unnecessarily suggestive.
In her testimony, the victim made clear that “other times” they had “gone to view suspects,” and that the detectives had made statements similar to “there’s a suspect” at some location.
Tellingly, the principal portion of the court’s opinion does not mention “good reason.” Rather, the court’s brief allusion to good reason comes only after it has already concluded that the showup was permissible and seems only grudgingly offered in response to the dissent.
The court cites no case as authority or even support for its new interpretation of Commonwealth v. Austin, 421 Mass. 357 (1995) (Austin).
The court today seems to adopt the position espoused by the dissent in Austin, which faulted the majority for analyzing whether good reason existed to justify the police showup. See id. at 368, quoting Commonwealth v. Thornley, 406 Mass. 96, 98-99 (1989) (O’Connor, J., dissenting) (“The question before the court... is not ‘whether good reason exists for the police to use a one-on-one identification procedure’ as the court. . . suggests,” rather, “[t]he question is whether [the procedure] was ‘so impermissibly or unnecessarily suggestive and conducive to irreparable misidentification as to deprive the defendant of his due process rights”).
The court warns that a true “good reason” requirement would be an “eccentric vesting of unnatural and unnecessary authority in the judicial branch.” Ante at 284.1 could not have crafted a better description of the now standard-less, multi-factorial, free-floating totality of the circumstances test that the court foists on our jurisprudence. Today’s decision seems to replace the search for “good reason,” and its traditional factors, with a wholly subjective totality *301of the circumstances determination as to whether a particular method of police investigation that led to an inherently suggestive showup was a “sensible approach” (ante at 284) taking into account the reliability and credibility of the witness and other unnamed and unexplained factors.
This section of the treatise explained the pre-Brathwaite Federal constitutional test for the validity of identification procedures. In Commonwealth v. Johnson, 420 Mass. 458, 462-463 (1995) (Johnson), we rejected the Federal standard announced in Manson v. Brathwaite, 432 U.S. 98 (1977) (Brathwaite), and held that we would continue to adhere “to the stricter rule . . . previously followed by the Supreme Court and first set forth in the [United States v. Wade, 388 U.S. 218 (1967), Gilbert v. California, 388 U.S. 263 (1967), and Stovall v. Denno, 388 U.S. 293 (1967), overruled in part on other grounds, Griffith v. Kentucky, 479 U.S. 314 (1987),] trilogy of cases.”
This reading of Austin appears to be the one we followed in Commonwealth v. Thompson, 427 Mass. 729, 735-736, cert. denied, 525 U.S. 1008 (1998). In that case, we quoted the two paragraphs from Austin set forth supra at 299-300, and then explained that the “identifications . . . were justified because they constituted parts of an efficient police investigation in the immediate aftermath of a crime.” Id. at 735. We then proceeded to explain why the officers had good reason to hold a showup. “[Ojfficers found two handbags and a wallet in the car [occupied by two men], suggesting that they were loot from recent robberies. The officers checked whether there had indeed been such robberies, and were informed about the Brookline incident, which had occurred just one hour before. The identification procedure allowed the officers to establish a connection between the stolen goods and the men, and provided [the victims] a good opportunity to view suspects while their recollections were fresh.” Based solely on its conclusion that the police had good reason to hold the showup, the court concluded that “the defendant failed to . . . show[] . . . that the identifications . . . were unnecessarily suggestive.” Id. at 736.
Similarly, this reading has been followed in the trial court. See Commonwealth vs. DelRio, Essex Superior Court No. ESCR2002-0647 (April 29, 2003), citing Austin, supra at 361 (“Determining whether a show-up is unnecessarily suggestive must include an inquiry into whether good reason existed for the use of the show-up” [emphasis added]).
I do not disagree that there may be other factors relevant to whether the police had “good reason” to hold a showup. However, this does not alter the principal proposition that the police need some “good reason” to engage in an inherently suggestive identification procedure.
The Supreme Court of Wisconsin recently joined our court in rejecting the Brathwaite reliability standard in State v. Dubose, 285 Wis. 2d 143 (2005). In doing so, the court found that “[i]t is now clear to us that the use of unnecessarily suggestive evidence resulting from a showup procedure presents serious problems in Wisconsin criminal law cases.” Id. at 164. The case involved a one-on-one showup of a suspect caught while fleeing the area of a reported armed robbery minutes after it occurred. The court focused on whether a more reliable identification procedure could have been used, and adopted a new standard for showups emphasizing such a requirement. Id. at 176-177. While I do not suggest that we presently adopt the Wisconsin standard, which, among other things, would fundamentally change our jurisprudence on showups conducted in the immediate aftermath of a crime, its analysis of the subject is *304worthy of consideration.
See, e.g., Johnson, supra at 471 (noting that the flaw in the Federal test is that it analyzes “the probable accuracy of a challenged identification, instead of the necessity for its use”). The police may very well, in certain cases, have a sufficiently good reason to hold a showup even if it is not absolutely necessary and even if less suggestive procedures are available. I would simply require the police to have a good reason for proceeding in this manner. See note 22, infra.
As the Appeals Court noted, there is nothing at all to distinguish this case from the routine police investigation that follows any serious criminal incident. See Commonwealth v. Martin, 63 Mass. App. Ct. 587, 592 (2005) (“We are aware of no case in which a showup identification was admitted on facts that, like these, suggest no special need. Although it has created certain exceptions, the Supreme Judicial Court has not, as yet, indicated that one-on-one identification procedures have ceased to be ‘disfavored’ ”).
The immediate aftermath exception applies “when the police, promptly after a criminal episode, show a defendant singly to a person for the sole purpose of identifying the wrongdoer.” Commonwealth v. Leaster, 395 Mass. 96, 102 (1985). In these cases, the police often begin searching for and find a suspect within minutes or hours after the reported crimes. See, e.g, Commonwealth v. Thompson, 427 Mass. 735 (1988), cert. denied, Thompson v. Massachusetts, 525 U.S. 1008 (1995) (showup held within approximately one hour of crime); Commonwealth v. Moffett, 383 Mass. 201, 213 (1981) (showup held “within half an hour of the assault and battery”); Commonwealth v. Bowden, 379 Mass. 472, 479 (1980) (showup held approximately “two hours after the murder”); Commonwealth v. Barnett, 371 Mass. 87, 88-89 (1976), cert. denied, 429 U.S. 1049 (1977) (showup of wounded suspect — taken to hospital shortly after wounded victim taken to hospital — within hour of shooting). See also Commonwealth v. Connolly, 356 Mass. 617, 623-624, cert. denied, 400 U.S. 843 (1970); Commonwealth v. Bumpus, 354 Mass. 494, 497-502 (1968), cert. denied, 393 U.S. 1034 (1969); Commonwealth v. Storey, 378 Mass. 312, 316-318 (1979), cert. denied, 446 U.S. 955 (1980); Commonwealth v. Venios, 378 Mass. 24, 29 (1979). In such circumstances, “the police procedure of arranging [a showup] is recognized as usual and natural and justified by the need for efficient investigation in the immediate aftermath of crime.” Commonwealth v. Barnett, supra at 92. “[P]rompt confrontation yielding a negative result, besides freeing the innocent, informs the police that a possible predisposition on their part is or may be in error and releases them quickly to follow another track.” Id. Additionally, allowing for an identification immediately after a crime “provides the witness with good opportunity for an accurate identification” because the witness can “view the suspect while his recollection or mental image of the offender is still fresh” and “before other images ... or his attempts to verbalize his impressions . . . distort the original picture.” Id. That the showup in the instant case was not conducted in the “immediate aftermath” of the crime is readily apparent. The police did not begin driving the victim around until the day following the attack, and the one-on-one identification occurred five days later.
The court would essentially relieve the police of any obligation to consider a less suggestive identification procedure in every case where they lack probable cause to arrest and, thus, where any contact between police and a suspect might increase the risk of flight. This would remove almost all barriers to the widespread incorporation of the showup in routine police procedure.
The court characterizes the “good reason” requirement as creating a best evidence rule. Such a caricature is, quite obviously, inaccurate. I do not suggest that if a less suggestive means of identification exists, it must always be used. However, whether such means exist and whether the police have explored the possibility of using them is certainly relevant to whether there was a good reason to hold a showup.
The court would place the burden of proof on the defendant to establish that he was willing to accompany police for a lineup or have his photograph taken. This, of course is beside the point. The question is not whether the defendant would have acquiesced to an alternate procedure, but whether the police had good reason to use an inherently suggestive one. In this case, whether the police explored or considered a nonsuggestive alternative is particularly relevant to the existence of good reason. Regardless, the only evidence we have on the subject shows that the defendant was compliant with the police. Although the court notes that the defendant asked to leave the beach area, it does not mention the fact that the defendant obediently followed the police request that he not leave. There was no evidence that he would not have agreed to additional police requests.
There was no evidence at the suppression hearing of a series of attempted assaults, rapes, or kidnappings — other than the one reported by the victim — that suggested the attacker would continue his assaults until he was caught.