This case is again before us after the United States Supreme Court, in Greineder v. Massachusetts, 133 S. Ct. 55 (2012), vacated the judgment and remanded Commonwealth v. Greineder, 458 Mass. 207 (2010) (Greineder), to this court. The remand came with instructions to give the case further consideration in light of the recent Supreme Court decision in Williams v. Illinois, 132 S. Ct. 2221 (2012) (Williams). Our review proceeds accordingly, and we conclude that Williams does not require us to change our jurisprudence.1 Therefore, we affirm the defendant’s conviction and the order denying his motion for a new trial.
1. Greineder. A recitation of the underlying facts is unnecessary as the facts have been fully set forth in our decision in Greineder, supra, which we incorporate by reference. We do, however, briefly recount our previous determination that the *582trial judge properly admitted the expert opinion of Dr. Robin Cotton, the forensic laboratory director of Cellmark Diagnostics laboratory (Cellmark), a private deoxyribonucleic acid (DNA) testing laboratory, that the defendant’s DNA matched the DNA found on a knife and two gloves recovered from the crime scene, despite Cotton’s reliance on the DNA test results obtained by a nontestifying analyst to form the basis of her opinion. Greineder, supra at 236. We held that such expert opinion testimony did not violate the defendant’s confrontation right pursuant to the Sixth Amendment to the United States Constitution2 “because the expert witness [was] subject to cross-examination about her opinion, as well as ‘the risk of evidence being mishandled or mislabeled, or of data being fabricated or manipulated, and as to whether the expert’s opinion [was] vulnerable to these risks.’ ” Id., quoting Commonwealth v. Barbosa, 457 Mass. 773, 791 (2010), cert, denied, 131 S. Ct. 2441 (2011) (Barbosa). We did, however, determine that Dr. Cotton’s testimony concerning the details of the nontestifying analyst’s DNA test results (illustrated by charts marked as chalks) was admitted in error. Greineder, supra at 237 (testimony concerning data, “while providing basis for [the] opinion, was hearsay”). Our review proceeded under the prejudicial error standard, and (for reasons we elaborate on infra) we determined that the defendant suffered no prejudice on account of the erroneous admission of the data that formed the basis of Dr. Cotton’s expert opinion. Id. at 239. As to that issue, we accordingly declined to grant relief under G. L. c. 278, § 33E, and affirmed the defendant’s conviction and the denial of his motion for a new trial. Id. at 255-256.
2. Massachusetts evidentiary rule on bases of expert opinion testimony. We begin our analysis of the continued validity of our approach to protecting the confrontation right of criminal defendants with a brief review of the Massachusetts evidentiary rule on expert opinion testimony and its permissible bases. The traditional rule was that an expert’s opinion had to be based either on evidence in the record or on facts of which the expert *583had direct, personal knowledge. See Barbosa, supra at 784-785. In Department of Youth Servs. v. A Juvenile, 398 Mass. 516, 531 (1986), we expanded the permissible bases of expert opinion testimony to include “facts or data not in evidence if the facts or data are independently admissible and are a permissible basis for an expert to consider in formulating an opinion.” Compare Mass. G. Evid. § 703 (2012), with Fed. R. Evid. 703 (expert opinion may be based on “inadmissible” facts or data). To determine “independent admissibility,” courts consider whether the underlying facts or data would be admissible through an appropriate witness. See Commonwealth v. Markvart, 437 Mass. 331, 337-338 (2002), quoting Department of Youth Servs. v. A Juvenile, supra. See also Mass. G. Evid., supra at § 703, note at 219 (DNA analyst may testify to tests personally conducted). As long as “independently admissible,” expert opinion may be based on facts or data not actually admitted in evidence.
Although facts and data not in evidence may form the basis of an expert witness’s opinion testimony, the expert may not present on direct examination the specific information on which he or she relied, see Commonwealth v. McNickles, 434 Mass. 839, 857 (2001), and cases cited, because expert testimony to the “fact[s] of the test results obtained by someone else . . . [is] hearsay.” Commonwealth v. Evans, 438 Mass. 142, 152 (2002), cert, denied, 538 U.S. 966 (2003). See Commonwealth v. Cohen, 412 Mass. 375, 393 (1992), quoting McCormick, Evidence § 246, at 729 (3d ed. 1984) (defining hearsay as statement, other than one made by declarant testifying at trial or hearing, offered in evidence to prove truth of matter asserted). The expert may, however, be required to disclose the facts or data that formed the basis of the expert opinion on cross-examination. See Mass. G. Evid., supra at § 705, at 223. Disallowing direct testimony to the hearsay basis of an expert opinion helps prevent the offering party from slipping out-of-court statements not properly in evidence in through the “back door” (citation omitted). Department of Youth Servs. v. A Juvenile, supra. Our prohibition of expert testimony concerning a nontestifying analyst’s test results (even where those results formed the basis of the expert’s opinion), on grounds that such basis evidence is offered for its truth and, therefore, is hearsay, differs from evidentiary rules in some other jurisdictions. Under Rule 703 of the Federal Rules of Evidence, for example, an *584expert witness may disclose the facts or data that formed the basis of the expert opinion during direct examination for the limited, nonhearsay purpose of explaining the bases for the expert opinion. See Williams, supra at 2239-2240 & n.10, quoting Notes of Advisory Committee on Proposed Rules to Fed. R. Evid. 703, 28 U.S.C. app. 361 (2000). The same is true in some States. See Williams, supra at 2234, citing Ill. Rule Evid. 703 (State evidence rule permitted introduction of basis of expert opinion where basis evidence not offered for its truth).
In Massachusetts, we draw a distinction between an expert’s opinion on the one hand and the hearsay information that formed the basis of the opinion on the other, holding the former admissible and the latter inadmissible. See Department of Youth Servs. v. A Juvenile, supra at 531-532. See also Barbosa, supra at 783-784 (“Where a Commonwealth expert testifies to her own opinion, the opinion is not hearsay, because the declarant of the opinion is testifying at trial”). The admission of expert opinion but exclusion of its hearsay basis protects a criminal defendant’s Federal and State constitutional right to confront witnesses. Id. at 783 (“our evidentiary rules and the Sixth Amendment are in harmony”). Expert opinion testimony, even if based on facts and data not in evidence, does not violate the right of confrontation because the witness is subject to cross-examination concerning his or her expert opinion and the reliability of the underlying facts and data. Id. at 785-786.
3. Nexus between rules of evidence and right of confrontation. To illustrate “the intersection between our common-law rules of evidence concerning expert testimony and the constitutional right of confrontation,” id. at 783, we highlight several foundational cases decided in the wake of the Supreme Court’s monumental Sixth Amendment decision in Crawford v. Washington, 541 U.S. 36 (2004) (Crawford),3 that concern a criminal defendant’s confrontation right in the context of expert opinion testimony that relies on a nontestifying analyst’s forensic test results.
Several years after Crawford, supra, we considered the admissibility of expert opinion and testimony to its basis in Com*585monwealth v. Nardi, 452 Mass. 379, 387-396 (2008) (Nardi). One question presented was whether an expert witness could testify to the autopsy findings of a pathologist who performed the autopsy but was unavailable to testify at trial. Id. at 383, 391-396. We concluded that the expert’s direct testimony about the autopsy findings of the unavailable pathologist was inadmissible for two reasons.4 Id. at 391-396. We concluded that the findings were inadmissible because they constituted hearsay — the witness was plainly asserting the truth of the statements in the autopsy report — and did not fall within a hearsay exception. Id. at 394. We also determined that the autopsy report was testimonial; thus, admission of testimony of an expert who did not author the autopsy report regarding the details of the report violated the defendant’s confrontation right.5 Id. at 392, 394. Nardi also presented the question whether the expert’s independent opinion regarding the victim’s death that was based on findings in the nontestifying pathologist’s report was admissible. Id. at 387-391. We concluded that the expert’s independent opinion was admissible because the autopsy report provided a permissible basis for its formulation. Id. at 388-389, quoting Commonwealth v. Markvart, 437 Mass. 331, 337 (2002). Moreover, we held that admission of the expert’s opinion testimony did not run afoul of the defendant’s confrontation right under either the Federal or State Constitution because the defendant had the opportunity to question the expert on the foundation of his opinion on cross-examination. Nardi, supra at 390-391, citing Commonwealth v. Daye, 411 Mass. 719, 743 (1992).
Two years later,6 in Barbosa, supra at 780-793, we employed *586the same reasoning that we used in Nardi in the context of an expert witness testifying to DNA test results, as opposed to autopsy results, of a nontestifying DNA analyst. The Commonwealth conceded that the portions of the expert witness’s testimony that relayed the opinion and recited the results of the nontestifying DNA analyst were inadmissible because they were hearsay.7 Id. at 782-783. As in our decision in Nardi, we nonetheless concluded that the defendant’s right of confrontation was not violated by admission of expert opinion based on inadmissible hearsay absent the analyst’s testimony, because the defendant “had a fair opportunity to confront [the expert] as to the reasonable basis for that opinion.” Barbosa, supra at 786, citing Nardi, supra at 390. We went on to address the unique factual context of objective chemical analysis, like DNA testing,8 as distinct from the more subjective science of autopsy reporting. Barbosa, supra at 787-791. We noted that, “with DNA analysis, the testing techniques are so reliable and the science so sound *587that fraud and errors in labeling or handling may be the only reasons why an opinion is flawed.” Id. at 790, citing National Research Council, Strengthening Forensic Science in the United States: A Path Forward 130 (2009). Furthermore, we “reject[ed] the premise that, in DNA analysis, there is no meaningful distinction between the opinion and the underlying fact finding.” Bar-bosa, supra at 789. To the contrary, we determined that data indicating the presence of certain alleles that match a defendant’s DNA profile is quite distinct from evidence of the statistical significance of the match.9 Id. We concluded that “[e]vidence of a [DNA] match ... is meaningless without evidence indicating the significance of the match.” Id., quoting Commonwealth v. Lanigan, 419 Mass. 15, 20 (1994).
In 2011, the Supreme Court addressed the constitutional sufficiency of “surrogate” testimony, or testimony of a person who did not perform or observe the laboratory analysis described in a blood alcohol report the prosecution sought to admit, as a vehicle for the admission of the nontestifying analyst’s findings. Bullcoming v. New Mexico, 131 S. Ct. 2705, 2709, 2714-2717 (2011) (Bullcoming). In its five-to-four decision in Bullcoming, the Court held that the nontestifying analyst’s report indicating *588the defendant’s blood alcohol concentration, offered through the laboratory supervisor, was admitted in error. Id. at 2710. The Court reasoned that, far from “raw, machine-produced data,” forensic test results are the product of independent judgment, and, as such, “are meet for cross-examination.” Id. at 2714. Cross-examination of a surrogate witness, even one with supervisory authority over the nontestifying analyst, was, according to the court, insufficient. Id. at 2714-2715. The surrogate in Bullcoming could not “expose any lapses or lies” or “reveal whether incompetence, evasiveness, or dishonesty” may account for the forensic test results. Id. at 2715. Therefore, the Court determined that the Sixth Amendment precludes admission of forensic test results through surrogate testimony. Id. at 2710, 2713. Significant for our analysis is that Justice Soto-mayor wrote separately to address the limits of Bullcoming: “[Tjhis is not a case in which an expert witness was asked for his independent opinion about underlying testimonial reports that were not themselves admitted into evidence. . . . [T]he State does not assert that [the surrogate witness] offered an independent, expert opinion.” (Emphasis added.) Id. at 2722 (Sotomayor, J., concurring in part).
Later that year, in Commonwealth v. Munoz, 461 Mass. 126, 131-138 (2011), vacated and remanded, 133 S. Ct. 102 (2012) (Munoz),10 we considered the continued validity of Barbosa, supra, in light of the Supreme Court’s decision in Bullcoming, supra. Munoz concerned the expert testimony of a chemist who relied on a nontestifying chemist’s raw data to form an independent opinion on the weight and composition of the cocaine seized from the defendant’s vehicle.11 Id. at 130-131. We drew a *589distinction between the facts in Munoz and those in Bullcoming: Munoz, like Barbosa, involved an expert’s “independent opinion about underlying testimonial reports that were not themselves admitted into evidence” (emphasis added), Munoz, supra at 132, quoting Bullcoming, supra at 2722 (Sotomayor, J., concurring in part), whereas Bullcoming concerned the admission, through a surrogate witness, of underlying data itself. Munoz, supra at 132-133. We then noted that the outcome of Bullcom-ing — that a surrogate witness cannot introduce a nontestifying analyst’s test results on direct examination — is in accord with the Massachusetts common-law prohibition of direct testimony about the underlying information that formed the basis of an expert’s opinion testimony because such information is offered for its truth and therefore is hearsay. Id. at 132 (“Our cases . . . distinguish^ between a substitute analyst’s testimony as to independent opinions based on data generated by a nontestify-ing analyst, and a substitute analyst’s testimony as to the testing analyst’s reports and conclusions”). See Nardi, supra at 392-393; Commonwealth v. McNickles, 434 Mass. 839, 857 (2001). Because the evidence the Bullcoming Court determined was erroneously admitted would have been excluded in Massachusetts, we concluded that Bullcoming did not undermine our bifurcated approach to safeguarding a defendant’s confrontation right, in which we admit an expert’s opinion testimony but exclude its hearsay basis on direct examination. Munoz, supra at 133. In Munoz, as in Barbosa, supra, we again rejected the view that “there is no meaningful distinction between the [expert] opinion and the underlying fact finding.” Munoz, supra at 136 n.13, quoting Barbosa, supra at 789. We also asserted that a surrogate expert witness can be meaningfully cross-examined on the risks of error in the underlying forensic testing on which he or she relied. Id. at 133-136. For these reasons, we concluded that an expert witness could still permissibly testify to his independent opinion, even if based on a nontestifying analyst’s test results, without offending a criminal defendant’s confrontation right under the Sixth Amendment of the United States Constitution or art. 12 of the Massachusetts Declaration of Rights. Id. at 130-131 & n.7.
*5904. Consideration in light of Williams. We turn now to our central task: to determine the import of Williams and then apply it to the facts of the present case. Williams concerned the testimony of a forensics expert who did not perform and was not even affiliated with the laboratory that performed the underlying DNA testing, but who used a nontestifying analyst’s DNA test results in formulating her opinion that the DNA recovered from the victim matched the defendant’s DNA. Williams, supra at 2227, 2229. In the course of her testimony, the expert referred to underlying DNA test results not admitted in evidence, despite her lack of personal knowledge that the DNA profile created by the nontestifying analyst was based on an analysis of the correct DNA sample. Id. at 2230,2236. Disclosure of such basis evidence was permitted under the relevant State rules of evidence for the limited, nonhearsay purpose of showing the basis for the expert opinion.12 Id. at 2231-2232, 2234-2235 & n.2. See Notes of Advisory Committee on Proposed Rules to Fed. R. Evid. 703, 28 U.S.C. app. 361 (2000).
The Williams Court held that the admission of statements that disclosed information about underlying DNA testing performed by a nontestifying analyst that formed the basis of the expert opinion did not violate the Sixth Amendment. Id. at 2228, 2240, 2244. The five Justices that reached this conclusion, however, differed in their reasoning. The plurality (four Justices) determined that the expert’s basis evidence was not offered for its truth. Id. at 2228, 2231-2232, 2235, 2239-2240. Therefore, the underlying facts that formed the basis of the opinion were admissible under Crawford because Crawford “does not bar the use of testimonial statements for purposes other than establishing the truth of the matter asserted.” Id. at 2235, quoting Crawford, supra at 59-60 n.9. Justice Thomas determined that admission of the expert’s testimony to the underlying basis of her expert opinion did not violate the Sixth Amendment because the *591nontestifying analyst’s report was not testimonial;13 therefore, he concurred in the judgment. Id. at 2255, 2259-2260 (Thomas, J., concurring in the judgment). Justice Thomas differed from the plurality in that he concluded that the expert’s basis testimony was, in fact, introduced for its truth. Id. at 2258-2259 (Thomas, J., concurring in the judgment). This view, that “[tjhere is no meaningful distinction between disclosing an out-of-court statement so that the factfinder may evaluate the expert’s opinion and disclosing that statement for its truth,” id. at 2257 (Thomas, J., concurring in the judgment), was shared by the four Justices writing in dissent. Id. at 2268-2269 (Kagan, J., dissenting). The four dissenters also determined that the laboratory report that included the DNA test results ostensibly offered for the limited purpose of explaining the basis of the expert opinion was testimonial and, therefore, statements that implicated it were admitted in violation of the confrontation clause. Id. at 2265, 2273 (Kagan, J., dissenting). Importantly, however, the dissent concluded:
“There was nothing wrong with [the expert’s] testifying that two DNA profiles — the one shown in the . . . report and the one derived from [the defendant’s] blood — matched each other; that was a straightforward application of [the expert’s] expertise. Similarly, [the expert] could have added that if the . . . report resulted from scientifically sound testing of [a DNA sample recovered from the victim], then it would link [the defendant] to the assault. What [the expert] could not do was what she did: indicate that the . . . report was produced in [a particular] way . . . ” Id. at 2270 (Kagan, J., dissenting).
Thus, no member of the Supreme Court plainly concluded that the expert opinion testimony was improper. See id. at 2235-2236 *592(plurality opinion), 2245-2246 (Breyer, J., concurring), 2255 (Thomas, J., concurring in the judgment), 2270 (Kagan, J., dissenting).
5. Import of Williams. Of great significance for our present purposes is that Williams focused on the admissibility of evidence of the basis of the expert’s independent opinion, and not the admissibility of the expert opinion itself. See id. at 2227, 2236, 2240 (plurality opinion). Five members of the Supreme Court concluded, albeit for different reasons, that such basis evidence is admissible without violating a defendant’s confrontation right. See id. 2244 (plurality opinion), 2245, 2248 (Breyer, J., concurring), 2255 (Thomas, J., concurring in the judgment). As we explained earlier, under Massachusetts jurisprudence, a forensic expert’s opinion that relies on the data or conclusions of a nontestifying analysis is bifurcated from its basis. We allow an expert to testify to his or her independent opinion even if based on data not in evidence; we do not allow expert witnesses to testify to the specifics of hearsay information underlying the opinion on direct examination. See Nardi, supra at 392-393; Commonwealth v. McNickles, 434 Mass. 839, 857 (2001). Thus, our rules of evidence differ from the Illinois evidentiary rule implicated in Williams that, like the Federal rule, allows an expert to disclose the data on which the expert’s opinion is based for the nonhearsay purpose of explaining its basis. Williams, supra at 2234-2235, 2239-2240 (discussing Ill. R. Evid. 703 and Fed. R. Evid. 703). The expert witness’s testimony in Williams that implicated the DNA test results of a nontestifying analyst “would have been . . . improper under long-standing Massachusetts law.” Commonwealth v. Munoz, supra at 133. Moreover, five Justices concluded that an expert’s disclosure of the underlying facts from an unadmitted DNA report went to its truth and not the limited purpose of explaining the basis for the expert opinion. See Williams, supra at 2256-2257 (Thomas, J., concurring in the judgment); id. at 2268 (Kagan, J., dissenting). This reasoning supports the Massachusetts rule that an expert’s testimony to the fact of a nontestifying analyst’s test results is hearsay. See Nardi, supra at 392.
Thus, our rules of evidence and the protections they afford *593are not inconsistent with Williams.14 Williams does not interpret the confrontation clause to exclude an expert’s independent opinion testimony, even if based on facts or data not in evidence and prepared by a nontestifying analyst. See Williams, supra at 2270 (Kagan, J., dissenting) (“straightforward application” of expertise does not offend confrontation clause). See also Bullcoming, supra at 2722 (Sotomayor, J., concurring in part) (independent expert opinion based on facts not in evidence is admissible and does not implicate confrontation clause). Moreover, where five Justices concluded that admission of underlying facts that formed the basis of an expert’s opinion did not offend the confrontation clause, Williams allows even more than an expert’s independent opinion in evidence. See Williams, supra at 2244 (plurality opinion), 2245-2246 (Breyer, J., concurring), 2255 (Thomas, J., concurring in the judgment). As is clear by now, our evidentiary rules afford a defendant more protection than the Sixth Amendment. See Nardi, supra at 392-394. See also Department of Youth Servs. v. A Juvenile, 398 Mass. 516, 531 (1986). In fashioning more protective evidentiary rules than the Supreme Court would require, we necessarily satisfy the mandates of the Sixth Amendment. We can, and often do, afford criminal defendants greater protections, both under our common-law rules of evidence and the Massachusetts Constitution. See Smith v. Commonwealth, 420 Mass. 291, 295 (1995), citing Commonwealth v. Aponte, 391 Mass. 494, 506 (1984) (“Our State *594Constitution often provides to criminal defendants broader protection than does the Federal Constitution under similar provisions”). Compare Commonwealth v. Tavares, 385 Mass. 140, 150, 152, cert, denied, 457 U.S. 1137 (1982), and cases cited (Massachusetts law requires both judge and jury to find voluntariness of confessions and admissions beyond a reasonable doubt), with Jackson v. Denno, 378 U.S. 368, 378 (1964) (under Federal Constitution, only judge determines whether confession or admission is voluntary), and Lego v. Twomey, 404 U.S. 477, 489 (1972) (voluntariness of admission or confession by preponderance of evidence). One vehicle for ensuring that a criminal defendant realizes the constitutional right of confrontation is our evidentiary rule that only an expert’s opinion and not its testimonial hearsay basis is admissible on direct examination.15 This rule strikes an appropriate “balance between the jury’s need to learn the factual basis of an expert’s opinion and the danger that the Commonwealth would use an expert’s opinion to inform the jury of facts not in evidence .... [This approach] is consistent with the right of confrontation . . . even where the expert bases his opinion on testimonial evidence that is admissible but not admitted.” Barbosa, supra at 785.
6. Application. Forensic expert opinion testimony, even if based on independently admissible facts or data not in evidence, does not violate a criminal defendant’s confrontation rights for two principal reasons. First, an expert witness is subject to *595cross-examination regarding that testimony. Second, the expert witness also can be meaningfully cross-examined about the reliability of the underlying data. We consider this two-fold reasoning in light of the facts of the present case.
First, as concerns the opportunity to cross-examine a forensics expert about his or her opinion, the defendant argues that the expert witness here did not, in fact, offer her independent opinion, but instead merely parroted the conclusions in Cellmark’s report of the DNA test results. The multiple reports submitted by Cell-mark included statistical calculations on the frequency that the genetic makeup recovered from evidence collected from the crime scene appears in the population, as well as conclusions regarding whether the defendant could be excluded as a source of DNA obtained from these samples. Nonetheless, and as we determined previously, the record reflects that Dr. Cotton reviewed the nontestifying analyst’s work, including six prepared reports, and then conducted an independent evaluation of the data.16 Greineder, supra at 236. She then “expressed her own opinion, and did not merely act as a conduit for the opinions of others.” Id. Moreover, we will not exclude expert opinion just because statistics indicating the significance of the genetic information have been included in the report provided to the expert. See Nardi, supra at 390. In fact, similar views on the statistical significance of allelic information extrapolated from test samples only lends credence to an expert’s independent opinion.
Second, the Massachusetts Association of Criminal Defense Lawyers (association), as amicus curiae, contends that an expert *596witness cannot adequately testify to the reliability of the underlying forensic data that formed the basis of the expert’s independent opinion. The association highlights the fallibility of DNA analysis, including its vulnerability to mishandling, manipulation, fabrication, and bias. But see National Research Council, Strengthening Forensic Science in the United States: A Path Forward, 7 (2009) (“With the exception of nuclear DNA analysis, ... no forensic method has been rigorously shown to have the capacity to consistently, and with a high degree of certainty, demonstrate a connection between evidence and a specific individual or source” [emphasis supplied]). The association also asserts that an expert cannot be cross-examined effectively on the subjective decisions a forensic analyst makes in developing a DNA profile.17 Although we acknowledge that forensic test results as well as expert opinion testimony (like all testimony) may be imperfect, we nonetheless maintain, consistent with our jurisprudence, that forensics experts can be meaningfully cross-examined about the underlying data and the procedures and processes that led to its production. See Barbosa, supra at 791. See also Munoz, supra at 134 (“Reasonable reliance . . . implies that the expert will have ascertained that the data on which he or she relies are adequate and appropriate to the task and were prepared in conformity with accepted professional — here scientific — practices and procedures”). In fact, the expert witness in this case may be in the best position to speak to the testing processes at Cellmark and the resultant DNA profiles generated. Dr. Cotton was an expert in biology and DNA, and she was also the forensic laboratory director of Cellmark at the time of the trial. As such, she was in a unique position to speak to the DNA testing process employed by Cellmark, including the chain *597of custody and evidence-handling protocols,18 and other standardized procedures of Cellmark analysts. For example, and as we addressed in Greineder, supra at 238, Cellmark was “using a standard minimum interpretation threshold setting on its computer software” to analyze whether someone was included as a possible source of DNA found on evidence recovered from a crime scene. Midway through the testing of items in this case, Cellmark raised this minimum interpretation threshold. Id. Only a senior person within Cellmark could testify to the reasoning behind raising the standard interpretive threshold; the individual analyst who conducted the DNA testing would not have been similarly able. Not only was Dr. Cotton in a better position than an individual analyst to testify to the reliability and accuracy of Cellmark’s testing procedures, she was also in a better position to testify meaningfully than the expert witness in Williams, whose testimony, which implicated the underlying forensic report, satisfied the mandates of the Sixth Amendment. Williams, supra at 2228, 2240, 2244. The forensics expert in Williams did not even work for Cellmark (which, as in this case, was the independent laboratory that produced a DNA profile from a sample collected from the victim), but instead worked for a laboratory run by the Illinois State Police. Id. at 2229. Although the expert in Williams could testify to the process of DNA analysis generally, she was not in as strong a position to testify to the details of Cellmark’s DNA testing procedures as was Dr. Cotton.19 Because there was a greater opportunity for *598meaningful cross-examination of the expert witness in the present case as compared to the expert in Williams, we conclude that her opinion testimony did not violate the defendant’s confrontation right.20
Here, not only did the defendant have a meaningful opportunity to cross-examine Dr. Cotton on the reliability of the data that formed the basis of her expert opinion, his experienced trial counsel used the opportunity effectively. Trial counsel reviewed Cellmark’s DNA testing process with Dr. Cotton on cross-examination, and had her acknowledge aspects of the process not governed by standard protocols. He challenged Dr. Cotton on Cellmark’s minimum interpretation thresholds that, although they were raised mid-analysis, were still below testing thresholds used elsewhere. Greineder, supra at 238. Trial counsel also elicited information from Dr. Cotton concerning how data filtering may affect DNA test results. A significant portion of trial counsel’s cross-examination of Dr. Cotton was devoted to attacking the reliability of Cellmark’s reported test results. Id. at 239. Thus, to the extent that DNA test results, with their “strong scientific underpinning,” see Barbosa, supra at 788 n.13, quoting National Research Council, Strengthening Forensic Science *599in the United States: A Path Forward 134 (2009), are vulnerable to error, the defendant did not want for opportunity to test the reliability of data underlying Dr. Cotton’s expert opinion. Greineder, supra at 238-239. See Williams, supra at 2239 (“match also provided strong circumstantial evidence regarding the reliability of Cellmark’s work”). Therefore, the confrontation clause’s purpose — to ensure fair criminal trials based on reliable evidence — was served. See Bullcoming, supra at 2725 (Kennedy, J., dissenting); Commonwealth v. Amirault, 424 Mass. 618, 631 (1997), quoting Maryland v. Craig, 497 U.S. 836, 846 (1990). That an expert may give an opinion but may not describe its hearsay basis on direct examination is “as sensible an approach to balancing the competing concerns as any.” Commonwealth v. Avilia, 454 Mass. 744, 762 n.18 (2009), quoting D.H. Kaye, D.E. Bernstein, & J.L. Mnookin, The New Wig-more: A Treatise on Evidence — Expert Evidence § 3.10.8, at 64 (Supp. 2009).
7. A bifurcated approach: admitting opinion, but excluding its hearsay basis on direct examination. The defendant challenges our bifurcation of admissible expert opinion from its inadmissible, hearsay basis. He argues that the two elements — the underlying facts that form the basis of an expert opinion and the expert opinion itself — are inextricably linked, at least in the context of DNA analysis. We disagree. There is a clear distinction between the allelic information that establishes genetic makeup and the statistical significance of the data that establishes how frequently a genetic combination appears in the population at large. See Commonwealth v. Vao Sok, 425 Mass. 787, 789, 790-791 (1997) (discussing alleles), 791 (statistical analysis). As we said in Barbosa, supra at 789, it is the statistical significance of a DNA match that is of greatest use to a jury; information about the prevalence of a particular gene combination gives meaning to the underlying fact of allelic presence. See Commonwealth v. McCowen, 458 Mass. 461, 484 (2010), citing Barbosa, supra (“the admission of the raw DNA testing results alone, even where the allele numbers match, is meaningless to a jury”). Admittedly, Justice Kagan, dissenting in Wz7-liams, approached the relationship between underlying facts and the resultant opinion in the DNA context from a different end. *600She emphasized the importance of the underlying DNA test results to aid the jury in evaluating the credibility of the expert witness and the validity of his or her opinion. Williams, supra at 2268-2269 (Kagan, J., dissenting) (“If the [underlying out-of-court statement] is true, then the conclusion based on it is probably true; if not, not. So to determine the validity of the witness’s conclusion, the factfinder must assess the truth of the out-of-court statement on which it relies”). Hearsay data underlying an expert’s opinion testimony are admissible if elicited by the defendant on cross-examination (who, in the act of cross-examination waives his confrontation right, see Barbosa, supra at 785-786). See Mass. G. Evid. § 705 (2012). Therefore, basis evidence that is hearsay may become available to the jury to evaluate a witness’s credibility. If a defendant does not open the door on cross-examination to the hearsay basis of an expert’s opinion, then the jury may properly accord less weight to the expert’s opinion. Williams, supra at 2239-2240. Should the prosecution wish to offer weightier testimony, then it should call either the author-analyst (assuming that person is qualified to testify to the statistical significance of the underlying data), or both the author-analyst and an expert, the former to testify to the underlying factual findings and the latter to interpret them.21
In support of its view that our evidentiary bifurcation of admissible expert opinion from its inadmissible hearsay basis on direct examination is not sufficiently protective of defendants’ confrontation right,22 the Committee for Public Counsel Services (CPCS), as amicus curiae, provides examples of cases where *601the prosecution has improperly elicited basis evidence that is hearsay from forensic experts on direct examination. Such cases, according to CPCS, demonstrate that an expert is unable to give a bare opinion without disclosure on direct examination of the nontestifying analyst’s underlying test results or conclusions. We disagree. Indeed, despite our repeated insistence that such basis evidence is inadmissible, prosecutors do elicit basis evidence from forensic expert witnesses on direct examination. See, e.g., Greineder, supra at 236-237. We then have had to evaluate the effect of the erroneous admission given the applicable standard of review.23 See, e.g., id. at 239. However, this does not indicate a flawed system. It merely reveals either a lack of emphasis on the prohibition of expert basis evidence on direct examination, or the scarcity of practical skills training to aid prosecutors in eliciting only opinion and not its hearsay basis from expert witnesses.24 Thus, we use the present opportunity to emphasize to judges that the hearsay bases of expert opinion testimony are inadmissible on direct examination. As concerns eliciting only the expert’s opinion, the prosecutor, as a general matter, may first ask the expert for an opinion, given the expert’s background and training and after review of the *602underlying data, whether the defendant, given his or her DNA profile authenticated by chain of custody evidence as having originated from the defendant, could be excluded as a possible source of DNA recovered from the crime scene, similarly authenticated by chain of custody evidence. See id. at 238. If the expert opines that the defendant could not be excluded as a possible contributor, a prosecutor may then elicit an expert’s opinion on the significance of the DNA evidence. R.C. Michaelis, R.G. Flanders, Jr., & RH. Wulff, A Litigator’s Guide to DNA: From the Laboratory to the Courtroom 297-298 (2008). The prosecutor may ask the expert for his or her opinion on the statistical likelihood that the DNA profile found on the relevant exhibit would be found in an individual randomly selected from the population. See note 9, supra. See also Williams, supra at 2270 n.2 (Kagan, J., dissenting) (assumptions, framed in hypothetical question “if X is true, then Y follows,” pass constitutional muster). See generally P.M. Lauriat & J.F. McHugh, Massachusetts Expert Witnesses § 6.1.2 (Mass. Cont. Legal Educ. 2d ed. 2010). Greater detail regarding specific allelic presence at various loci is unnecessary and, under our evidentiary rules, prohibited on direct examination of a forensics expert who did not conduct the underlying DNA testing.
Granted, in this case, Dr. Cotton did testify to the details and results of the nontestifying analyst’s DNA test results. Greineder, supra at 236-237. We held then and repeat now that this evidence elicited on direct examination was admitted in error. Id. at 237. We maintain, however, that the defendant was not prejudiced by the erroneous admission. As we determined in Greineder, supra at 239:
“Defense counsel made extensive use of the data prepared by [the nontestifying analyst] in his cross-examination of Dr. Cotton. This cross-examination buttressed two pillars of the defense, namely, the unreliability of Cellmark’s testing procedures, which necessarily affected the credibility of Dr. Cotton’s opinion as to the high probability of the defendant’s being a contributor to the DNA on the knife and brown gloves, and the defense theory that a third party committed the murder. . . . Both strategies depended entirely on the use of [the nontestifying analyst’s] data in *603cross-examination of Dr. Cotton. . . . The data were an integral part of the defense, and it was going to be the subject of inquiry by the defense whether [the nontestify-ing analyst] testified or not.”
8. Conclusion. For the reasons discussed above, Dr. Cotton’s expert opinion that the defendant’s DNA matched the DNA on items recovered from the crime scene was properly admitted. Expert opinion testimony, even that which relies for its basis on the DNA test results of a nontestifying analyst not admitted in evidence, does not violate a criminal defendant’s right to confront witnesses against him under either the Sixth Amendment or art. 12 of the Massachusetts Declaration of Rights. Nothing in the Supreme Court’s decision in Williams v. Illinois, 132 S. Ct. 2221 (2012), is to the contrary. Dr. Cotton’s direct testimony to the details of the DNA test results of a nontestifying analyst, however, was improperly admitted. Because the defendant was not prejudiced by the erroneous admission, we affirm his conviction and the denial of his motion for a new trial.
So ordered.
We acknowledge the amicus briefs submitted by the Committee for Public Counsel Services; the Massachusetts Association of Criminal Defense Lawyers; and the Attorney General and the district attorneys for the Berkshire, Bristol, Cape and the Islands, Essex, Hampden, Middlesex, Northwestern, Suffolk, and Worcester districts.
The defendant did not argue that art. 12 of the Massachusetts Declaration of Rights provides greater protection than the Sixth Amendment to the United States Constitution, and we did not address that issue. Commonwealth v. Greineder, 458 Mass. 207, 237 n.9 (2010) (Greineder).
The Supreme Court, in Crawford v. Washington, 541 U.S. 36, 59 (2004) {Crawford), held that “[tjestimonial statements of witnesses absent from trial [can be] admitted only where the declarant is unavailable, and only where the defendant has had a prior opportunity to cross-examine.”
There is an important distinction between satisfying the mandates of common-law evidentiary rules and satisfying the mandates of the confrontation clauses of the Federal and State Constitutions. In criminal cases, out-of-court statements are only admissible if they satisfy both; failure to satisfy either the applicable rules of evidence or the Federal and State Constitutions will result in the exclusion of evidence. See generally Mass. G. Evid., Introductory Note to art. VDI (2012), and cases cited.
Although we concluded that the expert’s testimony to the findings in the autopsy report on direct examination was admitted in error, we nonetheless determined that, in the circumstances of the particular case, its admission did not create a substantial likelihood of a miscarriage of justice and, therefore, a new trial was unwarranted. See Commonwealth v. Nardi, 452 Mass. 379, 395-396 (2008) (Nardi).
In the interim, in Melendez-Diaz v. Massachusetts, 557 U.S. 305, 308-309, *586311 (2009) (Melendez-Diaz), the Supreme Court applied Crawford, supra, to forensic evidence and held that the admission of a chemical drug test report without the testimony of the analyst who performed the test violated the defendant’s Sixth Amendment confrontation right. One question Melendez-Diaz left unanswered was whether “ ‘admission into evidence of affidavits containing testimonial hearsay, would necessarily require the government to call [an analyst] to testify [where] her report [is] not admitted into evidence,’ but is instead relied on as the basis for the independent opinion of a second analyst.” Commonwealth v. Munoz, 461 Mass. 126, 131 (2011), vacated and remanded, 133 S. Ct. 102 (2012) (Munoz), quoting United States v. Pablo, 625 F.3d 1285, 1295 (10th Cir. 2010), cert, granted and vacated, 133 S. Ct. 56 (2012). We determined that “Melendez-Diaz did not purport to alter the evi-dentiary rules governing expert testimony.” Commonwealth v. Barbosa, 457 Mass. 773, 787 (2010), cert, denied, 131 S. Ct. 2441 (2011) (Barbosa), citing Commonwealth v. Avila, 454 Mass. 744, 762-763 (2009). Melendez-Diaz, we concluded, “stands simply for the proposition that, under the Sixth Amendment right of confrontation, the Commonwealth may not offer an expert opinion in evidence unless the defendant has an opportunity to cross-examine the expert” (emphasis added). Barbosa, supra, citing Melendez-Diaz, supra at 311.
testimony to the hearsay basis underlying the expert opinion did not create a substantial likelihood of a miscarriage of justice and, thus, did not warrant retrial. See Barbosa, supra at 792-793 (same).
For descriptions of deoxyribonucleic acid (DNA) testing, see Commonwealth v. Vao Sok, 425 Mass. 787, 789 792 (1997), and Commonwealth v. Curnin, 409 Mass. 218, 227 231 (1991) (Appendix).
Here, Dr. Cotton presented her opinion on the statistical likelihood that the DNA recovered from the crime scene that matched the defendant’s DNA profile would match an unrelated person in the general population. Greineder, supra at 217-218. See Commonwealth v. Vao Sok, supra at 791 (statistical significance of DNA analysis). She testified, for example, that a statistical analysis based on the match at four loci of the defendant’s DNA to DNA recovered from the knife found at the crime scene “indicated that one in 1,400 unrelated African Americans and one in 2,200 unrelated Caucasians would be included with the defendant as possible secondary” sources of the DNA. Greineder, supra at 217. She further testified that “statistical analysis indicated one in fifteen million unrelated African Americans, and one in 680,000 unrelated Caucasians could be included with the defendant as possible contributors” to DNA recovered from a left hand glove; and that “one in 680 million unrelated African-Americans and one in 170 million unrelated Caucasians would be included with the defendant as possible secondary contributors” to DNA recovered from a right hand glove similarly recovered from the crime scene. Id. at 217-218. Such statistical information is necessary because approximately “99.9% of DNA is identical between any two individuals”; only the remaining .1% of DNA is particular to a given individual. Federal Judicial Center, Reference Manual on Scientific Evidence 491 (2d ed. 2000). See Commonwealth v. Vao Sok, supra at 789, quoting Commonwealth v. Cumin, 409 Mass. 218, 228 (1991).
As in Greineder, the Supreme Court vacated the judgment in Munoz, supra, and remanded it for further consideration in light of Williams v. Illinois, 132 S. Ct. 2221 (2012) (Williams). Munoz v. Massachusetts, 133 S. Ct. 102 (2011). The defendant in Munoz decided not to pursue the appeal before this court.
In Munoz, the Commonwealth conceded that the expert’s testimony on direct examination that implicated a nontestifying analyst’s data and conclusions violated the defendant’s confrontation rights and was admitted in error. Munoz, supra at 130, citing Barbosa, supra at 186. Thus, the issue in Munoz, supra at 131-133, was only whether Bullcoming v. New Mexico, 131 S. Ct. 2705, 2710 (2011) (Bullcoming), mandated a change in the Massachusetts evidentiary rules that permit expert opinion testimony even if based on testimonial hearsay. We concluded that it did not. Munoz, supra at 133.
The Williams plurality distinguished Williams from Bullcoming on the ground that, in Williams, the expert’s reference to the underlying forensic report, which was not introduced in evidence, was (at least, ostensibly) not offered to prove the truth of the matter asserted in the report. Williams, supra at 2240. In contrast, according to the Williams plurality, the forensic report admitted in evidence through the testimony of a “surrogate” witness in Bullcoming, supra, was unquestionably offered for its truth. Id.
The plurality, in an alternative conclusion, also determined that the expert’s basis testimony was admissible because the report that included the underlying DNA test results was not testimonial. Williams, supra at 2242-2244. However, the test the plurality employed in determining that the nontestifying analyst’s DNA test results were not “testimonial” differed from the test Justice Thomas employed to reach the same conclusion. Compare id. at 2242-2243 (primary purpose), with id. at 2255, 2259-2264 (Thomas, J., concurring in the judgment) (formality and solemnity). We address the issue whether the DNA test results in this case qualified as “testimonial.” See note 15, infra.
In determining that Williams, supra, does not undermine our jurisprudence, we mirror our approach in Munoz, supra at 131-136, in which we evaluated the continued validity of our bifurcation of expert opinion from its hearsay basis in light of the Supreme Court’s decision in Bullcoming, supra. There we noted that Bullcoming concerned the admission of evidence (a forensic report prepared by a nontestifying analyst via a surrogate witness) that would not have been admissible under Massachusetts law. Munoz, supra at 132-133. Our prohibition of an expert’s direct testimony about the underlying facts that formed the basis of the opinion “essentially, [is] a different way of stating the central holding of Bullcoming: that a substitute analyst cannot testify to, or otherwise introduce, the original analyst’s reports and conclusions on direct examination.” Id. at 133. Similarly, Williams concerns the constitutionality of the admission of expert basis evidence that would not have been admissible in Massachusetts because it is hearsay. See Williams, supra at 2256-2257 (Thomas, J., concurring in the judgment); id. at 2268 (Kagan, J., dissenting). Therefore, Williams does not undercut our evidentiary rule that sufficiently safeguards defendants’ confrontation right.
That our rules of evidence preclude an expert from testifying on direct examination to the hearsay information that formed the basis of his or her opinion is reason enough for its exclusion. Nevertheless, we also conclude that Dr. Cotton’s testimony regarding the DNA analyst’s test results was admitted in error for the additional reason that the nontestifying analyst’s report was testimonial. See Commonwealth v. McCowen, 458 Mass. 461, 483 (2010). Although the precise contours of the “primary purpose” test are arguably in flux following Williams, the report in the present case is distinguishable from the report in Williams because Cellmark Diagnostics laboratory (Cellmark) had the defendant’s DNA sample. See, e.g., Williams, supra at 2243. This indicates that the defendant was targeted as a suspect at the time Cellmark conducted DNA testing, and demonstrates that the DNA testing was conducted for the purpose of obtaining evidence for later use at trial. Thus, as we determined previously, “a reasonable person in [the nontestifying analyst’s] position would anticipate her findings and conclusions being used against the accused in investigating and prosecuting a crime.” Greineder, supra at 237, quoting Nardi, supra at 394. See Commonwealth v. Avila, 454 Mass. 744, 763 n.20 (2009), quoting Nardi, supra (same as to autopsies).
As it was in the present case, unfiltered or raw data, in addition to filtered data, should be made available to DNA expert witnesses, as well as to defense counsel during discovery. Unfiltered or raw data shows all peaks detected by the machine, whereas an analyst may limit the DNA peaks that a computer detects. Providing an expert with unfiltered data puts the expert in the best position to formulate an independent opinion and to respond to questions concerning the risk that evidence was fabricated or manipulated. See Barbosa, supra at 791. Beyond ensuring that data are made available to the expert witness, we decline to delve more deeply into an expert witness’s trial preparation process. We leave that to the “crucible of cross-examination." Crawford, supra at 61. Where the expert has not reviewed the raw data, defense counsel may use this to gain an advantage. Such was the circumstance of the present case: the expert had only reviewed the filtered, printed data (as opposed to the raw, electronic data), which defense counsel then used to impeach the expert’s testimony.
DNA test results may, for example, contain “artifacts,” or false indications of allelic presence. See Commonwealth v. Bly, 448 Mass. 473, 481 (2007). See also Commonwealth v. Buckman, 461 Mass. 24, 35-36 (2011), cert, denied, 132 S. Ct. 2781 (2012). Dr. Cotton testified that it is common for the computer to label a peak as an allele and the analyst to override that call. However, the fact that a nontestifying analyst may use his or her independent judgment in generating a DNA profile does not indicate that the resulting data are inherently unreliable or that an expert cannot be meaningfully cross-examined about the process of DNA analysis.
“Alleged defects in the chain of custody usually go to the weight of the evidence and not its admissibility.” Commonwealth v. Viriyahiranpaiboon, 412 Mass. 224, 230 (1992), citing Commonwealth v. White, 353 Mass. 409, 419-420 (1967), cert, denied, 391 U.S. 968 (1968). Of course, the prosecution is strongly encouraged to offer conventional chain of custody evidence to establish the source of the sample tested, such as the testimony of the person who recovered the DNA sample, the testimony of the person who had custody of the sample before it was sent to the laboratory, and shipping manifests to show the sample was sent to and returned from the laboratory. See Williams, supra at 2239. Although an expert witness may not be competent to testify to the chain of custody of the particular DNA sample, a DNA match is “itself striking confirmation” that the laboratory tested the correct sample. Id. at 2238 (“there is simply no plausible explanation for how [the laboratory] could have produced a DNA profile that matched [the defendant] if [the laboratory] had tested any sample other than the one [recovered from the crime scene]”).
In the context of DNA forensics, we strongly encourage the Commonwealth, *598as it did in this case, to call a qualified representative of the laboratory that conducted the underlying DNA analysis as its expert witness so as to afford the defendant greater opportunity to cross-examine the expert on the specific laboratory procedures leading to the production of the underlying data.
Although it is clear from the record that Cellmark had a known sample of the defendant’s DNA, it is not clear whether the nontestifying analyst who authored the DNA test reports accessed known samples of the defendant’s DNA while simultaneously interpreting the unknown samples. However, in response to the defendant’s memorandum in support of his motion for a new trial and a supplemental affidavit of Dr. Arthur Eisenberg, a former chairman of the United States DNA advisory board to the Federal Bureau of Investigations whom the defense retained to analyze Cellmark’s work in the present case, Dr. Cotton submitted an affidavit in which she stated that it was “decidedly contrary” to Cellmark’s practice and to accepted scientific practices to access known samples while interpreting unknown samples. Dr. Cotton also stated that, although the risk of “unconscious bias” exists, Cellmark’s protocols are designed to address that possibility, and “it would have been foolish to compromise our integrity by biasing our analysis toward a particular result.” Moreover, the fact that there was also evidence of a stranger’s DNA on a glove recovered from the crime scene, Greineder, supra at 238-239, demonstrates that DNA testing was not geared toward matching the defendant’s DNA profile.
“Only an expert can testify to the likelihood that more than one person possesses a particular DNA profile, based on her knowledge of the alleles selected for the DNA profile and the mathematical probabilities that more than one person may possess the same characteristics of those alleles.” Barbosa, supra at 789. See Williams, supra at 2252 & Appendix at 2253-2255 (Breyer, J., concurring) (distinguishing technician’s production of electropherogram indicating allelic presence at given loci from analyst’s conclusion about significance of DNA test results).
The Committee for Public Counsel Services, as amicus curiae, urges us to abandon our reliance on our evidentiary rule barring inquiry on direct examination about the testimonial hearsay basis of an expert opinion as a means of protecting criminal defendants’ confrontation right. It points us to language in Justice Thomas’s concurrence and the dissent in Williams to the effect that State evidentiary rules should not define constitutional requirements. See Williams, supra at 2256 (Thomas, J., concurring in the judgment); id. at 2272 *601(Kagan, J., dissenting). We decline to change our rule. Certainly, common-law rules of evidence are no substitute for a constitutionally protected right. However, where, as here, a State evidentiary rule provides even greater protection than is required under the Federal Constitution, by definiton, it provides sufficient safeguard of a defendant’s constitutional right.
In many of the cases we have decided since Crawford, supra, that implicated a defendant’s confrontation right in the context of expert opinion testimony, we upheld the defendant’s conviction. See, e.g., Greineder, supra at 239, 256. In Commonwealth v. Durand, 457 Mass. 574, 581 (2010), however, we held that the direct testimony of two doctors to autopsy report findings of a nontestifying examiner was admitted in error, and that the error was not harmless beyond a reasonable doubt. Thus, we reversed the conviction and remanded the case for a new trial. Commonwealth v. Durand, supra at 575, 601.
A Daubert-Lanigan hearing also protects a defendant from the risk of inaccurate forensics. Barbosa, supra at 790-791. At a Daubert-Lanigan hearing, a defendant may challenge the admissibility of an expert opinion that is based on evidence that the defendant expects was somehow tainted or not subject to the expert’s thorough review on ground that an expert must have a “permissible basis” to support his or her opinion. Id., quoting Department of Youth Servs. v. A Juvenile, 398 Mass. 516, 531 (1986). See Williams, supra at 2241 (noting additional safeguards preventing admission or improper use of expert basis evidence); id. at 2249 (Breyer, J., concurring) (professional guidelines ensure reliability of forensic test results).