The defendant, Russell Rumery, was convicted of operating a motor vehicle while under the influence of alcohol, G. L. c. 90, § 24. On appeal, he argues that the judge committed reversible error by (1) allowing two breath test readings from the implied consent form in evidence; and (2) denying the *686defendant’s request for an instruction on “margin of error.”1 We affirm.

On April 5, 2006, a patrol officer stopped the defendant on Route 2 in Erving after initially noticing that the defendant’s vehicle had a loud, defective exhaust. While looking for a safe place to pull the vehicle over, the officer observed the defendant’s vehicle twice cross over the double yellow lines in the road. Upon stopping the vehicle, the officer noticed that the defendant’s eyes were glassy and bloodshot, a number of empty beer cans were in the back cab of the vehicle, and a “strong odor of an intoxicating beverage” emanated from inside the vehicle. The officer then requested that the defendant perform a number of field sobriety tests. Dissatisfied with the defendant’s performance during several of the tests, the officer placed him under arrest. The defendant was taken to the police station where a breathalyzer test was performed. Two breath samples reading 0.08 per cent and 0.09 per cent were recorded.2

At trial, an implied consent form containing both results from the defendant’s breath test was admitted in evidence over his objection. The jury subsequently found the defendant guilty under both theories of statutory liability — operating a motor vehicle while under the influence of intoxicating liquor (impaired ability violation) and operating with a blood alcohol content of 0.08 or greater (per se violation).

Discussion. The defendant argues first that the judge committed reversible error by admitting both breath test results in evidence. In Commonwealth v. Steele, 455 Mass. 209, 212 (2009), *687the court held that, when prosecuting an individual for operating a motor vehicle while under the influence of intoxicating liquor (OUT), the Commonwealth is entitled to introduce only the lower of the two blood alcohol content results produced by the breath testing procedure. Ibid., quoting from 501 Code Mass. Regs. § 2.57 (2006) (“If the two adequate breath samples, expressed in the two decimal mode, differ within +/- 0.02 blood alcohol content units, the lower of the two adequate breath samples shall be taken as the individual under arrest’s blood alcohol level”). The court noted, however, that its holding did not address a situation in which “a defendant presents expert testimony suggesting that his blood alcohol level had risen from the time he operated a motor vehicle to the time the breathalyzer test was administered, and seeks to bolster that testimony with the results of both adequate breath samples” (emphasis added). Steele, supra at 213 n.8.

The Commonwealth contends that we are presented with an analogous situation here, and that it was within the judge’s discretion to allow both readings into evidence. We are unpersuaded by this argument, since the facts in this case are dissimilar to the circumstances contemplated by the Steele court. Although the defendant presented expert testimony permitting the jury to infer that his blood alcohol level could have been rising or falling, he did not seek to bolster that defense by introducing both results into evidence. In fact, the defendant made a timely objection to the introduction of both results at trial. Since such unique circumstances are not present here, there was no basis for the judge to disregard the clear language of the regulations. Accordingly, we conclude that it was error to allow both results into evidence.

Since we have determined that an error was committed, we must consider whether it was prejudicial. See Commonwealth v. Andujar, 57 Mass. App. Ct. 529, 531-532 (2003). An error is nonprejudicial or harmless only if the “conviction is sure that the error did not influence the jury, or had but very slight effect.” Commonwealth v. Flebotte, 417 Mass. 348, 353 (1994), quoting from Commonwealth v. Peruzzi, 15 Mass. App. Ct. 437, 445 (1983). “The standard of appellate review for an objected-to error is, in other words, whether [we] can say with a requisite degree of certitude that the erroneously admitted evidence played *688little or no role in the verdict.” Commonwealth v. Hubert, 453 Mass. 1009, 1009 (2009).

We conclude that the error committed here was harmless.3 Even without the 0.09 reading, there was substantial evidence from which the jury could conclude that the defendant was operating his vehicle while under the influence of alcohol. The jury heard evidence that the defendant was unsteady on his feet, could not maintain his balance, and had twice driven over the center line in the road. More importantly, the properly admitted reading of 0.08, by itself, permitted the jury to conclude that the defendant had a blood alcohol level that was above the legal limit. Furthermore, the judge mitigated his error by instructing the jury that they were to regard the 0.08 reading as the only result of the breath test administered to the defendant. Even after hearing this, the jury convicted the defendant under both the per se and impairment theories of OUI. Considering these circumstances, we conclude that the admission of the marginally higher reading of 0.09 played little or no role in the outcome.

The defendant also contends that the judge erred by denying his request for an instruction on “margin of error.” Specifically, the defendant argues that the jury should have been instructed that the breathalyzer machine had an inherent margin of error of 0.01 per cent.

The Secretary of the Department of Public Safety has promulgated various regulations to ensure the reliability of breath testing procedures. Pursuant to those regulations, a proper breath test “shall consist of a multipart procedure, involving at a minimum, the following sequence: (a) one adequate breath sample analysis, (b) one calibration standard analysis, and (c) a second adequate breath sample analysis.” 501 Code Mass. Regs. § 2.56(3) (2006). *689The “calibration standard analysis” refers to a process by which the Office of Alcohol Testing (OAT) confirms the reliability of a particular machine. In order to be approved by the OAT, a machine must be “capable of analyzing a reference sample of alcohol within +/- 0.01 blood alcohol content units of the known sample.” 501 Code Mass. Regs. § 2.38(2) (2006). The “known sample” is a simulated solution with an official alcohol concentration of 0.155 per cent. Thus, in order for test results to be valid, the calibration analysis must read 0.14 per cent, 0.15 per cent, or 0.16 per cent. 501 Code Mass. Regs. § 2.56(6) (2006). Assuming the calibration standard analysis meets these requirements, the results of the breath test “shall be valid and admissible in a court of law” so long as the “two adequate subject breath samples agree within +/- 0.02 blood alcohol content units.” Ibid.

The defendant’s argument here is flawed for two distinct but equally important reasons. First, the instruction sought by the defendant — that the breathalyzer machine had an inherent margin of error of 0.01 per cent — is factually incorrect. As explained above, the 0.01 per cent variation cited by the defendant is simply a minimum standard, set in the regulations, by which the OAT determines whether a particular machine is reliable. The figure does not represent the actual percent variation of any particular machine, nor does it imply that any machine has such a large margin of error. As set out in an affidavit attached to the Commonwealth’s motion to omit the margin of error instruction, the actual standard of error for all breathalyzer machines used by the Commonwealth is 0.0004 per cent, a figure significantly smaller than 0.01 per cent.4 In addition, the affidavit stated that when the machine in question was calibrated, it tested at 0.155 per cent, a variance of 0.00. Accordingly, an instruction such as the one requested by the defendant would have grossly misled the jury.

Second, the judge was under no obligation to instruct the jury on the intricacies of the breath testing procedure. The two-part procedure used to obtain a defendant’s blood alcohol level, including the calibration analysis, “pertains to the validity of the *690breathalyzer test and does not speak of evidentiary value.” Steele, 455 Mass. at 213. Once the procedure has been conducted under the appropriate conditions previously described, a breath test result is deemed valid and admissible, with the lower of the two truncated samples representing the defendant’s blood alcohol content. By the time a particular result has been admitted in a trial, any “margin of error” has already been accounted for by the multipart procedure. Any further mention of a “margin of error” would surely confuse the jury as well as undermine the purpose of the testing procedure set forth in the regulations.

Finally, if an instruction on margin of error was mandated, experts would be required in every OUI prosecution to provide complex testimony regarding each particular machine’s statistical variance. We decline to place such an unreasonable and onerous burden on litigants and jurors.

Judgment affirmed.