¶1 Tyrone Dentyroll Ford appeals his convictions for second and third degree child rape. We reverse and remand Ford’s conviction on count I for second degree rape of a child and affirm Ford’s conviction on count II for third degree rape of a child. Because we reverse count I, we vacate the lifetime no-contact order (NCO). But because we reverse on count I, we do not reach Ford’s arguments on ineffective assistance of counsel and remaining community custody conditions.
FACTS
¶2 During August and September 2006, Tyrone Ford had sex on two occasions with a minor, L.K.1 The State charged Ford with second degree child rape (count I) and third degree child rape (count II). The different degrees reflected that L.K. was 13 years old during the first incident and 14 years old during the second incident.
¶3 During voir dire, the State asked the prospective jurors if anyone was concerned about his or her ability to be *535fair and impartial. Several jurors raised their hands, including Wiggs and Siciliana. Siciliana stated that as a prior victim of sexual abuse, she might be “slightly biased.” Report of Proceedings (RP) (Aug. 27, 2007) at 39. Similarly, Wiggs stated that her prior experience as a victim of sexual abuse would affect her ability to be fair and impartial. Following further statements regarding their feelings about abuse allegations, the trial court struck both Wiggs and Siciliana for cause.
¶4 After L.K. testified, the State sought to amend the information to conform to the proof because L.K. provided more exact information regarding the incident dates during her trial testimony. Specifically, the State sought to change the first incident date from between September 1, 2006 and September 15, 2006 to on or about August 8. The State also proposed changing the second incident date from September 16 to September 17. Over objection, the trial court granted the State’s motion to file the amended information. The trial court held that filing the amended information would not change the substantive facts of the case.
¶5 After the parties rested, the jury retired to deliberate at 7:47 pm. The next day, the jury returned its verdict at 2:01 pm. The trial court asked the presiding juror if the jury had reached a unanimous verdict, and the presiding juror responded, “Yes.” IV RP at 389-90. The trial court began to read the verdict form, stating that the jury found Ford guilty of third degree child rape on count II. Then the trial court paused to review the documents and called for a sidebar.
¶6 A bench conference occurred off the record.2 The trial court stated, “I’m sending the jury back to the jury room. Verdict form No. 1 is completely blank. It must be filled in,” IV RP at 390, adding, “The defendant is remanded into custody at this time.” IV RP at 390. Next, the trial court stated, “I believe we have just a momentary delay [.] I think they just forgot to fill out the form.” IV RP at 390.
*536¶7 After the brief recess, the trial court stated:
I’m of the opinion that one of two things has happened. They have forgotten to fill in the form. Or in the alternative, they have reached a decision that either means they were deadlocked on Count One or that they reached a not guilty finding on Count One.
I’m inclined to [ ] tell them if they have a question to write the question out and submit it to us. Is that agreeable?
IV RP at 391.
¶8 Both parties agreed. The jurors did not receive this communication, however, because before the trial court could deliver it, the jury was already coming back from the jury room. The trial court asked the presiding juror if the jury had reached a verdict on count I, and the presiding juror responded, “Yes.” IV RP at 391-92. Then the trial court read the verdict form, stating that the jury unanimously found Ford guilty of second degree child rape, as charged in count I.
¶9 After trial, but before sentencing, the trial court granted Ford’s request for a new attorney who would move for a new trial based on his current counsel’s ineffective assistance. The trial court appointed another attorney to represent Ford. That attorney refused Ford’s request to move for a new trial and instead sent a letter to the trial court in which he explained that he would not move for a new trial because he did not have any credible evidence to support Ford’s claim that his trial counsel failed to notify Ford until after the trial had concluded about a plea offer that Ford would have taken.3 Apparently, the first attorney told the newly appointed attorney that he had notified Ford about the plea offer before the end of the trial, and the *537newly appointed attorney did not know who to believe. The record contains no other information about this plea offer or any other pretrial proceeding.
¶10 On count I (second degree child rape), the trial court imposed a minimum sentence of 126 months and a maximum sentence of life in prison. On count II (third degree child rape), the trial court ordered Ford to serve a concurrent 34-month sentence.
¶11 The trial court ordered community custody on count I “for any period of time the Defendant is released from total confinement before the expiration of the maximum sentence.” Clerk’s Papers (CP) at 51. The trial court also ordered community custody on count II for a period of 26-34 months, noting that the combined total time in community custody could not exceed the 60-month statutory maximum for third degree child rape.
¶12 These community custody conditions included prohibitions against (1) possessing alcohol; (2) being in places where alcoholic beverages are the primary sale item; (3) possessing paraphernalia for using or ingesting legal or illegal controlled substances; and (4) possessing, using, or owning “deadly weapons” as defined by a community corrections officer. The community custody conditions required Ford to take a medication called “antabuse” if instructed to do so by a community corrections officer. CP at 61. Additionally, the trial court signed a lifetime no-contact order to restrain Ford from L.K.
ANALYSIS
I. Jury’s Failure To Complete Verdict Form on Count I
¶13 First, Ford argues that the trial court erred in coercing the jury to return a verdict on count I (second-degree child rape) in violation of CrR 6.15. Ford asserts that although the written jury instructions advised the jury that it need not reach a unanimous verdict, the trial judge orally advised the jury that it must reach a verdict and sent the *538jury back to the jury room for further deliberation. Although Ford did not object to this oral instruction during his jury trial, he raises it now on appeal, arguing that he is entitled to a new trial on this basis. We agree.
¶14 We may review an alleged error raised for the first time on appeal if it is a manifest error affecting a constitutional right. RAP 2.5(a)(3); State v. Walsh, 143 Wn.2d 1, 7, 17 P.3d 591 (2001). To raise such an issue on appeal, the defendant must identify a constitutional error and show how the alleged error actually affected the defendant’s rights at trial. State v. Kirkman, 159 Wn.2d 918, 926-27, 155 P.3d 125 (2007). This showing of actual prejudice makes the error “ ‘manifest,’ ” allowing appellate review. Kirkman, 159 Wn.2d at 927.
¶15 Because the trial court is in the best position to determine if an irregularity at trial caused prejudice, we review the decision to grant or to deny a mistrial for an abuse of discretion. State v. Weber, 99 Wn.2d 158, 165-66, 659 P.2d 1102 (1983). An irregularity at trial is not prejudicial unless there is a reasonable probability that the trial’s outcome would have differed if the error had not occurred. State v. Thomas, 150 Wn.2d 821, 871, 83 P.3d 970 (2004). In determining the effect of an irregularity at trial, we examine (1) its seriousness, (2) whether it involved cumulative evidence, and (3) whether the trial court properly instructed the jury to disregard it. State v. Bourgeois, 133 Wn.2d 389, 409, 945 P.2d 1120 (1997). We must decide whether the record reveals a substantial likelihood that the trial irregularity affected the jury verdict, thereby denying the defendant a fair trial. State v. Hicks, 41 Wn. App. 303, 313, 704 P.2d 1206 (1985) (citing State v. Davenport, 100 Wn.2d 757, 762-63, 675 P.2d 1213 (1984)). A “strong, affirmative showing of misconduct is necessary in order to overcome the policy favoring stable and certain verdicts and the secret, frank and free discussion of the evidence by the jury.” State v. Balisok, 123 Wn.2d 114, 117-18, 866 P.2d 631 (1994); Br. of Resp’t at 5.
*539¶16 Citing State v. Boogaard, 90 Wn.2d 733, 585 P.2d 789 (1978), Ford argues that the trial court violated his right to a fair trial by coercing the jury to return a verdict on count I, contrary to CrR 6.15(f)(2). CrR 6.15(f)(2) provides, “After jury deliberations have begun, the court shall not instruct the jury in such a way as to suggest the need for agreement, the consequences of no agreement, or the length of time a jury will be required to deliberate.” This rule’s purpose is to prevent judicial interference with the deliberative process. Boogaard, 90 Wn.2d at 736.
¶17 In Boogaard, a prosecution for second degree theft, the jury began deliberating in midafternoon and had not reached a verdict by 9:30 pm. Boogaard, 90 Wn.2d at 735. Because it was getting late in the evening, the trial court called the jury to the courtroom and asked each juror whether he/she thought it was possible to reach a verdict in one-half hour. Boogaard, 90 Wn.2d at 735. All but one of the jurors answered affirmatively, and the trial judge instructed the jury to return to the jury room and continue its deliberations for one-half hour. Boogaard, 90 Wn.2d at 735. Thirty minutes later, the jury returned a guilty verdict. Boogaard, 90 Wn.2d at 735. After the trial court entered judgment on the verdict, Boogaard moved for a new trial, which the trial court denied. Boogaard, 90 Wn.2d at 735-36.
¶18 On appeal, the Washington State Supreme Court reversed the conviction and remanded, stating:
The questioning of individual jurors, with respect to each juror’s opinion regarding the jury’s ability to reach a verdict in a prescribed length of time, after the court was apprised of the history of the vote in the presence of the jurors, unavoidably tended to suggest to minority jurors that they should give in for the sake of the goal which the judge obviously deemed desirable — namely, a verdict within a half hour.
Boogaard, 90 Wn.2d at 736, 741 (internal quotation marks omitted).
¶19 Here, the trial court stated, “I’m sending the jury back to the jury room. Verdict form No. 1 is completely *540blank. It must be filled in.” IV RP at 390. The trial court’s instruction to the jury directly conflicted with the trial court’s jury instructions, which provided that the jury need not reach unanimous agreement on each charge. Instead, the jury could return guilty verdicts only if it unanimously agreed. Put simply, nothing in the jury instructions here required the jury to render a verdict.
¶20 The dissent makes much of the short time frame of the incident. The jury completed the form and then returned to the courtroom after only several minutes, allegedly supporting the argument that it had merely forgotten to fill in the sheet. But, it is equally possible that the jury walked back to the jury room, determined that it had already spent too much time deliberating, and any holdouts simply acquiesced rather than require the group to start over on count I. This is the same concern that the Boogaard court addressed. Boogaard, 90 Wn.2d at 736.
¶21 The trial court did not ask the jury to clarify whether it found Ford not guilty on form no. 1 or whether it was hung on the issue and openly speculated that either could be the case. The trial court did not ask the jury if it had simply overlooked filling in the verdict sheet. We simply cannot tell why the jury did not fill in the verdict form.
¶22 We note that several Washington cases have addressed blank verdict forms in cases involving lesser-included offenses but, so far, only in the context of double jeopardy challenges. See State v. Daniels, 160 Wn.2d 256, 156 P.3d 905 (2007); State v. Ervin, 158 Wn.2d 746, 147 P.3d 567 (2006). These cases indicate that a blank verdict form did not constitute an implied acquittal barring retrial on those charges. Daniels, 160 Wn.2d at 264; Ervin, 158 Wn.2d at 757, 758. The case here is distinguishable because the trial court specifically instructed the jury, after it had begun deliberations, that it must return a verdict and this case does not involve the lesser-included-offense scenario.
¶23 We hold that under the facts of this case, it is substantially likely that the court’s instruction affected the *541outcome of Ford’s trial. We hold that this constitutes a manifest error affecting a constitutional right. It is also reversible error based on a violation of CrR 6.15(f)(2) and the same concerns the Boogaard court expressed.
II. Lifetime No-Contact Order
¶24 Next, Ford argues that the trial court erred in imposing a lifetime NCO with L.K. as a condition of his sentence. At the conclusion of Ford’s jury trial, the jury found him guilty of second degree child rape (a class A felony) and third degree child rape (a class C felony). Ford assumes that the NCO applied to his class C felony.
¶25 As Ford notes in his briefing, although a lifetime NCO may be appropriate for class A felonies with a maximum term of life in prison, such orders cannot exceed the statutory maximum for the underlying offense. State v. Armendariz, 160 Wn.2d 106, 119-20, 156 P.3d 201 (2007). This provision must be vacated because we reverse count I and it is inapplicable to count II.
STATEMENT OF ADDITIONAL GROUNDS
¶26 In his statement of additional grounds (SAG), Ford argues (1) the trial court erred by allowing the State to amend the information; (2) the trial court allowed “expert-like” statements during voir dire, SAG at 11 (capitalization omitted); (3) his first counsel provided ineffective assistance; and (4) the cumulative error doctrine requires reversal. We have carefully reviewed each claim and hold that Ford’s SAG arguments lack merit.
¶27 Regarding Ford’s argument that the trial court erred by allowing the State to amend the information at trial, we find no error because the act of amending the dates did not change the substance of the offense or the degrees of the offenses.
¶28 Ford’s argument that the “expert-like statements presented during voir dire violated [his] right to an *542impartial jury trial” also fails. SAG at 11 (capitalization omitted). Ford contends that because two prospective jurors (Wiggs and Siciliana) spoke directly about a child’s incapability of lying and the importance of believing survivors, their bias tainted the resulting jury verdict, requiring automatic reversal. We reverse a trial court’s ruling on the scope of voir dire for an abuse of discretion if the defendant demonstrates that the abuse substantially prejudiced his case. State v. Brady, 116 Wn. App. 143, 147, 64 P.3d 1258 (2003) (citing State v. Davis, 141 Wn.2d 798, 825-26, 10 P.3d 977 (2000)), review denied, 150 Wn.2d 1035 (2004).
¶29 During voir dire, two jurors, Siciliana and Wiggs, stated that their past experiences as victims of sexual abuse would affect their ability to be fair and impartial. The trial court struck both Wiggs and Siciliana for cause. Ford did not object to this at trial.
f 30 Ford fails to provide any support for his argument, and instead he focuses on credibility issues, which we will not review on appeal. Thomas, 150 Wn.2d at 874. Even if we did consider his argument, it is evident that the trial court did consider the effects of Wiggs’s and Siciliana’s statements on the other prospective jurors. Indeed, the trial court removed both of these prospective jurors for cause.
¶31 Neither do we agree with Ford’s contention that these statements were “expert-like.” SAG at 11 (capitalization omitted). Wiggs and Siciliana made these statements based on their personal experiences with sexual abuse; neither of these women purported to offer an expert opinion. Additionally, Wiggs and Siciliana fully disclosed their viewpoints on sexual abuse during voir dire. We find no error here.
f 32 We find no merit in Ford’s argument that his first trial attorney rendered ineffective assistance in failing to (1) request a continuance when the trial court granted the State’s motion to amend and (2) conduct proper voir dire. To demonstrate ineffective assistance of counsel, a defendant must make two showings: (1) that counsel’s representation was deficient and (2) that counsel’s deficient *543representation caused prejudice. State v. McFarland, 127 Wn.2d 322, 334-35, 899 P.2d 1251 (1995). To show prejudice, a defendant must show that there is a reasonable probability that but for counsel’s unprofessional errors the result of the proceeding would have differed. State v. Crawford, 159 Wn.2d 86, 99-100, 147 P.3d 1288 (2006). In general, performance is deficient when it falls below an objective standard of reasonableness, but not when undertaken for legitimate reasons of trial strategy or tactics. State v. Horton, 116 Wn. App. 909, 912, 68 P.3d 1145 (2003) (citing McFarland, 127 Wn.2d at 334).
¶33 Ford argues that his first counsel rendered ineffective assistance by failing to request a continuance after the trial court allowed the State to file an amended information. Ford argues that this was not a legitimate or tactical choice because the attorney needed more time to assess the amendment’s effects before starting trial. But the State’s amendment made no substantive change to the offenses charged in the original information and had no effect on the trial court’s proceedings. Consequently, his attorney’s decision to proceed to trial was legitimate.
¶34 Ford argues that his first counsel rendered ineffective assistance by conducting improper voir dire. Ford contends that his attorney not only allowed two jurors to taint the jury pool with their “inflammatory and highly prejudicial statements,” but he also provoked Siciliana’s statement that “ The most important thing you can do to support survivors is to believe them.’ ” SAG at 24 (quoting RP (Aug. 27, 2007) at 49). Contrary to Ford’s argument, however, the attorney’s performance was not deficient because (1) Wiggs and Siciliana primarily discussed their experiences with sexual abuse in response to the State’s questions; (2) in asking Siciliana to clarify what she believed, Ford’s attorney did not directly provoke her response about survivors; and (3) the attorney’s decision to question Siciliana about her experience with sexual abuse was a proper tactical decision, considering that the purpose of voir dire is to determine whether any of the prospective *544jurors would have difficulty returning a fair and impartial verdict. Accordingly, Ford’s ineffective assistance of counsel claim fails.
¶35 Finally Ford argues that, “[t]aken together, the numerous errors in this case violated [his] right to due process. SAG at 26 (capitalization omitted). The cumulative error doctrine mandates reversal when the cumulative effect of nonreversible error(s) materially affects the trial outcome. State v. Newbern, 95 Wn. App. 277, 297, 975 P.2d 1041, review denied, 138 Wn.2d 1018 (1999). We find no cumulative error.
¶36 We reverse Ford’s count I conviction, vacate the lifetime NCO, and remand for a new trial. We affirm Ford’s count II conviction and remand for resentencing as to count II.
Armstrong, J., concurs.This court refers to victims by their initials.
The respondent’s brief also contains this portion of the trial record.
In the attorney’s letter, he stated:
Without further information of a credible nature, I cannot make a determination on the defendant’s claim in this regard. Therefore, I do not feel that I can in good conscience bring an ineffective assistance claim as part of a motion for a new trial at this time. Again, appellate counsel may be able to develop this claim further, if new evidence presents itself.
Clerk’s Papers at 43.