In this appeal from convictions of involuntary manslaughter by commission of public offense and child endangerment resulting in death, we primarily consider a claim of ineffective assistance of trial counsel based on the failure to challenge the sufficiency of evidence to support the submission of all four alternative means of committing the crime of child endangerment. On our review, we conclude trial counsel was ineffective, and a new trial must be granted. We affirm the decision of the court of appeals in part and vacate in part, reverse the judgment and sentence of the district court, and remand the case for a new trial.
I. Background Facts and Proceedings.
Zyriah Schlitter met Nicole King in 2006 and they entered into a relationship. They began sharing a residence and eventually had a daughter, K.S., on September 23, 2008. The relationship ended in late 2009. In February 2010, Schlitter and King agreed that Schlitter would be the temporary primary custodian of K.S. Schlitter was living with his grandparents at the time. He was also dating a woman named Amy Parmer. Schlitter would often stay overnight at Parmer’s apartment. Parmer had two children.
On March 1, 2010, Schlitter took K.S. to a medical clinic for a health checkup required for admission to a day-care center. A clinic nurse updated K.S.’s vaccines and found her to be in good health.
' K.S. was accepted by the day-care facility on March 2 and attended day care for the remainder of the workweek. Schlitter and K.S. then stayed with Parmer and her children at her apartment over the weekend. Parmer cared for K.S. on Sunday evening while Schlitter attended a financial management class at church for a couple of horns.
On Monday morning, March 8, Schlitter dropped K.S. off at the day-care center. Later that morning, a day-care worker *385observed a bruise on KS.’s forehead and around one eye. She also saw marks on the side of K.S.’s chin and discovered makeup had been applied to cover up the bruises. Parmer stopped by the day-care center during the afternoon to check on K.S. and was asked about the injuries. Parmer said K.S. bruised her eye from a fall and was accidentally struck on the forehead by a Pack ’n Play® falling out of a closet.
Schlitter did not take K.S. to the daycare center on March 9. K.S. had a fever, and Schlitter took her to the medical clinic. He told a nurse that KS. had not been sleeping well and had little appetite. The nurse inquired about the bruise on her forehead. Schlitter responded that K.S. fell into a coffee table. K.S. was diagnosed with conjunctivitis and prescribed Motrin® and eyedrops.
Over the next few days, Schlitter’s father and grandparents provided day care for K.S. K.S. would cling to Schlitter when he was present. On March 10, K.S. had a fever of 104°. Schlitter called the clinic to report the fever. He was told to continue the Motrin® and eyedrops and to call the next day if there was no improvement. On March 11, Schlitter called the clinic to report that K.S. vomited. He also reported the Motrin® would only briefly keep her fever under control, and an appointment was scheduled for the next day. Schlitter took K.S. to the clinic on March 12. Medical providers diagnosed K.S. with an ear infection and prescribed an antibiotic. No new bruising was observed. On March 13, her temperature returned to normal.
Schlitter and KS. again stayed at Parmer’s apartment on the weekend. King exercised visitation with K.S. for a period of time. She did not notice any bruises on her face or body.
Schlitter dropped KS. off at the daycare center on Monday, March 15. Workers at the center again observed bruising on her forehead and face. K.S. acted listless and sad. She slept more than normal, did not play, and did not want to interact. When Schlitter was asked about the new facial bruises, he responded that KS. liked to beat on herself. Workers at the daycare center reported their observations to the Iowa Department of Human Services (DHS). An investigator for the DHS met with Schlitter on March 16, Schlitter admitted to spanking K.S. and told the investigator that K.S. listened better to Parmer. K.S. was not removed from Schlitter’s care.
On March 17, K.S. spent the day with King. K.S. was detached and often cried. Schlitter called the medical clinic on March 18 to report that KS. was very sleepy. The next day, her condition seemed to improve.
Schlitter and KS. again spent the weekend with Parmer. On Sunday, March 21, KS. was sleepy, and she often cried. She also clung to Schlitter. At 5:15 p.m., Schlitter left KS. in the care of Parmer so he could attend the Sunday evening financial management class.
At' 7:45 p.m., Parmer called 911 and reported that K.S. was barely breathing. An ambulance arrived at the apartment and transported K.S. to a hospital in Cedar Rapids. Medical personnel at the hospital found her in a decorticate posture. Her pupils were fixed and dilated. The doctor observed hemorrhages in her eyes. She exhibited limited reaction to pain stimuli. After the doctors told Schlitter that her injuries were likely the result of child abuse, he entered the room where KS. was being treated and told her “I’m sorry.”
KS. was promptly airlifted to the University of Iowa Hospitals & Clinics. Fam*386ily members gathered to be with-her and tension surfaced between King and Schlit-ter. King blamed Parmer for the injuries, and Schlitter blamed the daycare.
Medical tests and scans of KS.’s brain 'showed significant swelling. Despite extensive medical efforts, K.S.’s condition continued to deteriorate. She remained in a coma, which doctors believed would likely never change. KS. was kept alive by a ventilator and a feeding tube. On Sunday, March 28, King and Schlitter agreed to the removal of life support systems.' KS. died.
On July .11, 2011, the State charged Schlitter and Parmer with murder in the first degree and child endangerment resulting in death. The trials were severed, and Schlitter went to trial on December 3, 2012. .
The medical testimony at trial described the injuries to K.S. as nonaccidental or abusive trauma. The medical professionals generally agreéd that KS. had suffered multiple head trauma events, The testimony came from the emergency room doctors and nurses at both hospitals, as well as neurologists, pathologists,- an ophthalmologist, radiologist, doctors in the pediatric intensive care unit, and the head of the child protection team.
The external injuries included bruises on KS.’s cheeks and under her chin; scrapes or red marks on her left shouldef, the nape of her neck, her left ear and cheek, upper right portion of her chest, and her right underarm area; contusions on her right upper arm and left' and right inner thighs; and an infected lesion on her left labia. KS.’s internal injuries included subdural hematomas in the brain and around the spinal cord,-as well as other brain injuries. An MRI revealed that KS. suffered a massive stroke on the left side of her brain, but revealed no evidence that KS.’s injuries were caused by strangulation.
Multiple doctors testified that the bruises on KS.’s face and body were different colors, indicating they had occurred at different times. The retinal hemorrhages, brain blood clots, and subdural membranes indicated injuries that could be up to a few weeks old. The blood found in her brain showed signs of fresh bleeding, older bleeding that had happened over two days before the recent injury, and a recent bleeding within hours of KS. becoming symptomatic. Moreover, .with repeated injuries to the same part of the brain, some of the new injury clouded evidence of the older injury.
The time frames suggested by the different doctors’ testimonies sometimes conflicted, The estimates ranged from minutes to hours, within a day, a twelve- to twenty-four-hour period estimate, and an “hours to days” time frame. One doctor stated she could not accurately estimate the timing, but that she had not seen any child awake with the kinds- of injuries found in KS. Almost all the medical professionals were clear that a specific time of the injuring event could not be pinpointed due to individual-specific rates of healing, the age of the patient, an unknown rate of bleeding, and uncertainty concerning the number-and frequency of injuries.
Dr. Resmiye Oral, the head of the child protection team, specializes in treating and consulting in cases of child abuse. She met with a statewide multidisciplinary team made up of the physicians, law enforcement, DHS employees, and medical examiners involved in KS.’s case. Dr. Oral, collected all of the reports of the physicians and examiners to make a* final medical determination regarding KS.’s injuries. Dr. Oral concluded that K.S. suffered at least two separate episodes of injury. She pinpointed the first injury as *387likely occurring one or two weeks before KS. entered the hospital, and the second injury as inflicted from minutes up to six hours before K.S. was brought to the hospital, noting the shorter time frame was more likely than the longer. The doctor stated it must have been an acute and forceful trauma to explain the injuries found.
The paramedic who responded to the 911 call on March 21 testified to statements made by Parmer in response to questioning about the condition of K.S. Parmer said she found K.S. unresponsive and struggling to breathe. She told a paramedic K.S. had a fever earlier in the week, but was .unaware of any falls or injuries.
Law enforcement investigators conducted several interviews with Schlitter and Parmer. Schlitter gave one interview at an Iowa State Patrol Office on March 30, 2010. During the interview, Schlitter acknowledged he was rough on K.S. at times in his discipline of her and was probably incriminating himself by maintaining that Parmer was a good caretaker and would not have harmed K,S. On another occasion, Schlitter told one investigator that while at the hospital, he had researched head trauma symptoms and that K.S. had exhibited some of the symptoms during the period of time prior to her hospitalization. Prior to trial, Schlitter had moved to suppress his statements made to law enforcement investigators during the March 30 interview at the state patrol office. The district court denied the motion, and the interview was entered into evidence.
Investigators also discovered Parmer had made inculpatory statements to two people. On one occasion, Parmer made a spontaneous statement to a coworker that she “might have killed a kid.” Another time, Parmer was in her apartment with the coworker and a man she was dating. Parmer suddenly started crying and-told the man, “You don’t want to get involved with me.” She then explained that she had taken an eighteen-month-old’s, life. She further explained that the child was K.S., and it involved a head injury.
At the close of all the evidence at trial, trial counsel for Schlitter moved for a judgment of acquittal. The motion, however, was limited to the sufficiency of the evidence to support the crime of first-degree murder. The trial court overruled the motion.
The jury found Schlitter guilty of involuntary manslaughter by commission of public offense (child endangerment) and child endangerment resulting in death. A general verdict was returned, and the jury did not identify the alternative theories relied , upon to support the guilty verdict for child endangerment. Schlitter moved for a new trial and arrest of judgment. After a hearing on February 20, 2013, the trial court denied Schlitter’s motions and sentenced Schlitter. The court merged the sentences for the two charges under the one-homicide rule. It imposed a mandatory indeterminate fifty-year sentence for child endangerment resulting in death and ordered $150,000 restitution to be paid to Nicole King. Although the State requested a thirty-year minimum sentence before parole eligibility, the court declined to require a minimum sentence before parole eligibility, leaving that question to the board of parole.
Schlitter appealed and raised four claims of error. First, he claimed the district court erred in failing to suppress his statements made during the interrogation on March 30. Second, he claimed his trial counsel was ineffective for failing, to challenge the- sufficiency of evidence to support the lesser included offense of involuntary manslaughter to the charge of first-degree murder and the alternative theories to the *388crime of child endangerment. Third, he claimed his trial counsel was ineffective for failing to timely object to improper comments by the prosecuting attorney during closing argument. Finally, he claimed his trial counsel was ineffective for failing to investigate properly.
We transferred the case to the court of appeals. It affirmed the judgment and sentence of the district court. It found Schlitter was not in custody during the interrogation on March 30, and the law enforcement officers were not required to give Schlitter his Miranda warnings. It also found Schlitter failed to preserve error on his secondary claim that the statements were involuntary. The court of appeals further found that trial counsel was not ineffective because sufficient evidence was presented to support all the charges. It also found trial counsel was not ineffective because, even if the prosecutor’s statements amounted to misconduct, no prejudice resulted. Finally, it found trial counsel was not ineffective for failing to conduct a proper investigation.
Schlitter sought, and we granted, further review. The only issue Schlitter raised was that his trial counsel was ineffective for failing to move for a judgment of acquittal for the crimes for which he was convicted.
II. Scope of Review.
Ineffective-assistance-of-counsel claims are reviewed de novo. State v. Tompkins, 859 N.W.2d 631, 636 (Iowa 2015). Ineffective-assistance-of-counsel claims require a showing by a preponderance of the evidence both that counsel failed an essential duty and that the failure resulted in prejudice. Anfinson v. State, 758 N.W.2d 496, 499 (Iowa 2008). We review sufficiency-of-the-evidence challenges for correction of errors at law. State v. Neiderbach, 837 N.W.2d 180, 190 (Iowa 2013).
We review constitutional issues, including Miranda violations, de novo. See State v. Kooima, 833 N.W.2d 202, 205 (Iowa 2013). We examine the totality of the circumstances in the entire record in our evaluation. State v. Baldon, 829 N.W.2d 785, 789 (Iowa 2013).
III. Analysis.
The right to effective assistance of counsel stems from the general right to counsel under the Sixth Amendment to the United States Constitution and article I, section 10 of the Iowa Constitution. State v. Ambrose, 861 N.W.2d 550, 556 (Iowa 2015). “To succeed on a claim of ineffective assistance of counsel, a claimant must establish by a preponderance of the evidence: ‘(1) his trial counsel failed to perform an essential duty, and (2) this failure resulted in prejudice.’ ” State v. Thorndike, 860 N.W.2d 316, 320 (Iowa 2015) (quoting State v. Adams, 810 N.W.2d 365, 372 (Iowa 2012)). The claimant must establish both elements of the claim. Dempsey v. State, 860 N.W.2d 860, 868 (Iowa 2015).
For the first element, we presume the attorney performed competently, requiring the claimant to rebut the presumption with evidence the attorney performed outside the standard of a reasonably competent practitioner. Id. To prove prejudice for the second element, the claimant needs to show the attorney’s errors functionally deprived the defendant of a fair trial and further show by a reasonable probability that the result of the proceeding would have been different without the errors by the attorney. State v. Ross, 845 N.W.2d 692, 698 (Iowa 2014).
A. Failure to Move for Judgment of Acquittal on Child-Endangerment Alternatives. “To preserve error *389on a claim of insufficient evidence^ a] defendant must make a motion for judgment of acquittal at trial....” State v. Truesdell, 679 N.W.2d 611, 615 (Iowa 2004). The motion must be made after the evidence on either side of the case has been presented. Iowa R.Crim. P. 2.19(8)(a). ,
When presented with a motion for acquittal, courts must view “the evidence in the light most favorable to the State and draw[] all fair and reasonable inferences from it, taking all the evidence into consideration, both direct and circumstantial.” State v. Duncan, 312 N.W.2d 519, 522 (Iowa 1981) (citations omitted). This standard requires courts to assume the truth of the evidence offered by the prosecution. Nguyen v. State, 707 N.W.2d 317, 327 (Iowa 2005). The evidence must be sufficient to convince a rational fact finder that the defendant is guilty beyond a reasonable doubt. State v. Shanahan, 712 N.W.2d 121, 134 (Iowa 2006). A fair inference of guilt is necessary, not merely suspicion, speculation, or conjecture. State v. Geier, 484 N.W.2d 167, 171 (Iowa 1992).
Counsel for Schlitter did not challenge the sufficiency of the evidence to support any of the alternative theories of guilt for a finding of child endangerment. We must consider if he failed to perform within “the range of normal competency” by determining if a competent attorney would have challenged the sufficiency of the evidence. State v. Graves, 668 N.W.2d 860, 881 (Iowa 2003). If counsel failed to raise a meritorious issue a normally competent attorney would have raised, and such failure cannot “be attributed to reasonable trial strategy, then we can conclude the defendant has established that counsel failed to perform an essential duty.” Id. at 870.
In its case against Schlitter, the State presented four alternatives of guilt to the jury on the charge of child endangerment. The trial court instructed on each alternative. The jury was told they could find Schlitter committed child endangerment if they found he had done any of the following alternatives:
a. Knowingly acted in a manner that created a substantial risk to [K.S.] ’s physical health or safety; or
b. By an intentional act or series of intentional acts used unreasonable force that resulted in bodily injury or was intended to cause serious injury; or
c. Willfully deprived [K.S.] of necessary supervision or medical care appropriate to her age, being reasonably able to make such necessary provisions, which deprivation substantially harmed [K.S.] ’s physical health; or
d. Knowingly permitted the continuing physical abuse of [K.S.].
Counsel did not move for judgment of acquittal on any of the alternatives presented based on insufficient evidence but, rather, conceded to a jury question on the child endangerment charge. Thus, if the evidence presented by the State at trial was insufficient to support any alternative, Schlitter’s trial counsel would have provided ineffective assistance by failing to raise the issue and permit the trial court to enter a judgment of acquittal on any alternative not supported by sufficient evidence.
We often do not address ineffective-assistance-of-counsel claims on direct appeal because a record is needed to fully develop the claim and identify the existence of any trial strategies that may have influenced the actions or inactions of trial counsel. See State v. Ondayog, 722 N.W.2d 778, 786 (Iowa 2006) (“[P]ostconviction proceedings are often necessary to discern the differ*390ence between improvident trial strategy and ineffective assistance.”)- However, no reasonable trial strategy could permit a jury to consider a crime not supported by substantial evidence. See State v. Brubaker, 805 N.W.2d 164, 174 (Iowa 2011). (holding counsel ineffective for failing to move for judgment of acquittal based on insufficient evidence to support a necessary element of the charged crime and noting such a failure “is not a trial strategy”). Therefore, we must review each alternative theory of the crime of child endangerment to determine if a reasonable trial counsel would have moved for judgment of acquittal on any of the four alternatives.
1, Knowingly acted in a manner that created substantial risk to physical health or safety. We first consider the sufficiency of evidence to support a finding that Schlitter knowingly acted in a manner that created a substantial risk to the physical health or safety of K.S. The term “knowingly” not only refers to the act, but also the creation of a substantial risk-to physical health or safety. State v. James, 693 N.W.2d 353, 356-57 (Iowa 2005). Additionally, “the definition of ‘substantial risk’ in the context of child endangerment” means “[t]he very real possibility of danger to a child’s physical health or safety.” State v. Anspach, 627 N.W.2d 227, 233 (Iowa 2001), The risk does not have to be likely, just real or identifiable. Id. at 232-33. The evidence offered by the State at trial targeted- Schlitter either as the abuser or complicit in abuse inflicted by Parmer by failing to intervene to stop or prevent it.
The State presented an -abundance of evidence that K.S. sustained bruises to her head on separate occasions in the weeks leading up to her death. A number of people noticed bruises the week of March 8 — including family members, day-care workers who saw her every day, and a nurse practitioner. Evidence was presented that either Schlitter or Parmer used makeup to cover bruising around KS.’s eye and forehead. One week later, a new bruise appeared on KS.’s forehead in the same location as the previous bruise.
Construing the evidence in favor of the State, a reasonable jury could find beyond a reasonable doubt that* even if Parmer was the abuser instead of Schlitter, he knew that K.S. was. at risk - of physical injury while in the sole care of Parmer. The jury could have, also found Schlitter knowingly acted in a manner that created a substantial risk to the physical health or safety of K.S. by leaving her in the care of Parmer, .
2, By an intentional act or series of intentional acts us'ed unreasonable force that resulted in bodily injury or was intended to causé' serious injury. ’ To prove the second alternative, the State must present sufficient evidence that Schlitter either committed an act resulting in the injury or had sole care of K.S. during the time in which the injury occurred. See Neiderbach, 837 N.W.2d at 219. The evidence presented at trial clearly supported a finding that a series of intentional acts of unreasonable force were inflicted on K.S. and that these acts resulted in the bodily injury she suffered. However, the evidence does not reasonably support a finding either that Schlitter committed the violent acts or that he had sole care of her when the injuries were sustained-. -During the period of time prior to discovery of the first bruises on K.S., numerous people other than Schlitter' had cared for her. These caretakers included Parmer, day-care providers, King, and several members of Schlitter’s family. Likewise, KS. had been in the care of several people prior to the time the second set of bruises was discovered. Additionally, K.S. had been in the . care of both , Schlitter and *391Parmer prior to the injuries that led to KS.’s hospitalization and death. Finally, Schlitter was not with K.S. during the two hours prior to the 911 call. There was no testimony that Schlitter had ever inflicted unreasonable force on K.S. in the past or that he had ever shaken her. To the contrary, the evidence was consistent that Schlitter may have yelled at her when frustrated, but he typically would leave the room to cope with his frustration.
In our careful consideration of all the evidence in the light most favorable to the State, we cannot conclude that a reasonable jury could find Schlitter inflicted the force on K.S. that resulted in her injuries. Such a finding could only be based on speculation. Speculation and conjecture cannot be used' to support a verdict. See State v. Webb, 648 N.W.2d 72, 76 (Iowa 2002) (“The évidence must raise a fair inference of guilt and do more than create speculation, suspicion, or conjecture.”). Thus, the second alternative could not support a guilty verdict for child endangerment, and Schlitter’s counsel was ineffective for failing to move for a judgment of acquittal on this alternative. The jury should not have been instructed to consider this- alternative in considering Schlitter’s- guilt, and Schlitter’s trial-counsel failed to perform an essential duty by failing to object’ to the submission of the alternative to the jury. Furthermore, we also find from this record that prejudice resulted to Schlitter when his trial counsel failed to move for a judgment of acquittal on this alternative. It is not- possible to know whether or not the jury relied on this alternative in reaching its verdict. See State v. Tyler, -873 N.W.2d 741, 753-54 (Iowa 2016) (holding we reverse a general verdict when not all theories are supported by sufficient evidence). Consequently, there is- no way to know if the jury're-framed from relying on this alternative in reaching their verdict.
Accordingly, Schlitter must be given a new trial based on ineffective assistance of counsel. A new trial cannot include the second alternative theory for the' crime of child endangerment.
3- Willfully deprived K.S, 'of necessary supervision or medical care. We now proceed to consider the sufficiency of the evidence to support the remaining two alternatives of child endangerment. If insufficient evidence was not presented, the alternative cannot be submitted at the new trial.
The third alternative required proof that Schlitter willfully deprived K.S. of necessary supervision or medical care he was reasonably able to provide and the deprivation substantially harmed her physical health. “Willfully” is defined either as “said of done deliberately or intentionally” or “established by proof of intentional and deliberate conduct undertaken with a bad purpose, in disregard for .the rights of another, or contrary to a known duty.” State v. Leckington, 713 N.W.2d 208, 214 (Iowa 2006) (quoting State v. Tippett, 624 N.W.2d 176, 178 (Iowa 2001) (first quote)) (finding either definition appropriate for this subsection of the child endangerment statute in that particular case). In Leck-ington, the defendant saw an intoxicated minor suffer an injury, left him alone in an unsupervised location without healthcare, and then tried to rémove him from her house while he was unconscious and foaming from the mouth rather than call for help in an effort to avoid a criminal investigation. Id. at 214-15, We found.the delay and the seriousness of the minor’s com dition satisfied- the requirement of willful deprivation of medical care. Id. at 215.
In this case, there was evidence that KS. exhibited numerous signs of abuse and head trauma. - On the other hand, she *392also exhibited signs of more normal childhood illness or infection. Schlitter took K.S. to the doctor on numerous occasions and called the medical clinic several times. He also administered medication prescribed by the doctor. Schlitter, however, did not seek medical care for KS.’s most serious symptoms. The doctors testified at trial that the symptoms of head trauma would have been obvious to anyone. In particular, Dr. Oral testified that the symptoms exhibited by K.S., such as lethargy, decreased appetite, pulling hair, nightmares, multiple bruises from distinct time periods, and lack of playfulness even after she had healed from the conjunctivitis and ear infection, combined with the repeated injuries' to her forehead were far enough outside normal child behavior that a reasonable caretaker would have sought medical care.
We conclude a reasonable jury could have found that Schlitter knew of the abuse occurring to K.S. and chose not to seek medical attention for the resulting injuries, such as facial bruising and other abnormal symptoms. A reasonable jury could have found Schlitter willfully deprived K.S. of medical care despite the ongoing symptoms of excessive sleep and failure to eat. It could have further found that Schlitter purposely did not take K.S. in for treatment to avoid the risk of exposure and an investigation, a risk he knew was possible after the March 16 visit with DHS regarding the bruise on KS.’s forehead.
4. Knowingly permitted the continuing physical abuse of K.S. Finally, we consider the alternative that Schlitter committed child endangerment by knowingly permitting the continuing physical abuse of K.S. In State v. Watkins, we held that continuous proximity to a child abused by a person was sufficient to find a defendant knowingly permitted the continuing physical abuse of the child. 659 N.W.2d 526, 536-37 (Iowa 2003). To make its case, the State had to show Schlitter.actually knew Parmer was abusing K.S., not just that K.S. always, ended up with odd, significant bruises after her care, even if plausible explanations for the bruises existed.
Construing the evidence in a light most favorable to the State, this alternative was supported by sufficient, evidence. Schlit-ter’s explanations for the origin of the forehead bruises were not consistent, and he provided no reason for his inconsistencies. Further, a reasonable jury could find he knew the bruising on March 8 was covered with makeup in an attempt to hide the injury. A jury could also reasonably believe a parent would not seek to hide bruises on a toddler with makeup. Construing the evidence in the light most favorable to the State, the jury could have inferred that K.S. was being abused and that Schlitter knowingly permitted the abuse to continue by failing to take action to remove her from the care of the abuser.
B. Failure to Move for Judgment of Acquittal on the Lesser Included Offenses of Murder. Schlitter also claims his trial counsel should have sought an acquittal on the lesser offense to murder of involuntary manslaughter by public offense because the State failed to establish sufficient evidence to prove the public offense of .child endangerment. Even if we recognized a duty to move for judgment of acquittal on lesser -included offenses after denial of a motion to acquit on the greater offense, because we find sufficient evidence to support three of the alternatives of child endangerment, this claim must fail.
C. Claim of Error by Prosecutor. We proceed to consider other issues raised by Schlitter on appeal to determine if they will impact the retrial. Because a new trial will be necessary, we will exercise our *393authority to promote efficiency and judicial economy by addressing those issues raised on appeal that will likely reoccur at the retrial.
During closing argument, the prosecutor made an emotional appeal to the jurors by telling them that the jury system gives control to “citizens to hold each other accountable for criminal behavior.” He also told the jurors that they had the “sacred duty of protecting the safety of the public and of the innocent by judging those that commit brutal acts of abuse and neglect against fellow humans to be guilty when it’s been shown beyond a doubt that’s reasonable.” Additionally, the prosecutor informed the jurors that they had an “important honor” to “protect the rights of citizens and acknowledge those rights and find [offenders] accountable through the rest of us.”
Counsel for Schlitter objected to these statements after deliberations had begun. The district court ultimately found the statements did not amount to prosecutorial misconduct. Based on this ruling, Schlit-ter raised a claim of prosecutorial misconduct on appeal.
At the outset, we observe that the term “prosecutorial misconduct” has gained a specialized meaning within the law:
to describe conduct by the government that violates a defendant’s rights whether or not that conduct was or should have been known by the prosecutor to be improper and whether or not the prosecutor intended to violate the Constitution or any other legal or ethical requirement.
ABA House of Delegates, Recommendation 100B, at 1 (2010), http://www. americanbar.org/content/dam/aba/dir ectories/policy/2010_am_100b.pdf [hereinafter ABA Recommendation], We have followed this approach by broadly describing trial conduct of a prosecutor in a criminal case that is claimed to deprive the defendant of a fair trial to be prosecutorial misconduct. See Graves, 668 N.W.2d at 870. The range of trial conduct by prosecutors falling into the category of claims referred to as “prosecutorial misconduct” includes questioning witnesses about others’ deceit, distorting testimony, making unsupported statements during closing argument, stating the defendant lied during testimony, diverting the jury from deciding the case based on the evidence, making other inflammatory or prejudicial statements about the defendant, and more. State v. Musser, 721 N.W.2d 734, 754-55 (Iowa 2006) (referring to improper closing argument that urges the jury to decide the case on something other than the evidence as prosecutorial misconduct); State v. Carey, 709 N.W.2d 547, 552 (Iowa 2006) (referring broadly to claims of improper closing argument by the prosecutor as claims of misconduct); Graves, 668 N.W.2d at 870-71 (collecting cases). While some of the conduct in these cases may have been intentional, other conduct can be the result of mistake or error during the heat of trial.
The problem with describing all claims as prosecutorial misconduct is that the term tends to conflate prosecutorial misconduct with professional misconduct as controlled by our Iowa Rules of Professional Conduct. ABA Recommendation, at •1; Shawn E. Minihan, Measuring Prose-cutorial Actions: An Analysis of Misconduct versus Error, Prosecutor, Dec. 2014, at 22, 23 [hereinafter Minihan]; see also Iowa R. Profl Conduct 32:8.4 (defining professional misconduct). The two phrases are not only similar in their language, but tend to connote similar meanings. Yet, professional misconduct generally applies to intentional misbehavior on the part of the attorney, while prosecutorial misconduct is not always intentional. Iowa R. *394Prof 1 Conduct 32:8.4; ABA Recommendation, at 2. In 2010, the American Bar Association (ABA) adópted a recommendation urging courts to be careful in distinguishing between prosecutorial misconduct and prosecutorial error and to attach different levels of culpability for each. ABA Recommendation, at 2-3.
One author has offered helpful guidance on how to distinguish between prosecutorial misconduct and pros-ecutorial error.1 Minihan, at 24-25. Prosecutorial - misconduct includes those statements-“where a prosecutor intentionally violates a clear and unambiguous obligation or standard imposed by law, applicable rule or professional conduct,” as well as “those situations where a prosecutor recklessly disregards a duty to comply with an obligation or standard'.-” Id. Prosecutorial error occurs “where the prosecutor exercises poor judgment” and “where the attorney has made a mistake” based on “excusable human error, despite the attorney’s use of reasonable care.” Id. at 25. This distinction also conforms to the general definitions for misconduct and a trial error. Compare Misconduct, Black’s Law Dictionary (10th ed.2014) (defining misconduct as “[a]n- attorney’s dishonesty or attempt to persuade a court or jury by using deceptive, of reprehensible methods”), with Trial Error, Black’s Law Dictionary (defining trial error as “[a] mistake in' or deviation from propef trial procedure during the presentation of a case to a jury”). Going forward, we adopt the ABA’s recommendation on our review of prosecutorial behavior and distinguish between incidences of prosecuto-rial error and prosecutorial misconduct. A prosecutor who has committed error should not be described as committing misconduct.
We discussed the role of the prosecutor in criminal cases hi Graves, 668 N.W.2d at 870. We also identified a multifactor test to evaluate the statements made during closing arguments in determining if there wás misconduct and_ if that misconduct was prejudicial. Id. at 877-78. These same factors easily translate to an evaluation of prosecutorial error.
In this case, the claim raised by Schlit-ter was actually describing error by the prosecutor, not prosecutorial misconduct. It is unnecessary, however, for us to apply the Graves factors to this claim or to address the additional -claim whether trial counsel was ineffective for failing to lodge a timely objection to the closing argument of the prosecutor. The claim of error by the .prosecutor based on the statements made during closing argument rests with the unique and particular choice of words, as well as the particular surrounding circumstances. It is unlikely the prosecutor will make the same choice of words or that the same circumstances will be repeated during the retrial. Accordingly, we do hot resolve the issue, but remind counsel on retrial to be mindful of the scope of closing arguments described in Graves.
D. Miranda Violation. Schlitter was interviewed by law enforcement officers on several occasions, including an interview at a state patrol office on March 30, 2010, He moved to suppress statements made to officers during this interview because he was not given the Miranda warnings and because his statements were involuntary *395based on promises of leniency. In particular, at the suppression hearing, Schlitter’s objections to the March 30 interview centered on two areas. First, he objected to the nature of the- interrogation. Second, he objected because the officers continued to question him after he asked them to stop once they began to graphically describe the possible ways K.S. could have received her injuries. On appeal, however, Schlitter primarily objected to the admission of his statements from the interview describing his frustration with K.S., the possibility that he had picked up K.S. roughly, and his implicit defense of Parmer.
The district court found the officers made no statements that resembled any promise of leniency. It also found Schlit-ter was not in custody during the interview, and the officers were not required to give him the Miranda warnings. The court held Schlitter was not in custody because he was allowed to and did leave the interview at his own will. This issue will be ’ raised again on retrial, and we proceed to resolve it on this appeal. In doing so, we agree the , record does not disclose any promises of leniency. Thus, we proceed to decide if Schlitter was in custody at any.time during the interview.
We begin by recognizing that Schlitter raised the Miranda issue under both-'the United States and Iowa Constitutions. He did not propose, however, that we consider a different standard for determining whether he was in custodyunder the Iowa Constitution than followed under the federal caselaw. As a result, with respect to the Iowa constitutional claim, we apply the prevailing federal standard, but - reserve the right to apply that standard in a different fashion from the federal caselaw. See State v. Becker, 818 N.W.2d 135, 150 (Iowa 2012).
Law enforcement officers are required to give Miranda warnings when a suspect is in custody and subjected to interrogation. State v. Tyler, 867 N.W.2d 136, 171 (Iowa 2015) (discussing the warnings police must give based on Miranda v. Arizona, 384 U.S. 436, 471, 478-79, 86 S.Ct. 1602, 1626, 1630, 16 L.Ed.2d 694, 722, 726 (1966)); “[Clustody must be determined based on how a reasonable person in the suspect’s situation would perceive [the] circumstances.” Yarborough v. Alvarado, 541 U.S. 652, 662, 124 S.Ct. 2140, 2148, 158 L.Ed.2d 938, 950 (2004). Custody occurs “upon formal arrest or under any other circumstances where the suspect is deprived of his or her freedom of action in any significant way.” State v. Ortiz, 766 N.W.2d 244, 251 (Iowa 2009). This standard seeks to apply the Miranda requirements to coercive atmospheres, not just coercive places. It uses a case-by-case evaluation of all the circumstances existing at the time of the interrogation. The factors used to determine custody include . . - . '
(1) the language used to summon the individual; (2) the purpose, place, and manner of interrogation; (3) the extent to which the defendant is confronted with evidence of her guilt; and, (4) whether the defendant is free to leave the place of questioning.
State v. Countryman, 572 N.W.2d 553, 558 (Iowa 1997).
In a Miranda claim, interrogation consists of the express questioning and words- and actions béyond those normally part of arrest and custody “that the police should know are reasonably likely to elicit an incriminating response from the suspect.” State v. Miranda, 672 N.W.2d 753, 761 (Iowa 2003) (quoting State v. Peterson, 663 N.W.2d 417, 424 (Iowa 2003)). The State has not separately addressed whether an- interrogation occurred and so has *396waived any argument to the contrary. See id. Therefore, if we determine Schlitter was in custody then the officers would have been required to inform him of his Miranda rights.
We first consider the circumstances concerning how the individual was summoned to the interrogation. • Countryman, 572 N.W.2d at 558. An officer called Schlitter and asked if he would be willing to come to his office at the state patrol office on a later date to answer some more questions. The officer did not physically approach Schlitter, bring him to the station in a police vehicle, or otherwise force Schlitter to the interview but rather made a request and arrangements for Schlitter to come in another day. Cf. State v. Bogan, 774 N.W.2d 676, 680-81 (Iowa 2009) (finding custody when principal pulled student out of class and walked him to the office followed by officers, and the student did not volunteer or acquiesce to speaking with police). There is no indication Schlitter attempted to decline the request or showed any reluctance to attend the interview.
We next consider the purpose, place, and manner of interrogation. Countryman, 572 N.W.2d at 558. With respect to the manner of questioning, we consider how long it lasted, “the number of persons conducting the questioning, the number of breaks taken during the questioning, the availability of restroom breaks or other breaks, and the type of questioning in which those conducting the interview engage.” Tyler, 867 N.W.2d at 172-73. In Tyler, we noted that even a three-hour interview was not necessarily .custodial. Id. at 172. On the other hand, even brief interviews that the individual knows will continue until the desired answer is given can be custodial. Miranda, 672 N.W.2d at 760. Interrogation at a police station is generally a more coercive environment than questioning a suspect away from the station, but “merely because questioning takes place at the police station” does not necessarily implicate custody. State v. Smith, 546 N.W.2d 916, 922 (Iowa 1996).
The purpose of the encounter in this case was to get Schlitter to confess to being the perpetrator of the physical injuries suffered by K.S. or to get him to implicate Parmer. Schlitter was a focus of the' interrogation, but so was another person. Schlitter suspected he and Parmer were targets of the investigation, as did others, including family members. Schlit-ter had talked with law enforcement investigators on several occasions prior to the March 30 encounter and had consistently denied any responsibility for the injuries inflicted on K.S. The questioning took place in an interview room and lasted about one hour and twenty minutes. The officers did not call Schlitter to the patrol office with the intent to detain or arrest him, nor did they indicate any such intent to Schlitter.
•During the interview, Schlitter sat in a chair against a wall between a desk and a table. A camera was located directly in front of him. Two officers were in the room during the interview, but only one asked Schlitter questions. The officers wore plain clothes. One officer was behind a desk, and the interviewing officer sat in a chair by the table facing Schlitter. This arrangement placed the officer between Schlitter and the door. He inched closer to Schlitter throughout the questioning, moving from around two feet away to nearly knee-to-knee, then moving back by the table.
It is clear the officers applied forceful verbal pressure on Schlitter as the questioning progressed. The pressure included a strong and graphic description of the injuries inflicted on K.S. The officer implied Schlitter inflicted the injury and con*397fronted Schlitter with the inconsistency between his denial of any responsibility and his declaration that Parmer was a good mother and never violent. The type and amount of pressure used by the officers tended to make the atmosphere coercive. The pressure was not just for Schlitter to implicate Parmer but also for him to confess in the alternative. Schlitter thought the aggressive pressure was unfair and asked the officer several times to stop.
The officers also asked Schlitter if he would consent to a polygraph examination; Schlitter said he would consent to a polygraph test but wanted to take it the following day because' it was getting close to dinnertime. After the officers pressed for Schlitter to immediately take the test, he requested to talk to his lawyer. When Schlitter was unable to reach his lawyer by phone, the officers again pressed for him to take the test, but then agreed it could be done the following day. Schlitter told the officers that he would come back the next day, stating, “[I]f that’s when you want me here.”
The third factor looks at “the extent to which the defendant is confronted with evidence of [his] guilt.” Countryman, 572 N.W.2d at 558. During this interview, the officer described the actions that could have caused K.S.’s injuries, such as striking her head, shaking her violently, or dropping her. The officer continued- to describe each of those scenarios in more detail. He told Schlitter that abused children cling to their abuser and do not run away from the one abusing them. The officer then continued asking if Schlitter somehow hit KS.’s head on-anything while carrying her or lifting her. He implied that Schlitter picked K.S. up too fast and squeezed her hard enough to cause the bruising without realizing the hold was too rough or accidentally squeezed her out of frustration. The officer asked to trace Schlitter’s hand, suggesting it could help identify the source of bruising to KS.’s face. -The officer told Schlitter his explanations were not credible and pointed out the bruising on K.S. only began after he became the custodial parent.
The amount of evidence of Schlitter’s guilt as the perpetrator presented to him during the interview was not significant. Schlitter did not make a confession, nor did the officer present any evidence to him showing Schlitter was directly responsible for KS.’s -injuries. Although the atmosphere became highly accusatory at a point, the evidence presented to Schlitter was circumstantial and speculative in nature.
The final factor considered to establish custody is whether the individual was free to leave the place of questioning. Id; One element of this is the degree of physical restriction placed on the individual. Smith, 546 N.W.2d at 925. Schlitter’s path to the exit was- partially blocked by the interviewing' officer. Additionally, the officers did not open the interview by telling Schlitter he was free to leave when he wanted. However, when the officers left the room, Schlitter had free access to the door. He was not handcuffed at any point during the interview, and the door to the room was not locked. He drove himself to' the station and was not dependent on the officers to drive him home. See Tyler, 867 N.W.2d at 174 (finding no custody even when the individual had been brought by officers to the police station when the individual was told she was,free to leave and that she would be given a ride). Although Schlitter became upset, during the interview, at no time did his demeanor indicate he felt he would not be allowed to leave. In fact,- when Schlitter told the officers towards the end of the interview that he could not remain long enough- to take a polygraph examination because he needed *398to leave for dinner with his family, they attempted to talk him into staying for the test, but allowed him to leave without doing so. See Countryman, 572 N.W,2d at 558 (finding no custody when individual was not restrained, never asked to leave, and officer testified, he would have tried to talk her out of leaving but would have allowed it). Importantly, this exchange indicated Schlitter did not consider himself to be in custody, but free to leave to have dinner with his family.
Considering the totality of the circumstances, we conclude Schlitter was not in custody at the time he entered the interrogation room of the patrol office. He cooperatively talked to officers in the days preceding the interview and, under the circumstances, would not have been alarmed to learn they wanted to talk to him again. He voluntarily went to the patrol office. The request to meet at the patrol office and to go into the interview room could not be viewed reasonably as a significant restraint on Schlitter’.s freedom of movement. The difficult question is whether the circumstances that followed deprived Schlitter at any point of his freedom to a degrqe similar to a formal arrest. See Miranda, 672 N.W.2d at 759 (noting Miranda safeguards apply as so.on as.the person is deprived of freedom to the level of a formal arrest). A coercive environment, whether by formal arrest or otherwise, gives rise to custody, which requires the protections of Miranda, See id,
The strength of Schlitter’s claim of custody is found in the aggressive and accusatory nature of the questioning. The approach taken by the investigating officers was consistent with the type of circumstances that can make suspects feel a coercive atmosphere of custody. The more an interrogating officer discloses evidence of guilt to a suspect and the more force the officer uses to express guilt to a suspect, the greater likelihood the suspect will be in custody for purposes of Miranda. Cf. Tyler, 867 N.W.2d at 173-74 (distinguishing between accusatory .and truth-seeking questioning); Smith, 546 N.W,2d at 925 (noting questions about conflicting stories was to find information rather than to confront the defendant with evidence of guilt); see also United States v. Griffin, 922 F.2d 1343, 1348 (8th Cir. 1990) (“[T]he fact that the individual has become the. focus of the investigation is relevant ‘to the extent that the suspect is aware of the evidence against him’ and this awareness contributes to. the suspect’s sense of custody.” (quoting United States v. Carter, 884 F.2d 368, 370 (8th Cir. 1989))). Yet, Schlitter understood the officers were asking him either to acknowledge his guilt or implicate Parmer. Even during the aggressive questioning, Schlit-ter understood the officers were looking at one or the other as the guilty party. Thus, if the officers wanted Schlitter to implicate Parmer, a necessary inference would be the officers lacked evidence of his guilt. Likewise, the request to trace his hand and to take a polygraph examination did not support custody under the circumstances, but' confirmed the ongoing nature of the investigation and the ongoing search for more evidence. Even though the officers wanted to press on with the questioning and with the polygraph test' when Sehlitter wanted to end the encounter, the questioning did promptly end, and Schlit-ter did agree to return the next day. Schlitter indicated he did not believe the. interview had evolved into a custodial setting by telling the officers' near the end of the interview he would meed to take the requested polygraph examination another time because he needed to be leaving for dinner. Under all the circumstances, and balancing all four factors, we conclude, as did the district court, the interrogation did not restrict Schlitter’s freedom to the point *399that it rendered him in custody for purposes of Miranda.
IV. Conclusion.
We conclude the district court did not err in denying the motion to suppress. However, we conclude insufficient evidence was presented at trial to support a conviction for child endangerment under the theory that Schlitter used unreasonable force that resulted in bodily injuries to K.S. As a result, trial counsel for Schlitter was ineffective for failing to preserve error. We therefore reverse and remand for a new trial. In light of the need for a new trial, it is unnecessary to address further the other issues raised by Schlitter on appeal. We allow the decision of the court of appeals to stand as a final decision on the claim of ineffective assistance of counsel relating to the failure to investigate.
DECISION OF COURT OF APPEALS AFFIRMED IN PART AND VACATED IN PART; DISTRICT; COURT JUDGMENT AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.
All justices concur except WIGGINS, J., who concurs in part and dissents .in part, and APPEL, J., who files a separate opinion concurring in part and dissenting in part in which WIGGINS and HECHT, JJ., join.