In this appeal, the defendant asks us to overturn State v. Comried, which interpreted Iowa Code section 321J.2(1)(e) (2001) (operating while intoxicated (OWI) statute) to ban driving a motor vehicle with any detectible amount of a prohibited drug in one’s body, regardless of whether the ability to drive was impaired. 693 N.W.2d 773, 778 (Iowa 2005). This defendant was stopped for driving over the cen-terline and admitted to smoking half of a joint and being under the influence of marijuana. A drug screen detected a nonim-pairing metabolite of marijuana in his urine. He filed a motion to dismiss the OWI charge, arguing Comried is no longer good law because it relied on an Arizona decision and that state’s supreme court later held an OWI conviction cannot be based solely on the presence of a nonim-pairing metabolite. The district court disagreed, denied his motion to dismiss, and convicted him of violating section 321J.2. The court of appeals affirmed his conviction based on Comried, noting it “will not diverge from supreme court precedent.” We granted the defendant’s application for further review.

For the reasons explained below, we reaffirm Comried based on the plain mean*179ing of the statutory text. The traffic stop and request for a urinalysis were lawful based on the defendant’s erratic driving and his admitted recent drug use and impairment. The defendant raises no constitutional challenge to the statute’s breadth, which permits a conviction based solely on the presence of a nonimpairing metabolite of. marijuana in the driver’s urine. Policy arguments that the statute is too harsh should be directed to the legislature..

I. Background Facts and Proceedings.

At 9:41. p.m. on June 20, 2014, Floyd County Deputy Sheriff Chad Weber was dispatched to Rockford City Park to respond to a report of narcotics activity involving a silver Hyundai Sonata. Upon arriving, he was approached by a man who reported smelling marijuana coming from a silver car and someone with dreadlocks driving off in that vehicle. Deputy Weber spotted a man with dreadlocks on foot and a silver Sonata backing out of a parking spot. Deputy Weber followed the silver Sonata. A check of the license plate number revealed the car’s registration was expired. He observed both left-side tires of the car crossing the centerline. Deputy Weber pulled the car over and identified the driver as Erik Childs. Deputy Weber’s report describes their encounter: ‘

I approached the vehicle and told the driver he was being stopped for crossing the center line and expired registration. I asked the driver where he had been tonight and he stated he was at the park playing basketball with his son. I then •told him that I had received a complaint of persons smoking marijuana in. that area in.a vehicle matching the description of this ■ vehicle. I then asked the driver if he was under the influence of drugs or alcohol. He said yes, in which I asked what substance and he said marijuana. I asked how much and he said half- a joint, I .asked how big the joint was and he held up his fingers showing me how big. . ■ ..

Deputy Weber also observed that when Childs “began to walk towards the back of the car [he] had his left hand on the vehicle to keep his balance.” Childs performed poorly on several field tests for sobriety, missing heel-to-toe steps and counting the number thirteen twice. At the police station, Childs consented to. a urine test, which revealed the presence of sixty-two nanograms per milliliter of a nonimpairing metabolite of marijuana, ll-nor-9-carboxy-delta-tetrahydrocannabinol (Carboxy-THC).1

Childs was charged with operating while intoxicated, first offense, in violation of Iowa Code section 321J.2(1)(a) (2014) (operating while under the influence of drugs) and (c) (operating a motor vehicle while “any amount of a controlled substance is present in the .-.. person’s blood or urine”). Childs filed a motion to dismiss, arguing he could not be convicted under section 321J.2 based solely on the presence of a nonimpairing metabolite of marijuana in his urine. Childs urged the court to overrule Comried, which interpreted section 321J.2(1)(c) (2001) to prohibit driving with “any amount” of a prohibited drug, that is, “any amount greater than zero.” 693 N.W.2d at 778. Comried was a statuto-*180^-interpretation case that relied on an Arizona decision addressing the same issue under the Arizona DUI statute. See id. at 775—76; see also State v. Phillips, 178 Ariz. 368, 873 P.2d 706, 708 (App. 1994). However, a later Arizona decision held “drivers cannot be convicted of [DUI] based merely on the presence of a non-impairing metabolite that may reflect the prior usage of marijuana.” State ex rel. Montgomery v. Harris, 234 Ariz. 343, 322 P.3d 160, 164 (2014). Childs argued that Phillips was no longer good law in Arizona, and accordingly, Comried should be overruled. Childs’s written motion asked for the statute to be reinterpreted to omit nonimpairing metabolites. At the hearing on the motion to dismiss, Childs echoed this argument:

We are asking for the case to be dismissed. When the Defendant was tested after he was pulled over and sobriety testing, he was found positive for a non-impairing metabolite of marijuana. Many states have already ruled this non-impairing metabolite is not a DUI; that only the impairing metabolite is.

[[Image here]]

[Phillips ] is the case that we actually based our OWI or marijuana law on, we used that case, and it’s cited throughout the case that decided that any amount of a controlled substance is an OWI in Iowa. They actually have distinguished that case, stating that ■ now it is the only—Only the impairing metabolite that is a DUI in [Harris ]. And based on the changes of law and based upon the fact that my client was not positive for the impairing metabolite, we are asking for the case to be dismissed.

The district court rejected this argument, stating,

Mr. Childs, again, your attorney is asking the Court to find that the law itself is unconstitutional; that there is no rational basis for the law here in Iowa.

I think that that’s a very, very high standard. I mean, to say that something is unconstitutional means that there is no—no reason at all to have this law in place, basically. And again, I think it’s an argument that I’m not going to agree with, but it’s something that could be appealed and maybe the Supreme Court or the Court of Appeals may find that they want to overturn this law and say that it’s not constitutional, but I’m not willing to do that.

I think that there is a rational basis to just say any marijuana in your system, whether it impairs you or not, that’s enough to say people shouldn’t be driving with that in their system.

Again, I understand the rationale of what your attorney is saying is that there should be some test as to whether or not it made you a bad driver, but Iowa hasn’t decided that that’s necessary. So, until someone tells me—someone else above me tells me it’s not constitutional, I’m going to And that it is.

So, I’m going to deny the Defendant’s Motion to Dismiss.

The district court filed a written order denying the motion to dismiss. Childs filed a motion to suppress, contending Deputy Weber lacked probable cause or reasonable suspicion for the traffic stop. The district court denied his motion, concluding the expired registration and driving over the centerline provided sufficient grounds. Childs ultimately was convicted on the minutes of testimony of operating while intoxicated, first offense, in violation of Iowa Code section 321 J.2.2

*181Childs appealed, and we transferred the case to the court of appeals. Childs’s appellate briefs raise no constitutional challenge to section B21J.2. Rather, Childs makes the same statutory-interpretation argument on appeal as he did in district court—Comried should be overruled and the statute reinterpreted to omit nonim-pairing metabolites. The court of appeals rejected his arguments and affirmed his conviction. We granted Childs’s application for further review.

II. Standard of Review.

“On further review, we can review any or all of the issues raised on appeal or limit our review to just those issues brought to our attention by the application for further review.” Papillon v. Jones, 892 N.W.2d 763, 769 (Iowa 2017) (quoting Woods v. Young, 732 N.W.2d 39, 40 (Iowa 2007)). We elect to confíne our review to Childs’s statutory-interpretation claim. The court of appeals decision affirming the denial of his motion to suppress shall stand as the final decision on that claim.

‘We review rulings on questions of statutory interpretation for correction of errors at law.” State v. Iowa Dist. Ct., 889 N.W.2d 467, 470 (Iowa 2017) (quoting State v. Olutunde, 878 N.W.2d 264, 266 (Iowa 2016)). “Similarly, we review a ruling on a motion to dismiss for correction of errors at law.” Ney v. Ney, 891 N.W.2d 446, 450 (Iowa 2017).

III. Analysis.

A. Preservation of Error. In district court and on appeal, Childs makes the same statutory-interpretation argument: that we should overrule Comried and hold section 321J.2 is not violated by the presence of nonimpairing metabolites of marijuana in a driver’s urine. The district court described Childs’s argument as a constitutional challenge in the colloquy at the hearing on the motion to dismiss and rejected it. Childs makes no constitutional claim on appeal. The State’s appellate briefing acknowledges that Childs preserved error on his statutory challenge. We agree.

We do not construe the district court’s discussion of the constitutionality of the statute to mean the court overlooked Childs’s statutory-interpretation argument that the statute did not apply to driving with a nonimpairing metabolite. To the contrary, the district court necessarily rejected Childs’s statutory-interpretation argument when it orally ruled the statute constitutionally applied to him, denied his motion to dismiss, and later found him guilty of violating section 321J.2. The court of appeals reached the same conclusion, stating “the district court did not err in interpreting section 321J.2 to include marijuana metabolites and in denying the motion to dismiss.” See EnviroGas, L.P. v. Cedar Rapids/Linn Cty. Solid Waste Agency, 641 N.W.2d 776, 782 (Iowa 2002) (holding error was preserved on both prongs of challenge to applicability of statute, presuming district court resolved both, even though it only discussed one); Meier v. Senecaut, 641 N.W.2d 532, 539-40 (Iowa 2002) (discussing appellate principle that “we assume the district court rejected each defense to a claim on its merits, even though the district court did not address each defense in its ruling”); see also City of Riverdale v. Diercks, 806 N.W.2d 643, 655 (Iowa 2011) (concluding district court, by awarding attorney fees, must have rejected city’s good-faith defense to the fee award); cf. State v. Hellstern, 856 N.W.2d 355, 360 (Iowa 2014) (“We are to decide the statutory issue first in order to avoid unnecessary adjudication of constitutional claims.”).

State v. Mitchell does not support the conclusion that Childs waived his statutory *182argument. 757 N.W.2d 431, 435 (Iowa 2008). Holly Mitchell was charged- with child endangerment because slie and her children lived with a registered sex offender. Id. at 434. Mitchell filed a motion to dismiss, raising two constitutional challenges (due process and equal protection) to the child endangerment statute. Id. at 435. The district court denied her motion by addressing only the equal protection claim. Id. We held Mitchell failed to preserve the due process claim for appellate review because she did not seek a ruling on that claim in district court before filing her appeal. Id. Mitchell is distinguishable; The district court in Mitchell could decide one constitutional claim without deciding the other. See id. The district court’s ruling rejecting the equal protection challenge was not implicitly dispositive of the due process claim. No statutory claim had been made. See id. at 434. By contrast, here, the district court could not uphold the constitutionality of the OWI statute as applied to Childs without necessarily interpreting the statute to apply to Childs.

The fact the State agrees Childs preserved error is another reason to conclude his statutory-interpretation claim is preserved for appellate review. See State v. Coleman, 890 N.W.2d 284, 286-87 (Iowa 2017) (relying on State’s concession that defendant preserved error). In its appellate briefing, the State recognized that Childs challenged the district court’s interpretation of Iowa Code section 321J.2 and that his “motion to dismiss and the district court’s ruling thereon preserved this issue for appellate review.” To hold otherwise would conflict with the lenient approach to error preservation in Coleman, which held the defendant preserved an argument under the Iowa Constitution for appellate review without mentioning the Iowa Constitution in district court. See id. at 286. Unlike the defendant in Coleman, Childs in fact made the same argument in district court in his motion to dismiss that he makes on appeal—an argument the district court ruled on by denying his motion and convicting him.

B. Statutory Interpretation—Com-ried Reaffirmed. We must decide whether to overrule Comried, which we decided twelve years ago. The district court and court of appeals correctly applied Comried, and Childs concedes that his conviction must be upheld if that case remains good law. We reaffirm Comried based on its reasoning, which applies the plain meaning of the operative statutory language.

The legislature recently amended the narcotics laws to allqw limited medical use of cannabis oil derived from marijuana, but chose to leave intact Iowa Code section 321J.2(1)(c).3 Childs does not claim he had a valid prescription for medicinal marijuana. See Bearinger v. Iowa Dep’t of Transp., 844 N.W.2d 104, 107-08 (Iowa 2014) (discussing prescription drug defense), Nor does Childs claim he only had the metabolite in his urine from prior drug use days earlier, such that he was not driving under the influence. To the contrary, he exhibited signs of current impairment and admitted to smoking marijuana and driving under its influence. He does not argue on- appeal that the statute as interpreted in Comried is unconstitutional.

*183Iowa Code section 321 J.2 provides that a person commits the offense of operating while intoxicated if the person “operates a motor vehicle in this state in any of the following conditions:” .

(а) While under the influence of an alcoholic beverage or other drug or a combination of such substances.

(б) While having an alcohol concentration of .08 or more.

(c) While any amount of a controlled substance is present in the person, as measured in the person’s blood or urine.

Id. § 321J.2(1)(a)-(c) (emphasis added). “Controlled substance,” ,in turn, is defined to include “any metabolite or derivative of the drug, -substance, or compound” listed in section 124.204. Id. § 321J.1(4) (emphasis added). Section 124.204 lists “[tjetrahy-drocannabinols ... meaning tetrahydro-cannabinol naturally contained in. a plant of the genus Cannabis” as a schedule I substance. Id. § 124.204(4) (u). Carboxy-THC is a metabolite of the tetrahydrocan-nabinol (THC) found in marijuana, a controlled substance.4 Carboxy-THC is found in the mine of a person who has smoked or ingested marijuana. See Darron J. Hubbard, Comment, Narcotics on Illinois’s Roadways: Drugged Driving’s III Effects After Martin, 62 DePaul L. .Rev. 591, 605-07 (2013) (reviewing the process by which body converts THC into Carboxy-THC). Therefore, Carboxy-THC falls within the definition of a prohibited “controlled substance” under Iowa Code section 321 J.l.

In Comried, we interpreted the text of section 321J.2(1)(c) to prohibit driving with “any amount” of a controlled substance detectable in one’s body. 693 N.W.2d at 778. We observed that the legislature amended section 321J.2 in 1998 to create a per se ban on driving with any amount of a controlled substance in the body, “whether or not they are under- the influence.” Id. at 776; see also Bearinger, 844 N.W.2d at 107 (interpreting Comried and noting- section 321J.2 creates a per se ban “regardless of whether a person is ‘under the influence’ of that controlled substance” (quoting Comried, 693 N.W.2d at 776)). We noted the purpose of chapter' 321J is “to reduce the holocaust on’our highways.” Comried, 693 N.W.2d at 775 (quoting State v. Kelly, 430 N.W.2d 427, 429 (Iowa 1988)). Relying on Arizona and Indiana precedent, we stated,

The legislature could reasonably have imposed such a ban because the effects of drugs, as contrasted to the effects of alcohol, can vary greatly among those who use them. One court has observed that,

since the manufacture and distribution of illicit drugs are unregulated and because the drugs’ potency varies, the effects are unpredictable. Therefore, ... there is no'level of use above which people can be presumed impaired or below which they can be presumed unimpaired. '

Id. at 776 (alteration in original) (quoting Phillips, 873 P.2d at 708). We also relied on Iowa precedent:

Our court of appeals has reached a similar conclusion in a license-revocation case based on driving with controlled substances in the body. The court, noting the difficulty in relating the. amount of drugs in the body to driving impairment, said:

*184Unlike the blood alcohol concentration test used to measure alcohol impairment there is no similar test to measure marijuana impairment. There is, though, as was used here, a test to measure the use of marijuana, a drug illegal in the State of Iowa, in a person’s body. There being no reliable indicator of impairment, the legislature could rationally decide that the public is best protected by prohibiting one from driving who has a measurable amount of marijuana metabolites.

Id. (quoting Loder v. Iowa Dep’t of Transp., 622 N.W.2d 513, 516 (Iowa Ct. App. 2000)).

Childs argues we should overrule Com-ried because one of the several decisions we relied on, Phillips, was subsequently narrowed by the Arizona Supreme Court in Harris. Harris, 322 P.3d at 164. Harris interpreted a subsection of that state’s OWI law to prohibit only substances that impair driving. Id. The next year, the Arizona Supreme Court clarified that the Arizona statute “casts a net that embraces drivers who have proscribed drugs or their impairing metabolites in their bodies but who may or may not be impaired,” while allowing a limited defense to patients certified for medicinal marijuana use who can prove they were not impaired. Dobson v. McClennen, 238 Ariz. 389, 361 P.3d 374, 377 (2015).

The Iowa legislature chose to cast a wider net, criminalizing driving with any amount of prohibited substances in one’s body, including the nonimpairing metabolite at issue commonly found in urine after marijuana use. The reasoning of Comried remains persuasive, as the operative text of the statute has not changed. See Iowa Code § 321J.2(c) (2014). Our court “may not ... enlarge or otherwise change the terms of a statute as the legislature adopted it.” State v. Iowa Dist. Ct., 730 N.W.2d 677, 679 (Iowa 2007) (alteration in original) (quoting State v. Miller, 590 N.W.2d 45, 47 (Iowa 1999)). “When a proposed interpretation of a statute would require the court to ‘read something into the law that is not apparent from the words chosen by the legislature,’ the court will reject it.” Id. (quoting State v. Guzman-Juarez, 591 N.W.2d 1, 2 (Iowa 1999)).

The premise for that legislative choice was the absence of reliable testing to determine whether a particular level of a narcotic impairs driving. That premise remains true today.

Unfortunately, there is no procedure comparable to the Standard Field Sobriety Test that a police officer can administer on a roadside to determine if a driver is under the influence of drugs. For example, marijuana diminishes a person’s temporal and spatial judgment, but the Standard Field Sobriety Test does not measure those effects. Police officers also rely on nystagmus to determine if a person is under the influence of alcohol, but drugs that dilate or constrict the pupils do not also cause nystagmus. There also is no device comparable to a breathalyzer to identify marijuana intoxication or the presence and amount of THC, the psychoactive ingredient in marijuana, in a driver’s blood. What is worse, even if that measurement could be done, there is no medical or scientific consensus regarding the amount of THC that would impair the average driver. That is true for a host of reasons, most of which stem from the fact that the relevant pharmaceutics are far more complicated for drugs than for alcohol.

Paul J. Larkin Jr., Medical or Recreational Marijuana and Drugged Driving, 52 Am. Crim. L. Rev. 453, 483 (2015) (emphasis added) (footnotes omitted). As the dissent in Harris recognized, “[T]he difficulty *185of detecting drug impairment justifies a flat ban.” 322 P.3d at 165 (Timmer, J., dissenting). “Hydroxy-THC [impairing] converts quickly to Carboxy-THC [nonim-pairing] .... [A] driver with Carboxy-THC in the blood at the time of testing may or may not have had Hydroxy-THC in the blood while driving.” Id. A “flat ban ensures that a driver who had an impairing substance in the body while driving is prosecuted even though that substance may have quickly metabolized into a non-impairing substance.” Id.

The harshness of Iowa’s flat ban is ameliorated by the fact that the motorist would be asked to submit to chemical testing only after the officer performed a lawful traffic stop and had reasonable grounds to believe the driver was impaired. See Iowa Code § 321J.6(1) (setting forth grounds for chemical testing). In this case, for example, Childs was driving over the centerline, had trouble with his balance upon exiting his car, performed poorly on field tests for sobriety, and admitted he was under the influence of marijuana after smoking half of a joint.

Childs does not argue we should rely on the absurd-results doctrine. We disagree with any claim that Comried’s interpretation of the Iowa OWI law produces an absurd result. We have cautioned that “the absurd results doctrine should be used sparingly because it entails the risk that the judiciary will displace legislative policy on the basis of speculation that the legislature could not have meant what it unmistakably said.” Sherwin-Williams Co. v. Iowa Dep’t of Revenue, 789 N.W.2d 417, 427 (Iowa 2010) (quoting 2A Norman J. Singer & Shambie Singer, Statutes and Statutory Construction § 45:12, at 105-07 (7th ed. 2007)); see also Bearinger, 844 N.W.2d at 110 n.3 (“The absurd-results doctrine should be used cautiously.”). We recently reiterated,

Establishing absurdity in an unambiguous statute is difficult for good reason. We have explained that “we will not ignore clear legislative language merely because it leads to a result that seems contrary to the court’s expectations.” The express language must produce a result that is “demonstrably at odds with the intention” of the legislature.

In re J.C., 857 N.W.2d 495, 503 (Iowa 2014) (citations omitted) (quoting Sherwin-Williams Co., 789 N.W.2d at 427, 429). It is not absurd for the legislature to enact a per se, or zero-tolerance, ban on driving with this marijuana metabolite in one’s body, given the absence of an available scientific test to determine what level of marijuana impairs driving.

Comried is not an outlier. Other states have interpreted equivalent OWI statutes to criminalize driving with any detectible amount of a prohibited drug, regardless of impairment. See Love v. State, 271 Ga. 398, 517 S.E.2d 53, 56, 57 (1999) (concluding that “a statute which makes it unlawful to drive while marijuana residue is circulating in the driver’s body fluids bears a rational relationship to ... protection of the public” but declaring law unconstitutional on equal protection grounds as prohibiting medicinal use); People v. Fate, 159 Ill.2d 267, 201 Ill.Dec. 117, 636 N.E.2d 549, 550, 551 (1994) (concluding a flat ban prohibiting “any amount of a controlled substance” was constitutional given that there was no standard for impairment); Bennett v. State, 801 N.E.2d 170, 176 (Ind. Ct. App. 2003) (“[A] flat ban on driving with any proscribed controlled substance in the body, whether or not capable of causing impairment, is permissible.”); Commonwealth v. Hutchins, 42 A.3d 302, 310 (Pa. Super. Ct. 2012) (“[A] conviction under [the OWI statute] does not require that a driver be impaired; rather, it prohibits the operation of a motor vehicle by any driver who has *186any amount of specifically enumerated controlled substance in his blood.” (quoting Commonwealth v. Etchison, 916 A.2d 1169, 1174 (Pa. Super. Ct. 2007))); State v. Smet, 288 Wis.2d 525, 709 N.W.2d 474, 479 (App. 2005) (concluding “proof of impairment is not necessary” under OWI statute).5

Only three states with per se bans, Arizona (as discussed above), Delaware, and Michigan, distinguish between active and inactive metabolites. Delaware does so because its OWI statute expressly states that it is illegal to drive with “any amount of a substance or compound that is the result of the unlawful use or consumption of an illicit or recreational drug” and, in turn, defines that term as “not inelud[ing] any substance or compound that' is' solely an inactive ingredient or inactive metabolite of such drug.” Del. Code Ann. tit. 21, § 4177(a)(6), (c)(9) (West, Westlaw current through 81 Laws 2017, chs. 1-20). Iowa’s OWI statute lacks such an-exclusion for nonimpairing metabolites. See Iowa Dist. Ct., 730 N.W.2d at 679 (“Statutory text may express legislative intent by omission as well as inclusion.”).

Michigan courts have struggled with the interpretation of that state's OWI law. The Michigan statute criminalizes driving a motor vehicle with “any amount of a controlled substance listed in schedule 1 under section 7212 of the public health code.” Mich. Comp. Laws Ann. § 257.625(8) (West, Westlaw current through P.A. 2017, No. 50 of 2017 Reg. Sess. of 99th Leg.). Marijuana is a controlled substance. Id. § 333.7212(1)(c). In People v. Derror, the Michigan Supreme Court examined whether the legislature intended Carboxy-THC, a nonimpairing metabolite of marijuana, to be considered a controlled substance included in the OWI statute, 475 Mich. 316, 715 N.W.2d 822, 825 (2006), overruled by People v. Feezel, 486 Mich. 184, 783 N.W.2d 67, 86 (2010). The,-court held that because Carboxy-THC is “a metabolite of THC in that it is produced when the body metabolizes THC,” it was properly considered a “derivative” of marijuana.6 Id. at 828. However, four years later, in Feezel, the court overruled Derror and concluded Carboxy-THC was not a derivative. 783 N.W.2d at 81, 86. The Feezel court noted the statutory definition was based on federal law and did not “contain the term ‘11-carboxy-THC’ ... [n]or do the statutes contain the term ‘metabolite.’” Id. at 83; see also 21 U.S.C. § 802(16) (2012). Three justices dissented in part, noting that the majority’s interpretation went against the *187plain language of Michigan’s statute. Id. at 87 (Young, J., concurring in part and dissenting in part).

In subsequent decisions, a justice noted that “[t]he trouble caused by the Feezel decision is worthy of this Court’s serious attention.” People v. Soares, 488 Mich. 918, 789 N.W.2d 854, 855 (2010) (Corrigan, J., dissenting); People v. Barkley, 488 Mich. 901, 789 N.W.2d 441, 442 (2010) (Corrigan, J., dissenting). The decision left law enforcement “in a legal limbo” because they could “arrest if we find marijuana on you, but it’s different if we find marijuana in you.” Soares, 789 N.W.2d at 855 (quoting Tom Greenwood, Ruling Clouds Pot Smoking, Driving Law, The Detroit News, July 29, 2010). Barkley illustrated the problem:

This case well illustrates the potential confusion wrought by the Feezel decision. Defendant, who was driving -with THC in her system, ran. a stop sign and collided with a pick-up truck that had the right of way at the intersection. Two passengers in defendant’s car—her six-year-old son and her adult friend—were killed. As a result, a jury convicted defendant of two counts of negligent homicide and one count of operating a motor vehicle and causing death while having a controlled chemical substance (marijuana) in her body, MCL 257.625(4) and. (8). Under Derror, defendant’s guilt of this last offense was clear. But Feezel. sir tempts to distinguish one metabolite of marijuana, 11-earboxy-THC, and prohibit it from being dubbed a controlled substance. Accordingly, the nature of defendant’s offense is now unclear. An expert testified' that defendant’s urine contained a sufficient amount of THC— at least 50 nanograms per milliliter—to test positive for the substance. But it is unclear from the record provided to this Court which metabolite or metabolites of THC were measured. All metabolites of THC indicate ingestion of marijuana, and defendant did not contest at trial which metabolite or metabolites appeared in her system.

789 N.W.2d at 442.

Unlike the Michigan statute, the Iowa legislature expressly added the words, “including tetrahydrocannabinols,” the psychoactive component of marijuana to the controlled-substances statute. Iowa Code § 124.101(19). Moreover, our OWI statute expressly criminalizes metabolites of that component in a way the Michigan statute did not. Compare Iowa Code § 321J.1(4), with Mich. Comp. Laws Ann. § 257.625. Accordingly, the Michigan cases do not support revisiting' Comried. We apply the Iowa statute as written and leave it to the legislature whether to revisit the zero-tolerance ban on driving with even nonim-pairing metabolites of marijuana.

IV. Disposition.

For these reasons, we affirm the decision of the court of appeals and affirm the district court’s judgment, conviction, and sentence.

DECISION OF COURT OF APPEALS AND DISTRICT COURT JUDGMENT AFFIRMED.

(concurring specially).'

I concur in the result. I would hold only that State v. Comried, 693 N.W.2d 773 (Iowa 2005), has decided the issue presented on appeal, and its rationale has not been undermined merely because case authority from another jurisdiction-we partially relied upon has , been overruled. *188Furthermore, a statute that criminalizes operating a vehicle while having the presence of a nonimpairing metabolite of marijuana in the blood system may seem to be based on a judgment that is wrong, even misplaced, but it is hardly absurd under the prevailing legal standard. While courts must not interpret ambiguous statutes in a way that will lead to an absurd result, see Iowa Ins. Inst. v. Core Grp. of Iowa Ass’n for Justice, 867 N.W.2d 58, 75 (Iowa 2015), an unambiguous statute is absurd only if its language produces “a result that is ‘demonstrably at odds with the intention’ of the legislature,” In re J.C., 857 N.W.2d 495, 503 (Iowa 2014) (quoting Sherwin-Williams Co. v. Iowa Dep’t of Revenue, 789 N.W.2d 417, 429 (Iowa 2010)). As we said in Comried, the legislature intended “to prohibit people from operating motor vehicles with controlled substances in their bodies, whether or not they are under the influence.” 693 N.W.2d at 776. The result of the statute here is not at odds with the legislature’s intent. Furthermore, no constitutional claim has been presented on appeal that requires us to address or even discuss whether the statute is rationally related to a legitimate government interest. On these limited grounds, I concur.

(dissenting).

I cannot join my colleagues in the majority because I believe error was not preserved on the question decided today. I also dissent from the majority’s expansive reading of State v. Comried, 693 N.W.2d 773 (Iowa 2005), because I believe it is flawed in several particulars. In detailing the reasons for my dissent, I will first examine the substance of Erik Childs’s position in the district court. Next, I will explain why the majority errs in concluding error was preserved on the question of statutory interpretation decided today, emphasizing the prudential reasons why further review should not have been granted in this case. Finally, I will detail my concerns with the majority’s reading of Com-ried.

I. The Motion to Dismiss.

Following a roadside stop, a sample of Childs’s urine tested positive for the presence of sixty-two nanograms per milliliter of an inactive secondary metabolite of tet-rahydrocannabinol (THC)—ll-nor-9-car-boxy-delta-THC (Carboxy-THC). Childs was charged with operating while intoxicated, first offense, in violation of Iowa Code section 321J.2(l)(c) (2014) (prohibiting the operation of a motor vehicle “[w]hile any amount of a controlled substance is present in the ... person’s blood or urine”).

Childs filed a motion to dismiss in which he took aim at our decision in Comried, 693 N.W.2d 773, which he read as “creating a [per] se ban on controlled substances while operating a motor vehicle.” The motion asserted Comried!a “[per]-se ban on operating a motor vehicle while under the influence of a controlled substance” is “questionable” because it “relied heavily upon” the opinion of the Arizona Court of Appeals in State v. Phillips, 178 Ariz. 368, 873 P.2d 706 (Ct. App. 1994), which was subsequently distinguished by the Arizona Supreme Court in State ex rel. Montgomery v. Harris, 234 Ariz. 343, 322 P.3d 160 (2014). The motion further asserted that

[d]ue to the fact that the Defendant only had the nonimpairing metabolite, Car-boxy-THC, in his system at the time of arrest, the case should be dismissed as he did not operate a motor vehicle under the influence at the time he was arrested.

At the hearing on the motion to dismiss, defense counsel reiterated the position that our decision in Comried was no longer *189good law because it relied on the Arizona court’s decision in Phillips. Noting that the Arizona Supreme Court had subsequently limited the vitality of the Phillips holding in Harris, defense counsel argued that Comried should not be viewed as authority in favor of the State’s position. Defense counsel neither addressed the plain meaning of the statutory text nor argued that any of our well-established rules of statutory interpretation should be applied when interpreting the text of section 321J.2(l)(c).

During the hearing colloquy, the district court revealed its misunderstanding that the motion to dismiss presented a constitutional challenge and explained its decision to deny the motion on rational-basis grounds:

Mr. Childs, again, your attorney is asking the Court to find that the law itself is unconstitutional; that there is no rational basis for the law here in Iowa.

I think that that’s a very, very high standard. I mean, to say that something is unconstitutional means that there is no—no reason at all to have this law in place, basically. And again, I think it’s an argument that I’m not going to agree with, but it’s something that could be appealed and maybe the Supreme Court or the Court of Appeals may find that they want to overturn this law and say that it’s not constitutional, but I’m not willing to do that.

I think that there is a rational basis to just say any marijuana in your system, whether it impairs you or not, that’s enough to say people shouldn’t be driving with that in their system.

Again, I understand the rationale of what your attorney is saying is that there should be some test as to whether or not it made you a bad driver, but Iowa hasn’t decided that that’s necessary. So, until someone tells me—someone else above me tells me it’s not constitutional, I’m going to find that it is.

So, I’m going to deny the Defendant’s Motion to Dismiss.

(Emphases added.) Defense counsel did not inform the court during the hearing that the court misunderstood the argument as a constitutional challenge; nor did counsel request a ruling on any statutory-interpretation issue.

The district court subsequently issued an order summarily denying the motion to dismiss. Childs did not file a posthearing motion requesting a ruling on any issue of statutory interpretation. He was convicted on the minutes of testimony of operating while intoxicated, first offense, in violation of Iowa Code section 321 J.2.7

II. Error Preservation.

The majority generously reads the defendant’s written .motion to dismiss as requesting a reinterpretation of section 321J.2(1)(c) to exclude the presence of inactive metabolites in one’s urine or blood as a basis for a conviction under the statute. The motion averred narrowly that Comried is no longer controlling authority in Iowa because the Arizona Supreme Court disavowed Phillips in 2014. See Harris, 322 P.3d at 160, 164 (interpreting statute prohibiting driving with “any drug ... or its metabolite in the person’s body” as requiring proof of driving “with any amount of THC or an impairing metabolite in the[] body” (quoting Ariz. Rev. Stat. § 28—1381(A)(3))). Aside from the fact that *190Comried interpreted part of section 321J.2(l)(c), there is no indication in the motion to dismiss that Childs was advancing any specific statutory-interpretation argument. The motion made no reference to the plain meaning of the statutory text; it advanced no argument that the statute is ambiguous; and it cited no rule of statutory interpretation supporting a determination that the presence of an inactive metabolite of marijuana in the defendant’s blood or urine can sustain a conviction under the statute.

The majority nonetheless concludes error was preserved on the proper interpretation of section 321J.2(1)(c) even though the district court ruled only on the constitutionality of the statute. The majority incorrectly reaches this conclusion by relying on the proposition that our court may assume, for purposes of appellate review, that the district court implicitly reached- a legal conclusion necessary to its ruling. We expressly rejected that proposition as a rule'of error preservation in Meier v. Senecaut, and I would do so again here. See 641 N.W.2d 532, 539-40 (Iowa 2002). Because the majority’s conclusion is wrong as a matter of law, and for prudential reasons arising from the state of the record, I dissent from the majority’s conclusion that error was presérved.

A. Rules of Error Preservation. “Error preservation is a fundamental principle of law with roots that extend to the basic constitutional function of appellate courts.” State v. Harrington, 893 N.W.2d 36, 42 (Iowa 2017). “Judges are not advocates who reach out to decide questions..,. ” Feld v. Borkowski, 790 N.W.2d 72, 83 (Iowa 2010) (Appel, J., concurring in part and dissenting in part). We do not consider issues for the first time on appeal and therefore only resolve issues preserved for appeal. State v. Coleman, 890 N.W.2d 284, 304 (Iowa 2017) (Waterman, J., dissenting).

Ordinarily, an issue is not preserved in a criminal case unless it has been both raised in and decided by the district court. State v. Manna, 534 N.W.2d 642, 644 (Iowa 1995); accord State v. Reilly, 104 Iowa 13, 14, 73 N.W. 356, 356 (1897). Similarly, an issue that is not asserted on appeal is generally waived.8 State v. Short, 851 N.W.2d 474, 479 (Iowa 2014). If an issue is raised but not decided in the district court, the issue is not preserved unless the party requests a ruling on the issue at a time when the court can take corrective action. State v. Krogmann, 804 N.W.2d 518, 524 (Iowa 2011); accord State v. Bricker, 135 Iowa 343, 345, 112 N.W. 645, 645 (1907). If the party fails to request the' ruling, error is not preserved. See State v. Schiernbeck, 203 N.W.2d 546, 547 (Iowa 1973).

These rules of error preservation promote sound judicial administration by promoting the prompt and orderly resolution of issues in a case and giving district courts an opportunity to fix mistakes prior to an appeal. See State v. Ambrose, 861 N.W.2d 550, 555 (Iowa 2015); see also Coleman, 890 N.W.2d at 304. The rules also protect parties from being surprised by issues on appeal, see Segura v. State, 889 N.W.2d 215, 219-20 (Iowa 2017), and from unprincipled judicial overreach, see Feld, 790 N.W.2d at 83-84 (discussing tension between judicial duties to decide concrete *191cases and to ensure coherent development of law); see also Coleman, 890 N.W.2d at 304-05 (discussing judicial overreach on appeal). The waiver doctrine, expressed in our rules of appellate procedure, similarly advances interests of judicial economy by limiting our review to issues actually argued on appeal. See Iowa R. App. P. 6.903(2)(g)(3) (stating, that the failure of appellant “to cite authority in support of an issue may be deemed waiver of that issue”).

Nonetheless, we recognize several exceptions to our rules of error preservation. For instance, we have recognized an exception to the requirement that issues must be decided by the district court for evidentiary rulings in certain instances. See, e.g., State v. Reyes, 744 N.W.2d 95, 99-100 (Iowa 2008) (considering statutory issue not decided by the district court but fully developed in supplemental briefing). But see, e.g., DeVoss v. State, 648 N.W.2d 56, 62-63 (Iowa 2002) (holding that general error preservation requirements do not prevent us from considering alternative grounds for the admission of evidence that was fully developed, but not decided, in the district court); State v. Howard, 509 N.W.2d 764, 769 (Iowa 1993) (“We conclude [the defendant] failed to preserve error on his hearsay claim.”). Further, “[w]hen there are parallel constitutional provisions in the Federal and State Constitutions and a party does not indicate the specific constitutional basis, we regard both federal and state constitutional claims as preserved,” even if the district court did not rule on both. State v. Gaskins, 866 N.W.2d 1, 6 (Iowa 2015) (quoting King v. State, 797 N.W.2d 565, 571 (Iowa 2011)); e.g. Coleman, 890 N.W.2d at 286-87 (mar jority opinion). We declined to recognize an exception where the district court concluded that a statute applied to a defendant but did not resolve an as-applied constitutional challenge to the statute as was urged in the district court and on appeal. See Adams v. City of Des Moines, 629 N.W.2d 367, 369 (Iowa 2001) (citing Ritz v. Wapello County Board of Supervisors, 595 N.W.2d 786, 789 (Iowa 1999), to emphasize that a motion to expand a ruling is “crucial to preservation of error” on a legal issue urged but not decided in the district court).

B. Application of Error Preservation Rules to This Case. Childs argued in the district court that our conclusion in Com-ried is “questionable” because it cited an Arizona Court of Appeals case that has since been distinguished by the Arizona Supreme Court. Comried involved a question of statutory interpretation. The district court, however, expressly denied Childs’s argument on constitutional grounds and did not address the defendant’s statutory' claim or cite Comried in its ruling on the motion to dismiss. On appeal, Childs reasserts that Comried was wrongly decided. Because Childs did not request a ruling on his statutory claim at a time when the district court could still take corrective action and because no constitutional claim was even asserted in the district court, I would conclude error was not preserved on the statutory claim. See Krogmann, 804 N.W.2d at 524; see also Bricker, 135 Iowa at 345, 112 N.W. at 645.

The factual scenario presented in this case does not fit any of the recognized exceptions to our rules of error preservation discussed above. The legal principles guiding our interpretation of Iowa Code section 321J.2(l)(c) were never developed in the record or- applied by the district court. Rather, the parties limited their arguments in the district court to the effect of changing Arizona caselaw on the con*192tinuing vitality of our decision in Comried.9 Even assuming for the sake of argument that an issue concerning the interpretation of section 321J.2(l)(c) was presented to the district court, the court’s ruling on that issue could not be reasonably understood as an evidentiary ruling.

The closest factual analogue to the error preservation issue presented in this case occurred in State v. Mitchell, 757 N.W.2d 431 (Iowa 2008). The defendant in that case pressed two theories of error on appeal. See id. at 434. The first theory was that a child endangerment statute violated the defendant’s due process rights. Id. The second theory posited that the child endangerment statute violated the defendant’s equal protection rights. Id. The district court only analyzed and decided one theory, the equal protection claim. Id. at 435. The defendant did not seek a ruling on the due process claim in the district court before filing her appeal. Id. We concluded the defendant failed to preserve error on her due process claim because the district court did not rule on it. Id.

As in Mitchell, we should conclude in this case that Childs failed to preserve error on his statutory-interpretation claim because he did not seek a ruling on it. Unlike the majority, I view the statutory-interpretation analysis as analytically distinct from the constitutional analysis actually undertaken by the district court.10 See Harris, 322 P.3d at 161 (rejecting argument that preliminary reading of statute in constitutional analysis bears on merits of statutory-interpretation claim); see also Adams, 629 N.W.2d at 369 (concluding error was not preserved on as-applied constitutional challenge to statute urged in district court and on appeal, even though district court determined statute applied to defendant); cf. King v. State, 818 N.W.2d 1, 42 (Iowa 2012) (Waterman, J., concurring specially) (arguing a court should not resolve textual issue by referring to source with no bearing on meaning of text at issue).

The majority concludes that the district court’s as-applied constitutional ruling necessarily decided an unspecified and undeveloped issue of statutory interpretation for purposes of error preservation. This conclusion is based on the faulty premise that we may assume for purposes of appellate review that the district court implicitly reached a legal conclusion necessary to its ruling. The majority cites Meier, 641 N.W.2d at 539-40, in support of this proposition. In fact, Meier expressly stands against it. Id. (rejecting argument that error is preserved on a legal issue necessary to district court’s legal conclusion but not addressed by the district court).

In Meier, we concluded error was not preserved on an issue of law raised in a motion to dismiss but not decided by the district court. Id. at 540-41. We rejected the argument that a district court implicitly rejects legal claims necessary to sustain its judgment for purposes of our rules of error preservation. See id. at 539-40 C‘[T]his assumption that the district court rejected claims not specifically addressed is not a rule of error preservation.... It is tied to our long-standing presumption that a district court found facts essential to sustain the judgment, and ... is not utilized as a means to preserve error.... ” (Citations omitted.)). We emphasized that *193our long-standing rule that we assume a district court decided facts necessary to support its decision “is not a rule of error preservation, but a rule governing our scope of review when an issue is raised and decided by the district court and the record or ruling on appeal contains incomplete findings or conclusions.” Id. at 539 (emphasis added). Because the rule only applies when an issue has been “raised and decided by the district court,” id., we held that it “is not a replacement for the requirement to preserve error and cannot be used in this casé to satisfy the preservation of error requirement that an issue on review be first decided by the district court,” id. at 540 (emphasis added).

The majority misplaces reliance on other cases in support of its conclusion on error preservation. EnviroGas L.P. v. Cedar Rapids/Linn County Solid Waste Agency, 641 N.W.2d 776, 782 (Iowa 2002), is just an iteration of the long-standing scope-of-review rule that Meier determined “cannot be used” to subvert the requirement that an issue of law is not preserved unless the district court rules on it. See Meier, 641 N.W.2d at 540. Although the majority cites EnviroGas as supporting the proposition that we may assume the district court implicitly reached a legal conclusion necessary to its ruling, the case stands only for the proposition that we may “presume [a] factual matter was resolved so as to support the court’s ultimate ruling.”11 See 641 N.W.2d at 782 (emphasis added); accord Bankers Trust Co. v. Fidata Trust Co., 452 N.W.2d 411, 413 (Iowa 1990) (“We, therefore, presume the court decided facts necessary to support its decision....” (Emphasis added.)). Similarly, in City of Riverdale v. Diercks, we assumed the district court found facts necessary to support an attorney-fee award under our state Freedom of Information Act. 806 N.W.2d 643, 655 (Iowa 2011). The majority uses the rule in this case to conclude the district court resolved an issue of law—the interpretation of a statute—not a question of fact.

In addition, the majority’s reliance on the doctrine of constitutional avoidance—in a case in which no constitutional argument was actually raised—is widely off base. The doctrine of constitutional avoidance is not a rule of error preservation and generally only applies when both statutory and constitutional questions are raised. See State v. Hellstern, 856 N.W.2d 355, 360 (Iowa 2014). Nor is it a rule of law that must be uniformly applied to every case— it is a prudential consideration of judicial restraint applied in many cases, but not all. See, e.g., State v. Storm, 898 N.W.2d 140, 145 n.1 (Iowa 2017) (declining to consider statutory question that could obviate need to reach constitutional question). The majority cites no authority for the proposition that the principle of constitutional avoidance has any bearing in a case in which the district court clearly misunderstood a statutory claim to be a constitutional claim. I reject as unsupported by law or fact the notion that the doctrine of constitutional avoidance provides support for the conclusion that the district court implicitly resolved a statutory-interpretátion claim when it resolved a constitutional claim that was neither raised nor briefed in the district court.

Finally, the majority errs in concluding that the district court’s constitutional anal*194ysis necessarily disposed. of the defendant’s statutory-interpretation claim on the merits. An as-applied constitutional analysis does not resolve issues of statutory interpretation. Harris, 322 P.3d at 161 (rejecting argument that reading of statute in as-applied constitutional analysis bears on merits of question of statutory interpretation).

C. Prudential Considerations. I further conclude that several prudential considerations should deter us from adopting and applying a new rule of error preservation to reach and decide the statutory-interpretation issue- on this record. First, neither the district court nor the parties cited a single canon of statutory interpretation or any exception to the plain-language rule. See Storm, 898 N.W.2d at 163, 165 (Hecht, J., dissenting) (outlining nonexhaustive list of exceptions to plain-language rule). Second, the parties did not raise, brief, or argue any constitutional theories, and the district court failed .to specify what constitutional question it believed it was deciding. Third, in reaching the merits, of an issue of law not decided by the district court, the majority violates our rules of error preservation and in so doing, risks “rewarding] trial counsel’s silence and gives all, defense counsel a perverse incentive to lay in the weeds in district court .., [and] deprives the district court of the opportunity to rule.” See Coleman, 890 N.W.2d at 304 (Waterman, J., dissenting). Finally, the new rule of error preservation applied by the majority raises implications that I am not sure the court is prepared to countenance. If Childs preserved error on the statutory-interpretation issue in this case, will the rule of the case be that a party preserves error on a' question of statutory interpretation by raising any as-applied constitutional challenge?

The inadequacy of the evidentiary record heightens the importance of my prudential concerns about the majority’s resolution of a statutory-interpretation issue that was neither seriously presented in the district court nor decided in the district court. The majority chooses to adopt a new exception to our rules of error preservation even though the record is devoid of basic information abo.ut marijuana, its constituent compounds and metabolites, and the ability of drug tests to reliably test for the presence of the drug. The majority is left to fill in the gaping holes in the eviden-tiary record with its own understanding of key scientific concepts and facts, posing the risk the court’s reasoning may be undermined by mistaken assumptions, impressions, and conclusions regarding marijuana. On such a weak record, this court should not resolve the important question of whether the language “controlled substance” in section 321J.2(l)(c) encompasses inactive metabolites of a controlled substance. >

Under these circumstances, I would also conclude further review was improvidently granted. Consistent with principles of judicial restraint, I would not disregard jurisprudential considerations by forging ahead with discretionary further review. See Short, 851 N.W.2d at 519 (Waterman, J., dissenting) (noting judicial restraint counters wide-open judicial activism and furthers “decisionmaking goals of clarity, efficiency, and principled reasoning”. (emphasis omitted) (quoting State v. Schwartz, 689 N.W.2d 430, 445 (S.D. 2004) (Konenkamp, J., concurring in result))); cf. also King, 818 N.W.2d at 39 (Waterman, J., concurring specially) (“I write separately to emphasize the importance of judicial restraint....”), Without reliable information in the record, courts risk making unsound decisions based on their own inadequately informed understanding of the scientific questions involved, aided only by sources they uncov*195er and their own assessments of the credibility of those sources. We should avoid that risk here by dismissing this application for further review as improvidently granted.

These prudential considerations caution against the majority’s conclusion that error is preserved on a statutory-interpretation claim that the district court failed to recognize and decide if the district court decides an unraised constitutional question. After the district court explained its denial of the motion based on the constitutionality of Iowa Code section 321 j.2(l)(c), Childs did not object nor did he request a ruling on the issue of statutory interpretation he now advances on appeal. I would thus conclude that error was not preserved. Although the State did not dispute error preservation on the poorly argued statutory-interpretation question, we should not allow the State’s posture to override our customary error preservation and prudential considerations.

Because I conclude Childs failed to preserve error on the statutory-interpretation issue, I would not address the merits of the question of how section 321J.2(l)(e) should be interpreted. I would leave for another day whether an inactive metabolite of a controlled substance is a controlled substance under section 321J.2(l)(c) and section 321J.1(4) (defining “controlled substance” as “any drug, substance, or compound that is listed in section 124.204 or 124.206, or any metabolite or derivative of the drug, substance, or compound”).

III. Comried.

The majority broadly reaffirms Com-ried’ s interpretation of “the plain meaning of the operative statutory language.” But it must be noted that the operative statutory language interpreted in that case is not the statutory language at issue in this case. In Comried, we interpreted the phrase “any amount” in Iowa Code section 321J.2(1)(c) to mean “any amount greater than zero.”12 693 N.W.2d at 778, The majority focuses on different language in this case, deciding whether an inactive secondary metabolite of marijuana is a “controlled substance” for purposes of section 321J.2(1)(c).

The majority cites the Harris dissent for the proposition that the flat ban imposed by Comried on driving with any amount of a controlled substance is justified by “the difficulty of detecting drug impairment,” given the rate at which impairing substances metabolize. See Harris, 322 P.3d at 165 (Timmer, J., dissenting). The majority also concludes that there is no roadside test that can measure for the presence of controlled substances in a driver’s body. Importantly, these factual conclusions have no evidentiary, basis in the record of this case and are subject to change based on the record in future cases, evolving information about marijuana, or the development of new methods of testing for the presence of any amount of a controlled substance.

In his special concurrence, Chief Justice Cady- cites Comried for the proposition that the purpose of Iowa Code section 321J.2(1)(c) is “to prohibit people from operating motor vehicles with controlled substances in their bodies, whether or not they are under the influence.” Comried, 693 N.W.2d at 776. Chapter 321J provides, *196however, that its legislative purpose is “to protect society, including drivers, from death or serious long-term injury.” Iowa Code § 321J.23(2). A conviction under section 321J.2(l)(e) “identifies [the defendant] as a risk to the health and safety of others, as well as to the intoxicated driver.” Id. § 321J.23(3). We “consider statutory text to be the best evidence of legislative intent or will” and construe statutes “to effect the expressed intent of the legislature.” See 2A Norman J. Singer & Shambie Singer, Statutes and Statutory Construction § 46:3, at 178 (7th ed. 2014). Consistent with this principle of judicial fidelity to expressed legislative intent, I am not convinced that the purpose of Iowa Code section 321J.2(l)(c) is to create an operating-while-intoxicated offense divorced entirely from the question of actual impairment, and thereby roadway safety.

Comried did not address what constitutes a controlled substance and thus does not control the question of whether the phrase “controlled substance” used in section 321J.2(l)(e) includes Carboxy-THC, an inactive metabolite of marijuana. To the extent that the majority reaches the question of whether Carboxy-THC is a controlled substance under section 321J.2(l)(c), I dissent.

IV. Conclusion.

After considering the briefs, record, and oral arguments in this case, I conclude that several problems undermine the soundness of the majority’s decision. In particular, the district court never ruled on the issue of statutory interpretation, the parties did not address the plain meaning of Iowa Code section 321J.2(l)(c) or assert any recognized exceptions to our plain-meaning rule, and the record is devoid of basic scientific evidence informing the court’s interpretation of the relevant statutory provisions. For these reasons, I conclude error was not preserved and further review was improvidently granted, and I respectfully dissent.

(dissenting).

I would dismiss this petition for further review as improvidently granted. I agree with Justice Hecht that the only question preserved in the district court was its constitutional holding, the only issue raised on appeal is a statutory claim, and as a result, neither is appropriate for our review. Further, the briefing on the statutory claim on appeal and in the district court was minimal. Erik Childs simply argues that our existing precedent, State v. Comried, 693 N.W.2d 773 (Iowa 2005), was wrongly decided because it relied on an Arizona case, State v. Phillips, 178 Ariz. 368, 873 P.2d 706 (App. 1994), which was subsequently distinguished in State ex rel. Montgomery v. Harris, 234 Ariz. 343, 322 P.3d 160, 164 (2014). That is the extent of the argument actually presented. This case is thus not a good vehicle for deciding some of the very important questions posed by Iowa Code section 321J.2 (2014). But the majority is determined to proceed to make its sweeping declarations about the statute. I find the case far more troubling than does the majority.

Iowa enacted the relevant provision of the present statute in 1998.1998 Iowa Acts ch. 1138, § 11 (codified at Iowa Code § 321J.2(1) (1999)). In Comried, we considered a vehicular homicide conviction under Iowa Code section 707.6A(1) (2001). Comried, 693 N.W.2d at 774. That conviction was based on a violation of section 321J.2, which provided that a person with any amount of a controlled substance in the body was guilty of intoxicated driving. Id. We held that “any” means “any.” Id. at 778. No constitutional issues were raised in Comried. See id. at 775-78. Childs invites us to reconsider the result in Comried. He *197relies in large part on developments in Arizona law, where an appellate court in Harris recently held that a similar statute should be narrowly construed to avoid absurd results. 322 P.3d at 164.

I start with the basic question—what is the purpose of the statute? That one is easy. We have said the purpose of the statute is to “promote public safety by removing dangerous drivers from the highways.” Bearinger v. Iowa Dep’t of Transp., 844 N.W.2d 104, 107 (Iowa 2014) (quoting State v. Vogel, 548 N.W.2d 584, 587 (Iowa 1996)). We have never found that a purpose of the statute was to stigmatize marijuana use or impose penalties on marijuana users because of their status.

The next question is whether the statute, if interpreted literally, fits the legislative purpose of addressing the danger of impaired drivers on the road. Here, we hit tougher terrain. The per se approach, which declares that the presence of any metabolite, active or inactive, is sufficient to support a criminal conviction and potential imprisonment, is clearly overbroad in light of the purpose' of the statute. The science is clear that the inactive metabolites of marijuana may remain in the body for weeks after consumption. See Nat’l Highway Traffic & Safety Admin., Drugs and Human Performance Fact Sheets, Cannabis/Marijuana, https://one.nhtsa. gov/people/injury/research/jobl85drugs/ eannabis.htm (last visited June 22, 2017) (stating detection time for THC metabolites in urine is well past the window of intoxication and impairment). Thus, many persons are subject to the statute even though their driving is not impaired in the least and their marijuana use was not recent. Assuming we behave rationally, we do not impose criminal penalties arising from behavior due to its danger when, in fact, the behavior is not dangérous.

The statute thus raises serious constitutional problems. I doubt that it is consistent with due process to subject a person to potential incarceration under 'a criminal law designed to prevent dangerous behavior when the behavior itself is not dangerous at all. It would be outrageous, in my view, to impose harsh sanctions on a driver who was exposed to marijuana weeks or months ago and poses no danger on the road, all in the name of highway safety. As noted by Justice Cavanagh in People v. Derror, “There is no rational reason to charge a person who inhaled. marijuana two weeks ago and who now decides to drive to the store to pick up a gallon of milk.” 475 Mich. 316, 715 N.W.2d 822, 846 (2006) (Cavanagh, J., dissenting), overruled by People v. Feezel, 486 Mich. 184, 783 N.W.2d 67, 86 (2010); see also Commonwealth v. Etchison, 916 A.2d 1169, 1174-78 (Pa. Super. Ct. 2007) (Bender, J., concurring in part and dissenting in part). These dissents emphasize that one cannot draw any reasonable conclusion of impairment solely from a positive test for canna-binoids. Derror, 715 N.W.2d at 846; Etchison, 916 A.2d at 1175; see also Feezel, 783 N.W.2d at 83, 86 (overruling the Derror majority and holding that a metabolite of THC is not a controlled substance under Michigan law). While it is true, of course, that no constitutional issues were raised in this appeal, we ordinarily interpret statutes to avoid constitutional problems. Simmons v. State Pub. Def., 791 N.W.2d 69, 73-74 (Iowa 2010).

Notably, we recently decided an important case which required that a defendant’s state of intoxication must be tied in a causal way to the injuries resulting in a case of homicide by vehicle. See State v. Adams, 810 N.W.2d 365, 371 (Iowa 2012). In Adams, the state argued that merely driving while intoxicated was sufficient to establish an offense under the statute. Id. at 368-69. We noted that criminal statutes *198are strictly construed against the state and that we would not produce an absurd result. Id. at 369. We concluded that the intoxication of the driver must be causally linked to the underlying death. Id. at 372; see Eric A. Johnson, Wrongful-Aspect Overdetermination: The Scope-of-the-Risk Requirement in Drunk-Driving Homicide, 46 Conn. L. Rev. 601, 605-06 & nn. 17-18 (2013) (describing the split in the courts on the question of whether scope-of-risk doctrine from tort law applies in criminal law setting of intoxicated driving). Applying Actoms-type logic here, the presence of a metabolite and use of marijuana must be a cause of harm or a cause of a risk of harm to support a criminal conviction. This argument, of course, was not raised-in this appeal.

Another constitutional problem with the statute is that it does not provide a person of ordinary intelligence with fair notice. Metabolites from marijuana can., be retained in a person’s system for days or weeks. A person who has consumed marijuana thus has no fair notice as to when he or she may legally drive a car. It may be a day, weeks, months, or even years. Consistent with the observation made by Justice Cavanagh in Derror, .the Childs majority’s interpretation “now criminalizes a broad range of conduct and makes criminals out of people who have no knowledge of the conduct that they must now seek to avoid.” 715 N.W.2d at 844. Suppose, for instance, one travels to Colorado on vacation and lawfully smokes marijuana. May that person lawfully drive back to Iowa when returning home? How long must the person wait before lawfully driving? Can anybody know? A driver with the majority opinion in their glove compartment will not find any useful advice on this issue.

Further, it is well established that metabolites of marijuana can be obtained through passive inhalation. See id.; cf. Daniel P. Mazo, Comment, Yellow Rows of Test Tubes: Due Process Constraints on Discharges of Public Employees Based on Drug Urinalysis Testing, 135 U. Pa. L. Rev. 1623, 1647 (1987) (“Research indicates that urinalysis also cannot discern active smoking of marijuana and hashish from passive inhalation.... ”); Kaye McDonald Sunderland & Coni S. Rathbone, Jar Wars: Drug Testing in the Workplace, 23 Willamette L. Rev. 529, 548 (1987) (“[Pjassive inhalation' must be considered as a possible sou'rce when interpreting low level test results.”). Under the State’s interpretation, a driver who had a trace of metabolite, based upon passive transmission, is subject to serious criminal offenses. But there is “no rational reason to charge a person who passively inhaled marijuana smoke at a rock concert a month ago and who now decides to drive to work.” Derror, 715 N.W.2d at 846.

In order to avoid all these problems, it might be asserted that the statute does not criminalize dangerous driving, but criminalizes the. status of being a recent user of marijuana. I doubt the legislature would bury a status crime in its driving statutes. In any event, an effort to justify the penalties on. marijuana use as a status offense would also run into serious constitutional problems. If the legislature sought to punish marijuana users for their status as marijuana users, the classification in the statute distinguishing marijuana users who happen to be driving from those who are, for instance, passengers, would be subject to attack as an irrational classification in violation of equal protection principles. The status of drivers and nondrivers who have metabolites of.marijuana would be the same, yet they are treated differently under the statute.

We have already invoked the absurdity doctrine in the area of drunk driving to avoid unintended convictions not related to *199the purpose of the statute of dealing with the danger of impaired drivers. In Bear-inger, we considered whether the prescription-drug defense applied to administrative actions involving the. revocation of drivers’ licenses. 844 N.W.2d at 105. Interestingly, the underlying criminal statute for OWI expressly contained such a defense, but the statute relating to revocations did not contain similar language. Id. at 107-08. Ordinarily, we would honor the legislative text. In Bearinger, however, we emphasized that the purpose of the statute was highway safety and that persons-who were driving while using prescription drugs as prescribed by a physician were not a danger and thus should not be subject to license revocation. Id. at 110.

In Bearinger, we did not rely on legislative text and call it a" day. Instead, we imported language into the legislative text to ensure that the purpose of the statute— namely, protecting the public against dangerous drivers—was advanced.-Why don’t we apply the same reasoning here and interpret the statute to mean an active metabolite? What accounts for the active and energetic approach in Bearinger to focus on actual public safety and the steadfast refusal to do so here? Is it a desire to express strong cultural disapproval of marijuana? If so, how is this rationally related to a statute designed to combat impaired drivers?

There is support in-the academic literature for a Bearinger/Ho/ms-type interpretation. The literature points out that the presence of minute amounts of a metabolite simply has no relationship with recent ingestion, let alone impaired driving. See Andrea Roth, The Uneasy Case for Marijuana as Chemical Impairment Under a Science-Based Jurisprudence of Dangerousness, 103 Cal. L. Rev. 841, 890 (2015) (“[A] prohibitionist approach is an awkward fit if the justification for- the law is the dangerousness of the drug’s impairing effects..'...”); Joshua C. Snow, The Unconstitutional Prosecution of Controlled Substance Metabolites Under Utah Code § 41-6A-517, 2013 Utah L. Rev. OnLaw 195, 203 (2013) (“[T]he presence of a metabolite in the body does not necessarily equate with present intoxication [and] does not even equate with recent ingestion.”).

" There is another policy-based rationale for giving the statute a narrow gloss. As Professor Steven Bender has observed, the history of marijuana legislation is based on racial stereotyping, and enforcement of open-ended marijuana laws leads to disproportionate enforcement against racial minorities. Steven W. Bender, The Colors of. Cannabis: Race and Marijuana, 50 U.C. Davis L. Rev. 689, 690 (2016). Bender traces the origin of strict marijuana legislation to “raeialized perceptions of users of color as threatening public safety and welfare.” Id. Bender notes the “disproportionate burden of marijuana enforcement on racial minorities.” Id. at 693. Racial minorities are subject to “Driving While Black” or “Driving While Hispanic.” Id. at 701-02; see also David A. Harris, “Driving While Black” and All Other Traffic Offenses:.The Supreme Court and Pretextual Traffic Stops, 87 J. Crim. L. & Criminology 544, 546 (1997). Thus, wittingly or not, inactive metabolite laws may be a contributing factor leading to disproportionate prison populations such as that experienced in Iowa. See generallyt Michelle Alexander, The New Jim Crow: Mass Incarceration in the Age of Colorblindness 59-96 (rev. ed. 2012).

The majority nonetheless concludes that the legislature intended to proceed with its sweeping regulation notwithstanding the problems cited above. The majority’s statutory approach requires persons with trace metabolites, but who pose no threat to public safety, to sacrifice personal free*200dom for the benefit of the community because more precise measurement tools have not been developed. Such persons are the statute’s roadkill under the majority’s interpretation. The people picking up the gallon of milk weeks after smoking marijuana or after being passively exposed to marijuana are not culpable under the statute, but their convictions under the statute amount to unfortunate collateral damage imposed because the State is looking for a convenient way to obtain convictions without the traditional methods of proving impairment on a case-by-case basis through ordinary evidentiary techniques.

I am not so sure. I recognize the difficulties in interpreting the statute in light of the specific statutory text. But did the legislature in 1998 intend for this absurd result? I recognize the standard of absurdity is a high bar. See Brakke v. Iowa Dep’t of Nat. Res., 897 N.W.2d 522, 534 (Iowa 2017). But one wonders whether the legislature was fully aware of the evolving science and the implications of the statutory text. Certainly some of the statute’s applications are absurd. Does the statute in full context introduce enough ambiguity to avoid untoward results?

A case can be made, perhaps, for upholding Comried based upon legislative acquiescence or stare decisis. The case for legislative acquiescence and stare decisis was much stronger in State v. Williams, 895 N.W.2d 856 (Iowa 2017), where an interpretation of the meaning of the term “arrest” had been repeatedly endorsed in multiple opinions over a thirty-seven-year period, the most recent of which, State v. Wing, 791 N.W.2d 243 (Iowa 2010), overruled by Williams, 895 N.W.2d at 867-73, was thoroughly reasoned. The Comried decision, however, is cryptic, does not explore the troublesome contours of a per se interpretation, does not recognize the constitutional issues, and has not been repeated in thorough opinions. Further, as has been pointed out by Justice Hecht, the statutory language has been amended since Com-ried.

The notion that broadly framed statutes can be narrowly interpreted is not a new concept. In Iowa Insurance Institute v. Core Group of Iowa Association for Justice, we held that a statute, which on its face required disclosure of “all information ... concerning the employee’s physical or mental condition relative to the claim,” did not include information protected by the work-product doctrine. 867 N.W.2d 58, 69, 79 (Iowa 2015) (quoting Iowa Code § 85.27(2) (2011)). We held that the statute should not be evaluated solely based on isolated words. Id. at 72. Instead, we insisted on looking at the statute’s larger context. Id. As we noted, there are many occasions when we have narrowed the apparently unqualified isolated terms of a statute. Id. at 73-74.

That is the type of reasoning I would apply here. In looking at the totality of the statute, its structure, and its purposes, one begins to question whether the legislature intended to include inactive metabolites notwithstanding the unqualified but isolated language used in the statute. I would be inclined to cinch up the statute in some fashion to avoid the untoward results that I doubt the legislature intended, either by requiring the presence of an active metabolite as in Harris, 322 P.3d at 164, or by requiring a causal link as in Adams, 810 N.W.2d at 371.

In any event, the cheers and jeers that will no doubt arise from today’s decision may be premature. The approach taken today may eliminate a less intrusive statutory-interpretation solution to the obvious problems of the statute. But weighty constitutional problems remain. Can criminal sanctions arise from application of this drugged-driving statute to someone who, *201in fact, poses no danger at all arising from consumption of marijuana, or maybe poppy seed rolls, in the past, the consumption of which demonstrably has no relationship to impaired driving? Where the inactive metabolite has no causal relationship to impaired driving, would any conviction be an impermissible status offense? Does the presence of a metabolite in any amount under the statute present an irrebuttable presumption contrary to due process? Does conviction of such persons under a jurisprudence of dangerousness serve any legitimate penal purpose under the Eighth Amendment or article I, section 18 of the Iowa Constitution? Would an enhanced criminal penalty under Iowa’s statute for repeat offenders be subject to a Bruegger-type challenge, where a very broad law involving a wide variation of conduct is combined with an escalating criminal sanction? See State v. Bruegger, 773 N.W.2d 862, 884 (Iowa 2009). Do due process and cruel-and-unusual-punishment concepts require the State to prove impaired driving on a case-by-case basis, like so many other crimes? If interpreted as a status crime, can the distinction between drivers and others who have recently ingested marijuana be defended from an equal protection challenge?

The statutory shoe has been dropped. The constitutional shoe will drop in future cases. The practical effect of today’s decision may well be to kick the can down the road and escalate future disputes to a constitutional dimension.

No one doubts, of course, the ability of the legislature to enact statutes that protect the public from drivers who are actually impaired. The question for the future is whether the legislature can establish a regime to control dangerous drivers that, in many applications, relies on a sweeping generality that is unsupported by science and does not utilize the traditional American way of requiring individualized guilt based on moral culpability before criminal sanctions are enforced.