BRACHTENBACH, J.

This case involves the Legislature's 1986 amendments to three statutes involving driving while under the influence of intoxicants (DWI): RCW 46.61.502 (defining DWI), RCW 46.61.504 (defining actual physical control of vehicle while under the influence of intoxicants), and RCW 46.61.506 (setting forth admissible evidence and tests to establish driving or physical control of vehicle while under the influence of intoxicants) (hereinafter referred to as 1986 amendments). The district court found the 1986 amendments unconstitutional. The superior court affirmed. We granted direct review and reverse.

Prior to the 1986 amendments, a person driving a vehicle in this state with "0.10 percent or more by weight of alcohol in his blood as shown by chemical analysis of his breath, blood, or other bodily substance" was guilty of DWI. See former RCW 46.61.502(1); see also former RCW 46.61.504, .506. Effective June 11, 1986, the Legislature redefined the "per se" DWI offense so as to provide that a person is guilty if he "has 0.10 grams or more of alcohol per two hundred ten liters of breath, as shown by analysis of his breath, blood, or other bodily substance . . ." (Italics ours.) See former RCW 46.61.502; Laws of 1986, ch. 153, §§ 2-5; see also former RCW 46.61.504 and. 506.

Under the former statutory scheme, breath tests were used to establish the requisite blood-alcohol ratio for the per se DWI offense. Thus, evidence of the relationship between breath alcohol (the test used) and blood alcohol (the statutory element of the per se DWI offense) was relevant to the issue of whether the State had proved the necessary blood alcohol ratio beyond a reasonable doubt.

The 1986 amendments, by redefining the per se DWI offense in terms of breath-alcohol ratio, render evidence of the relationship of breath alcohol to blood alcohol irrelevant because breath alcohol content now forms an element of the per se DWI offense. The State need not prove any blood alcohol level under the 1986 amendments. Instead, the State must prove the necessary breath-alcohol ratio beyond a reasonable doubt.

Each respondent involved here was charged with driving while under the influence of intoxicants pursuant to RCW 46.61.502(1) and (2) (as amended in 1986). Two respondents are females and one is a black male. Each respondent submitted to a breath test and was found to have at least 0.10 grams of alcohol per 210 liters of breath. Each case was assigned to the Seattle District Court for trial. At various pretrial proceedings, respondents challenged the constitutionality of the Legislature's 1986 amendment of RCW 46.61.502(1). Respondents moved to dismiss the charges pending against them, and, alternatively, to bar prosecution under former RCW 46.61.502(1) (the per se offense). Respondents also moved to suppress their breath test results as evidence of intoxication under any subsection of former RCW 46.61.502.

The district court granted respondents' motion to bar prosecution under former RCW 46.61.502(1) and to suppress the breath test results as evidence in prosecutions pursuant to the remaining sections of former RCW 46.61.502. The trial court concluded that the 1986 amendments were unconstitutional. Specifically, the court held that the 1986 amendments violate due process because they (1) include an improper "conclusive presumption" amounting to guilt by machine; (2) foreclose defendants' rights to present witnesses and evidence; and (3) violate principles of fundamental fairness. Moreover, the court ruled that the statutory scheme violates defendants' rights to confrontation and equal protection. The court found that the Legislature's intent in amending the statutes was to eliminate defendants' expert testimony relating to blood-breath ratios and thereby shorten trials and make convictions easier.

Based on the district court's rulings, the State sought and was granted review by the King County Superior Court. The superior court affirmed the district court's rulings.

SCIENTIFIC BACKGROUND

Alcohol contained only in the breath does not cause intoxication. It is the impact of alcohol on the central nervous system, particularly on the brain, that causes the physical and psychological changes associated with impairment. Alcohol reaches the central nervous system through the blood. When used to establish blood alcohol levels, breath testing devices use a mathematical constant to approximate the percentage of alcohol in the blood based on the amount of alcohol present in a breath sample. All breath-testing devices currently used in the United States, including the two devices certified for use in Washington — the Smith and Wesson Breathalyzer 900A and the BAC Verifier DataMaster, use a conversion factor of 2100:1. The 2100:1 conversion factor is an assumed blood-breath ratio. The blood-breath ratio represents the relationship between the number of alcohol molecules in the bloodstream to the number present in the breath when both substances are tested simultaneously. Thus, a 2100:1 conversion factor assumes that for each molecule of alcohol in a given volume of breath, there are 2,100 molecules of alcohol in the same volume of blood.

Because blood-breath ratios vary both between individuals, and at different times in the same individual, a breath test based on a 2100:1 blood-breath ratio may not accurately represent a particular individual's blood alcohol level. If the actual blood-breath ratio is lower than 2100:1, a breath test will overestimate blood alcohol, and vice versa. Factors influencing an individual's blood-breath ratio include body temperature, hematocrit level (the ratio between red blood cells and blood plasma), and the time at which alcohol was consumed in relation to the time breath alcohol is measured. Higher than normal body temperatures resulting from fevers, exercise, and menstrual cycle variations in women result in a lower blood-breath ratio than normal. If all other factors are the same in a given individual, a breath test based on that individual's normal blood-breath ratio, given when body temperature is elevated, will overestimate that individual's actual blood alcohol level.

Moreover, because of the way alcohol is transferred from the blood to the lungs, the amount of alcohol in an individual's breath is not constant. Breath testing machines are designed to test the last portion of a person's breath. Typically, there is less alcohol in the first part of the breath than in the last part, which comes from deeper portions of the lungs, the alveolar sacs, where the alcohol is transferred from the blood to the lung air.

I DUE PROCESS

The State challenges the district court's conclusion that the 1986 amendments violate defendants' due process rights by "creating a conclusive presumption that breath alcohol equals blood alcohol, and hence substitutes the judgement of a machine for the judgement of the jury." See conclusion of law 4, Clerk's Papers, at 69. The superior court affirmed this finding, concluding the statute was "arbitrary in the sense that it is guilt by computer". Oral Decision of Superior Court, at 20, Clerk's Papers, at 202.

[1, 2] A conclusive presumption requires the trier of fact to find the existence of an elemental fact upon proof of a basic fact. Thompson, The Constitutionality of Chemical Test Presumptions of Intoxication in Motor Vehicle Statutes, 20 San Diego L. Rev. 301, 306 (1983). A conclusive presumption cannot be used against a defendant in a criminal trial. See Morissette v. United States, 342 U.S. 246, 275, 96 L.Ed. 288, 72 S.Ct. 240 (1952).

The State argues that the 1986 amendments do not create any presumption that breath alcohol equals blood alcohol. Instead, the State contends the 1986 amendments redefine the DWI crime in terms of breath alcohol, rendering blood alcohol irrelevant. The State concludes that the Legislature merely replaced one indicator of driving impairment (blood alcohol) with another (breath alcohol).

Respondents argue that blood alcohol is not only an indicator of impairment as is breath alcohol, but actually causes impairment. Thus, by criminalizing breath alcohol, the Legislature attempts to indirectly regulate blood alcohol levels. Respondents conclude that this indirect regulation creates a conclusive presumption by allowing the State to prove a basic fact — breath alcohol — which requires the trier of fact to presume an elemental fact — blood alcohol and thus impairment.

The district court ruled that "alcohol in the blood, not the breath, . . . impacts behavior." Findings of fact on undisputed facts 19, Clerk's Papers, at 67. Respondents assert that this finding establishes that blood alcohol causes impairment. The State argues that finding 19 establishes only that blood alcohol relates to impairment because blood is the transporter of alcohol to the brain. The State concludes that the record contains no evidence to support a finding that blood alcohol causes impairment.

The record establishes that central nervous system alcohol actually causes impairment, Report of Proceedings, at 11, 108-09, and that blood is a transporter of alcohol to the central nervous system tissues, Report of Proceedings, at 13. While evidence exists that alcohol in the bloodstream always indicates presence of alcohol in the central nervous system, and thus may cause intoxication in a but/for sense, the evidence establishes that only central nervous system alcohol directly causes intoxication. See Report of Proceedings, at 10-13. Even respondents acknowledge that "[s]hort of performing a test of brain tissue in an autopsy, a test of the blood is the most direct possible measure of a person's alcohol level." Brief of Respondents, at 55. Therefore, although the district court's finding 19 establishes that blood alcohol is a more direct measure of intoxication than breath alcohol, it does not establish that criminalizing breath alcohol implicitly regulates blood alcohol as respondents argue. See also Oral Decision of Superior Court, at 18: ("[I]t is blood alcohol rather than breath alcohol that affects conduct . . . that may not be quite — very well explained, but what . . . [the district court judge] says is quite so.") Clerk's Papers, at 200.

This court upheld the former blood alcohol per se offense as an alternative method of committing a DWI offense in State v. Franco, 96 Wn.2d 816, 823, 639 P.2d 1320 (1982). Under the prior statutory scheme, a driver with .10 percent by weight of alcohol in his blood was per se DWI if the State proved the elements of the offense beyond a reasonable doubt. This court held that the per se statute did not create any presumption of impairment at that level of blood alcohol, but rather defined one alternative offense of DWI in those terms. See Franco, at 832. Similarly, the 1986 amendments simply redefine the former per se crime in terms of breath alcohol and proof of the offense is essentially the same under the amendments as it was in the statutory scheme upheld in Franco. See also Forte v. Texas, 686 S.W.2d 744, 747-48 (Tex.Ct.App. 1985), rev'd in part, 707 S.W.2d 89 (Tex.Crim.App. 1986); State v. Wilcox, 10 Ohio App.3d 11, 460 N.E.2d 323 (1983); State v. Ulrich, 17 Ohio App.3d 182, 478 N.E.2d 812 (1984); State v. Vannata, 8 Ohio Misc.2d 22, 456 N.E.2d 1358 (1983). We conclude that the 1986 amendments do not create an impermissible conclusive presumption.

We also disagree with respondents' contention that by redefining the per se DWI offense in terms of breath alcohol, the 1986 amendments take away a defendant's right to present evidence to dispute the reading of the breath testing device. While the 1986 amendments render evidence of blood alcohol irrelevant to the extent it is introduced to disprove the relationship between blood and breath alcohol (the State no longer must prove blood alcohol levels), defendants may still present evidence challenging the accuracy and validity of the breath test reading. Moreover, the State must still prove beyond a reasonable doubt the elements of the per se DWI offense. The 1986 amendments only redefine the element of blood alcohol in terms of breath alcohol.

Defendants have no due process right to the former definition of the DWI per se offense. See Godfrey v. State, 84 Wn.2d 959, 963, 530 P.2d 630 (1975); Seattle-First Nat'l Bank v. Kawachi, 91 Wn.2d 223, 244, 588 P.2d 725 (1978). Breath test evidence alone is not conclusive proof of the per se offense. See State v. Keller, 36 Wn. App. 110, 113, 672 P.2d 412 (1983). The State still must establish the foundational requirements of the breath test. Cf. State v. Baker, 56 Wn.2d 846, 355 P.2d 806 (1960).

WAC 448-12-230 states in part:

To obtain a valid breath test, it must be determined (a) that the subject has had nothing to eat or drink for at least fifteen minutes prior to the administration of the test, and (b) that the subject does not have any foreign substances, not to include dental work, fixed or removable, in his/her mouth at the beginning of the fifteen minute observation period.

The prosecution must also prove beyond a reasonable doubt that the machine was in proper working order and that the 0.10 percent reading was a correct one.

In concluding the test, the operator is to follow the instructions displayed by the instrument. The temperature of the solution in the simulator must be 34 Centigrade, plus or minus .2 Centigrade, prior to the time the test is given. The reading from the simulator test must be between .090 and .110 inclusive.

WAC 448-12-230. The officer who gives the test must be certified and the State must prove his competence to administer the test. See WAC 448-12-230 through -330.

The defense has the same opportunity to attack those results as it always has had under the prior presumptions and the immediate prior statutory scheme. See Franco, at 828; State v. Peterson, 100 Wn.2d 788, 674 P.2d 1251 (1984). The defendant may introduce evidence refuting the accuracy and reliability of the test reading. The fact finder has the duty to weigh all the evidence to determine, beyond a reasonable doubt, whether a defendant was driving with 0.10 grams or more of alcohol per 210 liters of breath.

Alternatively, respondents argue that if the 1986 amendments are interpreted so as to make the relationship between blood and breath alcohol irrelevant, then the amendments fail as being in excess of the state's police powers. Both courts below ruled that the 1986 amendments constituted a valid exercise of the state's police power. See conclusion of law 1, Clerk's Papers, at 68; Oral Decision of the Superior Court, at 15, Clerk's Papers, at 197.

[3] Pursuant to article 1, section 1 of the Washington Constitution, the State Legislature may prescribe laws to promote the health, peace, safety, and general welfare of the people of Washington. See, e.g., State v. Moore, 79 Wn.2d 51, 483 P.2d 630 (1971); State v. Melcher, 33 Wn. App. 357, 655 P.2d 1169 (1982). "[B]road discretion is thus vested in the legislature to determine what the public interest demands under particular circumstances, and what measures are necessary to secure and protect the same." Reesman v. State, 74 Wn.2d 646, 650, 445 P.2d 1004 (1968). A statute is a valid exercise of police power if it (1) tends to correct some evil or promote some interest of the State, and (2) bears a reasonable and substantial relationship to accomplishing its purpose. State v. Conifer Enters., 82 Wn.2d 94, 96-97, 508 P.2d 149 (1973).

A legislative enactment is presumed constitutional and the party challenging it bears the burden of proving it unconstitutional beyond a reasonable doubt. Caminiti v. Boyle, 107 Wn.2d 662, 732 P.2d 989 (1987); Convention Ctr. Coalition v. Seattle, 107 Wn.2d 370, 730 P.2d 636 (1986); Granat v. Keasler, 99 Wn.2d 564, 663 P.2d 830, cert. denied, 464 U.S. 1018 (1983); Duckworth v. Bonney Lk., 91 Wn.2d 19, 586 P.2d 860 (1978); State v. Conifer Enters., supra. Moreover, if a court can reasonably conceive of a state of facts to exist which would justify the legislation, those facts will be presumed to exist and the statute will be presumed to have been passed with reference to those facts. Conifer Enters., at 97; Moore, at 54. Thus, where scientific opinions conflict on a particular point, the Legislature is free to adopt the opinion it chooses, and the court will not substitute its judgment for that of the Legislature. State v. Dickamore, 22 Wn. App. 851, 855, 592 P.2d 681 (1979).

Here, the parties do not dispute that the purpose of the 1986 amendments is to reduce the drunk-driver hazard on the public highways. See, e.g., Fury v. Seattle, 46 Wn. App. 110, 111, 730 P.2d 62 (1986), review granted, 108 Wn.2d 1022 (1987); State v. Franco, 96 Wn.2d 816, 639 P.2d 1320 (1982); State v. Moore, 79 Wn.2d 51, 483 P.2d 630 (1971); Fritts v. Department of Motor Vehicles, 6 Wn. App. 233, 492 P.2d 558 (1971). Respondents contend, however, that the chosen means of furthering this goal — breath alcohol content — is too indirectly related to impairment to justify the statute. Respondents argue that the State has not established any evidence that directly links breath alcohol to impairment. Respondents point out their post-hearing objection to the district court's finding of fact (labeled undisputed) 20 which states "[m]ost studies done to determine the impact of alcohol were done using breath tests." Clerk's Papers, at 67. Respondents assert that no substantial evidence supports the court's finding. We disagree.

Witness David Predmore of the Washington State Toxicology Laboratory testified that the correlation between the amount of alcohol in a driver's breath and his ability to drive is "roughly the same as . . . [for] blood alcohol." Report of Proceedings, at 362. Moreover, the State points to several studies indicating that breath tests will underestimate blood alcohol levels 80 to 91 percent of the time while overestimating them only 5 to 6 percent of the time. See State's Exhibit 38 (Biasotti, The Role of the Forensic Scientist in the Application of Chemical Tests for Alcohol in Traffic Law Enforcement, 29 J. Forensic Sci. 1164, 1167-68 (1984)); State's Exhibit 40 (Parsons Dallosta, A Comparative Field Study of Blood and Breath Analysis With the Intoxilyzer 5000 (Sacramento Cy. District Attorney's Laboratory of Forensic Services)); State's Exhibit 39 (Jones, Determination of Liquid/Air Partition Coefficients for Dilute Solutions of Ethanol in Water, Whole Blood, and Plasma, 7 J. Analytical Toxicology 193 (1983)); State's Exhibit 35 (Washington State Patrol Breath Test Correlation Data); State's Exhibit 37 (Washington State Toxicology Laboratory, Blood-Breath Comparison Study (1984)). The State contends that this evidence alone shows a reasonable and substantial relationship between breath alcohol and impairment based on its relationship to blood alcohol, which was found to be an acceptable means in State v. Franco, supra. The State also asserts that the record contains substantial evidence that breath alcohol is directly related to impairment. There is testimony that the vast majority of impairment studies correlating alcohol consumption with driving abilities have relied on measures of breath alcohol, not blood alcohol. Report of Proceedings, at 250-51, 361-62. Washington State Patrol Sergeant Rod Gullberg explained that "breath becomes just from a statistical standpoint a better predictor of impairment than does blood simply because of the predominance of the data." Report of Proceedings, at 251. The record also contains legislative memoranda indicating that "[m]ost studies of alcohol-related driving impairment have in fact used . . . grams per liter of breath measurement rather than a blood level measurement." Memorandum to House Judiciary Committee, Clerk's Papers, at 61. See also House Bill 1499, Clerk's Papers, at 56-57 (virtually all scientific data supporting DWI per se laws are based on correlations between breath alcohol content and driving impairment).

Respondents contend that the State's evidence does not demonstrate any information directly linking breath alcohol readings with impairment. Although respondents point out various factors affecting the relationship between blood alcohol and breath alcohol, contending that although breath alcohol was used to study impairment those studies ultimately correlated the breath levels to blood levels, respondents do not point to any evidence in the record to rebut the State's position. While the record may establish that breath is a less direct measure of blood alcohol levels, it does not establish a lack of a reasonable and substantial relationship between breath alcohol and impairment. We conclude that the 1986 amendments are a valid exercise of police power.

We also conclude that the breath standard used in the 1986 amendments gives fair notice and is not unconstitutionally vague. While the trial court concluded that the word "breath" as used in the 1986 amendments is not unconstitutionally vague, see conclusion of law 3, Clerk's Papers, at 69, the superior court apparently ruled that the breath standard is too vague to give fair notice and thus violated due process. The superior court ruled, "we have a floating standard and that is a concept that I find very difficult to accept in connection with a criminal statute as the right of people to know what they are defending against". Oral Decision of the Superior Court, at 20, Clerk's Papers, at 202.

[4] In determining whether a statute is unconstitutionally vague on its face, the court applies a 2-part test. First, does the statute "provide fair notice, measured by common practice and understanding, of that conduct which is prohibited, so that persons of reasonable understanding are not required to guess at the meaning of the enactment[?]", and second, does the statute "contain ascertainable standards for adjudication so that police, judges, and juries are not free to decide what is prohibited and what is not, depending on the facts in each particular case[?]." State v. Carter, 89 Wn.2d 236, 239-40, 570 P.2d 1218 (1977). Respondents agree that the second part of the Carter test is met by the 1986 amendments. They argue, however, that the amendments are unconstitutionally vague because they fail to convey a sufficiently definite warning of proscribed conduct to defendants. Respondents contend that no external standards are available from which a driver can predict when his breath level is at or above the prohibited level. Moreover, because of the imperfect correlation between breath and blood alcohol, respondents assert that blood alcohol charts are insufficient to give notice of the prohibited conduct. We disagree.

In State v. Franco, 96 Wn.2d at 824-25, this court held that the former 0.10 percent blood alcohol standard was not unconstitutionally vague. The court found that the standard gave fair warning in that it was reasonable to assume that a driver was impaired at that level of blood alcohol and that charts were available showing the number of drinks necessary to produce the prohibited level. Franco, at 825.

Likewise, we find it reasonable to assume that drivers are impaired at the statutory breath level and that drivers still have notice under the 1986 amendments that driving while impaired by alcohol is a crime. We note that the breath level offense is one alternative DWI offense; another alternative is to drive while impaired. See former RCW 46.61.502(2). Overall, the 1986 amendments give fair notice of the forbidden conduct.

[5] Additionally, we do not consider respondents' argument regarding the State's toxicologist's definition of breath adopted in WAC 448-12-220. Respondents' argument is a challenge to the administrative regulation and is not properly before this court. The issue was not raised by the State on this appeal and is not properly raised by respondents. The outcome of the issue does not provide on alternative grounds for upholding the trial court's decision. Respondents' argument amounts to a vagueness challenge to a legislative enactment based on a challenge to rules promulgated by an administrative agency. We conclude that the 1986 amendments do not violate respondents' due process rights.

II EQUAL PROTECTION

The State challenges the trial court's conclusion that the 1986 amendments violate equal protection by discriminating against persons with blood-breath ratios below 2100:1. See conclusion of law 5, Clerk's Papers, at 69. The State asserts that the 1986 amendments criminalize a particular breath alcohol level without regard to blood alcohol levels. Because no conversion of breath readings into blood readings is required, a person's blood-breath ratio has no impact on guilt. The State argues that the statute distinguishes only between those who have the prohibited amount of alcohol in their breath and those who do not. The State concludes that breath alcohol is rationally related to impairment and does not violate equal protection guaranties.

Respondents argue that, although facially neutral, the 1986 amendments discriminate in fact against persons who have blood-breath ratios below 2100:1 because although those persons may have a prohibited breath alcohol level, that level does not correlate with a blood alcohol level demonstrating impairment. Respondents assert that this discriminatory impact violates the equal protection rights of several classes of individuals. Respondents contend that because blacks have smaller lung capacities than same-sized white persons, blacks' blood-breath ratios are lower and a breath alcohol test will overestimate their blood alcohol levels. Moreover, persons suffering from certain chronic lung diseases also may have smaller than normal lung capacities also resulting in overestimated blood alcohol levels. Finally, respondents point out that denture wearers, who retain a greater amount of mouth alcohol, can have a breath-alcohol level that overestimates their blood alcohol.

Respondents urge this court to apply to their equal protection claims an "invigorated rational basis test" under which a statute will be upheld only if it "may fairly be viewed as furthering a substantial interest of the State." See Plyler v. Doe, 457 U.S. 202, 217-18, 72 L.Ed.2d 786, 102 S.Ct. 2382 (1982). Respondents rely on State v. Phelan, 100 Wn.2d 508, 671 P.2d 1212 (1983). In Phelan, a Board of Prison Terms and Paroles sentencing practice denied credit for jail time served to persons too poor to make bail. The court applied an intermediate scrutiny test because the presentence detention involved "both a deprivation of liberty in addition to that which would otherwise exist, and a classification based solely on wealth . . ." Phelan, at 514. See also State v. Cook, 37 Wn. App. 269, 271, 679 P.2d 413 (1984) (applying intermediate scrutiny to pretrial detention of juveniles).

The State disagrees, arguing that this case is distinguishable from Phelan on the basis that no classification of persons according to any suspect or semi-suspect class is made and because, unlike the situation in Phelan, no liberty interest is affected until after a conviction under the 1986 amendments. Indeed, this court has applied a rational basis standard to an equal protection challenge to a statute denying jury trials to juveniles. In State v. Schaaf, 109 Wn.2d 1, 21, 743 P.2d 240 (1987), the court distinguished Phelan, and declined to apply an intermediate scrutiny standard because the statute did not directly implicate physical liberty. The State argues that because the 1986 amendments are facially neutral and because there is no evidence that the breath alcohol standard has a disparate impact on a suspect class, the classification should be subject to a rational basis standard of review.

[6] Here, the record demonstrates no classification, express or implied, based on a suspect or semi-suspect class. Moreover, no liberty interest is directly affected by the amendments. Unlike the defendant in Phelan, who was already incarcerated, a defendant under the 1986 amendments may or may not be convicted and suffer a loss of liberty. See Schaaf, at 21 (juveniles may or may not be convicted, court refused to speculate on the effect a jury would have and thus applied rational basis test). Therefore, we apply a rational basis standard of review.

[7] Under a rational relationship test, a legislative classification is subjected to minimal scrutiny and will be upheld unless it rests on grounds wholly irrelevant to the achievement of a legitimate State objective. Schaaf, at 17; Phelan, at 512. Here, the parties agree that the purpose of the 1986 amendments is to reduce the hazards of drivers who are impaired by alcohol. The classifications are those persons who have the prohibited breath alcohol level and those who do not. The record contains substantial evidence linking driving impairment with the statutorily prohibited breath alcohol level. Although the record contains evidence that breath alcohol may not accurately show blood-alcohol in particular individuals, the record does not establish that proscribing a particular level of breath alcohol is wholly irrelevant to achieving the purpose of reducing DWI hazards.

Moreover, the record does not establish that a particular race or class of persons is affected by the classification. The State maintains that to the extent the trial court's conclusion suggests a discriminatory impact on those groups, no findings of fact support it. We agree. While the record indicates that the average black person has a 13 percent smaller lung capacity than the average white person, the record contains no evidence suggesting that breath alcohol readings inaccurately measure impairment among blacks, nor that the average statistic affects particular black individuals. A defense expert testified that blood-breath ratios in blacks had never been studied. Report of Proceedings, at 124. Similarly, although the record demonstrates that some chronic lung disease and the presence of dentures can result in breath alcohol levels that overestimate blood alcohol levels, nothing in the record establishes that these characteristics are peculiar to a particular class. The relationship between the legislative classification and its goal is rational. We conclude that the 1986 amendments do not violate equal protection guaranties.

III EQUAL RIGHTS AMENDMENT

Respondents contend that because the 1986 amendments have a disparate impact on women, the amendments violate this state's Equal Rights Amendment, Const. art. 31. Respondents raise this issue as an additional ground supporting the lower court rulings finding the 1986 amendments unconstitutional. See Tropiano v. Tacoma, 105 Wn.2d 873, 718 P.2d 801 (1986); Peterson v. Hagan, 56 Wn.2d 48, 351 P.2d 127 (1960). The district court did not rule on the merits of respondents' ERA claim. See conclusion of law 8, Clerk's Papers, at 70.

Article 31, of the Washington Constitution provides: "Equality of rights and responsibility under the law shall not be denied or abridged on account of sex." The protections provided by the ERA go beyond those of the equal protection clause of the federal constitution and the privileges and immunities clause of the state constitution. Darrin v. Gould, 85 Wn.2d 859, 540 P.2d 882 (1975). The former equal protection approach to gender-based discrimination no longer applies, having been "swept away by the equal rights amendment." Marchioro v. Chaney, 90 Wn.2d 298, 305, 582 P.2d 487 (1978), aff'd, 442 U.S. 191 (1979). Under the ERA, if equality is restricted or denied on the basis of sex, the classification is discriminatory. Marchioro, at 298. Respondents argue that facially neutral legislation having a disparate impact on women violates the ERA.

Respondents argue that the 1986 amendments, although facially neutral, discriminate against women on the basis of their blood-breath ratios. Respondents' argument is based on the premise that alcohol contained only in the breath has no impact on intoxication. See findings of fact on undisputed facts 19, Clerk's Papers, at 67; Report of Proceedings, at 46-69, 163, 165. Instead, only when breath alcohol levels reflect blood alcohol levels is breath alcohol an indicator of intoxication because blood is the transporter of alcohol to the central nervous system and central nervous system alcohol causes intoxication. Respondents contend that because alcohol in the blood always indicates alcohol in the central nervous system, blood alcohol can also be said to "cause" intoxication. Moreover, because the correlation between breath alcohol and blood alcohol depends upon the relationship of volume of breath to volume of blood in a particular individual, the blood to breath ratio used determines the accuracy of a blood alcohol level calculation made from a breath test.

Respondents contend that although the 1986 amendments purport to criminalize only breath alcohol, the standard adopted in 1986 amendments implicitly assumes a 2100:1 blood-breath ratio because studies of impairment were made by giving breath tests using that ratio, which were then used to calculate blood alcohol levels. Respondents point out that the 2100:1 blood-breath ratio may be inaccurate for any particular individual.

Respondents also suggest that the 2100:1 ratio used by breath testing machines to convert a breath alcohol reading into a blood alcohol reading was chosen based on studies of mostly male subjects. See Brief of Respondents, at 8; Defense Exhibit 16 (3 percent of the people examined in various published blood-breath correlation studies available for review were women). Because women tend, on an average, to have lower blood-breath ratios than men, findings of fact on undisputed facts 14(d), respondents reason that the State's use of a male based norm (the 2100:1 blood-breath ratio) in breath testing has the effect of overestimating blood alcohol levels for women.

Respondents thus argue because breath tests overestimate women's blood alcohol on this basis, the breath alcohol standard has a disparate impact on women. The district court found that "[t]heoretically, the average female would have a blood-breath ratio 8-15 percent lower than the average male." Findings of fact on undisputed facts 14(d), Clerk's Papers, at 66.

Respondents note also that the blood-breath ratio is affected by lung size. Respondents point out that women have a smaller average lung capacity than do men. Report of Proceedings, at 69-71. A person with smaller lungs will exhale more of their lung capacity into a breath machine than will a person with average lung capacity. (This assumes that the people are the same size except for lung capacity). Thus, a smaller lunged person will exhale more deep lung air than would a person of the same size with average lung size by virtue of emptying more of his or her lungs. The result is that a greater proportion of deep lung air is tested in a person with smaller lungs than a person with average lungs. Because the deep lung air contains more alcohol, the person with smaller lungs will have a higher breath alcohol reading because of lung size.

Respondents also contend that women's blood-breath ratios are lower than the average male's because women tend to have higher average body temperatures. See findings of fact on undisputed facts 12, Clerk's Papers, at 66. Body temperature is one of the primary components of the blood-breath ratio; the blood-breath ratio is affected 6 percent for each degree in variance from the average human temperature of 37 degrees centigrade. Thus, other factors being equal, a person with a temperature of 38 degrees would have a breath alcohol level 6 percent higher than actual blood alcohol levels. Respondents contend that the variation between average body temperature in men and women means that the presumed blood-breath ratio of 2100:1 overestimates blood alcohol by approximately 4 to 6 percent for women test subjects. Report of Proceedings, at 58. Respondents thus argue that because the 1986 amendments implicitly incorporate the 2100:1 blood-breath ratio in the breath alcohol standard, the amendments have a disparate impact on women as a class.

The State argues that the record before the court does not support a finding of disparate impact. The State acknowledges that the district court found that women theoretically have lower blood-breath ratios than do men and that women tend to have higher body temperatures on average than men. However, the State contends that no scientific evidence establishes conclusively that these characteristics are exclusively gender-based. In fact, the district court found that "[t]here have been no studies specifically undertaken concerning differences in blood-breath ratios between men and women." Findings of fact on undisputed facts 14(c), Clerk's Papers, at 66. The State concludes that respondents' postulated disparate impact, drawn from two findings of fact, and unsupported by evidence in the record is insufficient to invalidate the 1986 amendments as violative of the ERA. We agree.

[8] Although the record demonstrates that breath alcohol levels may not accurately represent blood alcohol levels in particular individuals, this evidence alone does not establish that the breath alcohol standard falls more harshly on women as a class. Respondents bear the burden of showing that the Legislature's decision to use a breath alcohol standard to prove impairment falls more harshly on women. Here, no studies have been made of the differences in blood-breath ratios between men and women. At best the record shows that the use of blood-alcohol levels may have a hypothetical disparate impact on women. Here we are concerned with breath alcohol only. The record shows no evidence that breath alcohol levels represent different levels of impairment between men and women as a class. We conclude that, on this record, respondents have failed to satisfy their burden to establish disparate impact, and, thus, respondents have not shown that the 1986 amendments violate the ERA.

IV LEGISLATIVE INTENT

The State challenges the district court's conclusion that the Legislature's intent in enacting the 1986 amendments "was to eliminate expert testimony by the defense relating to blood breath ratios and thereby shorten trials and make convictions easier." Conclusion of law 7, Clerk's Papers, at 65. The district court did not state whether this conclusion was considered as a basis for holding the 1986 amendments unconstitutional, and the superior court did not address it.

The State is correct that a court may not strike down an otherwise constitutional statute on the basis of an alleged illicit legislative motive. See United States v. O'Brien, 391 U.S. 367, 20 L.Ed.2d 672, 88 S.Ct. 1673 (1968); cf. Convention Ctr. Coalition v. Seattle, 107 Wn.2d 370, 375, 730 P.2d 636 (1986) (statement by one legislator does not establish legislative intent).

Respondents, nevertheless, contend that legislative intent may be a basis for holding a law unconstitutional if the legislative history discloses that a "substantial or motivating factor" behind enactment of the law was an intent to discriminate, and the State cannot demonstrate that the law would have been enacted without this factor. See Hunter v. Underwood, 471 U.S. 222, 85 L.Ed.2d 222, 105 S.Ct. 1916, 1920 (1985).

Here, the district court based its conclusion on legislative history. The court focused on comments by one senator, a representative from the Washington Association of Prosecuting Attorneys, and a Washington State Patrol sergeant, all to the effect that the Legislature meant to eliminate defense experts' testimony about blood-breath ratios, confusing the jury, and dragging things out. See Memorandum Decision of the District Court, Clerk's Papers, at 73.

The legislative history also discloses, in a House Bill Report, that the Legislature intended to relate the DWI definition more closely to scientific data supporting the law and to reduce state resources expended in prosecuting DWI defendants. See House Bill 1499, Clerk's Papers, at 56-57. Overall, the legislative history does not indicate any intent to discriminate against a particular class of individuals. In any event, the comments of one senator and two members of the public are insufficient to demonstrate that any such intent was a substantial motivating factor in the Legislature's enactment of the 1986 amendments. See Convention Ctr. Coalition, at 375.

Therefore, to the extent the district court relied upon legislative intent as a basis for holding the amendments unconstitutional, we reverse. Moreover, the conclusion that the Legislature intended to eliminate defense testimony as to blood-breath does not support a conclusion that the Legislature intended to discriminate against a certain class of persons.

V ADMISSIBILITY OF BREATH TEST EVIDENCE

Because we conclude that the 1986 amendments are constitutional, we reverse the district court's ruling barring the State's use of breath test evidence under former RCW 46.61.504. Moreover, we disagree with respondents' contention that breath test evidence remains inadmissible because defendants cannot obtain independent tests as provided for in RCW 46.61.506(5). Respondents rely on the district court's finding that "as a practical matter, virtually all breath testing devices" are police controlled. See findings on disputed facts 5, Clerk's Papers, at 68. Respondents argue that the State thus impermissibly "thwarts" defendant's attempts to exercise their rights to additional tests. See Blaine v. Suess, 93 Wn.2d 722, 612 P.2d 789 (1980). In Suess, a DWI defendant's conviction was reversed because police officers "thwarted" his attempts to obtain an independent blood test. The defendant indicated he desired an independent blood test after being informed of his right to do so. Police initially indicated that they would transport him to a hospital for testing, but then took him to jail instead. The court concluded that although "law enforcement authorities have no duty to volunteer to arrange for testing, they must not thwart an accused's attempts to make such arrangements." Suess, at 728.

The facts of this case are distinguishable from those before the court in Suess. Here, respondents have shown no evidence of acts by the State that misled or interfered with their attempts to obtain additional tests. Moreover, the record shows that breath testing devices are present in nonstate controlled locations, including detoxification centers, hospitals, and work release facilities. Report of Proceedings, at 236. One defense expert testified that breath testing instruments are present in hospitals. Report of Proceedings, at 170, 200. Finally, we note that RCW 46.61.506 specifically states that a defendant's inability to obtain an independent test will not bar admission of a properly administered breath test. See RCW 46.61.506(5). The statutory DWI scheme, along with the warnings given to defendants, adequately informs defendants of their rights to independent tests. Breath test evidence is admissible to prove violations of the DWI statutes.

Reversed.

PEARSON, C.J., UTTER, DOLLIVER, DORE, CALLOW, GOODLOE, and DURHAM, JJ., and CUNNINGHAM, J. Pro Tem., concur.