762 So.2d 601 (2000)
STATE of Louisiana
v.
Joseph Davis FERRIS.
No. 99-KA-2329.
Supreme Court of Louisiana.
May 16, 2000.
Rehearing Denied June 16, 2000.
Richard P. Ieyoub, Atty. Gen., Brent C. Coreil, Dist. Atty., Reggie Paul Dupre, Jr., Houma, Mary Ellen Hunley, Baton Rouge, Counsel for Applicant.
Allen Bruce Rozas, Mamou, Counsel for Respondent.
LEMMON, Justice.[*]
This is a direct appeal from a judgment of the district court that declared unconstitutional La.Rev.Stat. 14:98.1, which defines the crime of "underage driving under the influence" as the operating of a motor vehicle when the operator is under the age of twenty-one and has a blood alcohol concentration (BAC) of 0.02 percent or more. The issue on appeal is whether Section 98.1 violates the equal protection guarantee of La. Const. art. I, § 3 against arbitrary discrimination based on age, since a person age twenty-one or over with a BAC of 0.02 percent cannot be criminally charged under either this statute or La. Rev.Stat. 14:98, which declares criminal the operation of a motor vehicle only when the operator's BAC is 0.10 percent or more.
Facts
On May 22, 1998, defendant was arrested and charged with violating La.Rev.Stat. 14:98.1. At the time, defendant was eighteen years old and was operating a motor vehicle with a BAC of 0.07 percent.
Defendant filed a motion to quash the bill of information, alleging that La.Rev. Stat. 14:98.1 violates the equal protection rights of persons under twenty-one years of age contrary to La. Const. art. I, § 3.[1] Defendant asserted that the statute unconstitutionally discriminates on the basis of age in that it declares criminal the operation of a motor vehicle by a person under the age of twenty-one with a BAC of 0.02 percent or more, while a person twenty-one and older can legally operate a motor vehicle with a BAC between 0.02 and 0.10 percent.
At the trial of the motion, the judge refused the State's offer into evidence of the record in Manuel v. State, 95-2189 (La.3/8/96), 692 So.2d 320.[2] On the evidence admitted into the record, the trial judge declared the statute unconstitutional and granted the motion to quash. The State filed a direct appeal to this court.
Finding error in the trial court's exclusion of clearly relevant evidence (the record in Manuel), this court set aside the judgment and remanded the case to the trial court to conduct an evidentiary hearing, at which all relevant evidence under La.Code Evid. art. 402 was to be admitted as to whether Section 14:98.1 arbitrarily, capriciously, or unreasonably discriminates against persons under twenty-one years of age. 98-2442 (La.5/18/99), 747 So.2d 487.
On remand, the trial judge again declared the statute unconstitutional as violative of the Louisiana constitutional prohibition of arbitrary age discrimination. This direct appeal followed. La. Const. art. V, § 5(D).
Underage Driving under the Influence
In 1994, the Louisiana Legislature amended La.Rev.Stat. 14:98, the general statute on driving while intoxicated, to include a provision setting forth a lower maximum BAC content for minors driving motor vehicles. Specifically, Section 98, as amended, defined the crime of operating a vehicle while intoxicated (DWI) as the operating of any motor vehicle when "[t]he operator's blood alcohol concentration is 0.10 percent or more by weight, or is 0.04 percent or more by weight if the operator is under the age of eighteen years ...." (emphasis added).[3]
In 1997, the Legislature repealed the portion of former La.Rev.Stat. 14:98 underscored in the previous paragraph and enacted an entirely separate statute, La. Rev.Stat. 14:98.1, which provides in pertinent part:
A. The crime of underage operating a vehicle while intoxicated is the operating of any motor vehicle, aircraft, watercraft, vessel, or other means of conveyance when the operator's blood alcohol concentration is 0.02 percent or more by weight if the operator is under the age of twenty-one based on grams of alcohol per one hundred cubic centimeters of blood.
B. Any underage person whose blood alcohol concentration is found to be in violation of R.S. 14:98(A)(1)(b) shall be charged under its provisions rather than under this Section.
C. On a first conviction, the offender shall be fined not less than one hundred nor more than two hundred fifty dollars, and participate in a court-approved substance abuse and driver improvement program.
D. On a second or subsequent conviction, regardless of whether the second offense occurred before or after the first conviction, the offender shall be fined not less than one hundred fifty dollars nor more than five hundred dollars, and imprisoned for not less than ten days nor more than three months. Imposition or execution of sentence shall not be suspended unless:
(1) The offender is placed on probation with a minimum condition that he serve forty-eight hours in jail and participate in a court-approved substance abuse and driver improvement program; or
(2) The offender is placed on probation with a minimum condition that he perform ten eight-hour days of court-approved community service activities, at least half of which shall consist of participation in a litter abatement or collection program and participate in a court-approved substance and driver improvement program.[4]
Arbitrary Age DiscriminationThe Manuel Decision
In Manuel v. State, 95-2189 (La.7/2/96), 692 So.2d 320 (on reh'g), this court upheld, over constitutional age discrimination arguments almost identical to those raised in this case, the legislative acts raising the minimum drinking age to twenty-one years.
The Louisiana Constitution's prohibition of arbitrary age discrimination must "be tempered with the concept that the young and the old can be given preferential treatment." Lee Hargrave, The Louisiana Constitution: A Reference Guide 25 (1991)(noting that even the Constitution draws several lines based on age). In determining a violation of this constitutional prohibition of age discrimination, this court has adopted an intermediate level of scrutiny. The Manuel decision explained this standard, as follows:
La. Const. art. I, § 3 sets up a spectrum for analyzing equal protection challenges based on discriminatory classifications....
. . . .
In the middle of the spectrum are laws that classify persons on the basis of the six grounds enumerated in the third sentence of Section 3 [one of which is age]. A law containing a statutory classification based on any of the six enumerated grounds does not enjoy the usual presumption of constitutionality. Moreover, with that reversal of the ordinary presumption of constitutionality comes a reversal of the rule that ordinarily places the burden of proof on the party seeking a declaration of unconstitutionality....
. . . .
.... Because age classification is specifically enumerated in Section 3 and because an age classification must have a nonarbitrary basis, the burden of proof is on the proponent of constitutionality to show that the statute establishing such a classification substantially furthers an appropriate governmental purpose.
95-2189 (on reh'g) at pp. 4-5, 692 So.2d at 339-340 (emphasis in original).
In Manuel, we applied this standard of middle level scrutiny to statutes that prohibited the sale of alcoholic beverages to persons eighteen to twenty years old, as well as the purchase and public possession by such persons. Stating that the State is assigned the burden of justifying such a classification by showing that it substantially furthers an appropriate state purpose, we noted that the principal governmental purpose of the discriminatory statutes was improving highway safety. We acknowledged as undisputed that improving highway safety is an important governmental purpose and focused the analysis on whether the selected method substantially furthered that purpose. We explained that the term "substantially furthered" means the governmental purpose must be a substantial, as opposed to an incidental, reason for the classification. 95-2189 (on reh'g) at p. 6, 692 So.2d at 340 n. 5.
While the opinion on original hearing had focused on Louisiana data and on comparing three-year age groupings, we held on rehearing that the proper focus was to compare licensed drivers in the eighteen-to-twenty age group with the group of licensed drivers who were twenty-one and over. The issue thus framed was whether drawing the line at twenty-one substantially furthered the State's significant interest of promoting highway safety. Explaining our focus on the discrimination against adults in the lowest range of adulthood, we stated:
[T]he challenged statutes focus on the problem of youthful drinking and driving. If youthful drinking and driving is a substantial problem in highway safety and if increasing the minimum drinking age is an appropriate means of attacking that problem, the minimum drinking age can only be increased at the lowest level the eighteen, nineteen and twenty-year-old group. Thus, the statutes do not single out the members of this three-year age group for unequal treatment as compared to other three-year age groups. This is the only age group that can be affected by a three-year increase in the minimum drinking age.
95-2189 (on reh'g) at p. 8, 692 So.2d at 341 n. 7 (emphasis in original).
We concluded that there was a sufficient relationship between the discriminatory classification resulting from raising the minimum drinking age and the statutory objective of reducing youthful drinking and driving to improve highway safety, primarily relying on national statistical data, as corroborated by experience, logic and common sense. Based on experience, we reviewed national highway statistics which established, among other things, that the disadvantaged group (drivers from eighteen to twenty years old) was involved in twice as many accidents per capita as the general group of licensed drivers and that Louisiana ranked higher than forty-six other states in the percentage of alcohol-related fatalities involving drivers under twenty-one. Based on logic and common sense, we reasoned that the increase in the drinking age establishes a "[p]rohibition of drinking by persons who are proportionately the most dangerous group of drinking drivers" and therefore "has to increase highway safety substantially, as opposed to incidentally." 95-2189 (on reh'g) at p. 10, 692 So.2d at 342. We quoted a seasoned state trooper's apt remark that the eighteen-to-twenty-year-old age group "is not only inexperienced at driving but is also inexperienced at drinking." 95-2189 (on reh'g) at p. 8, 692 So.2d at 341.
We further took judicial notice that all of the other states had enacted similar legislation raising the drinking age to twenty-one years to comply with the National Minimum Drinking Age Act, 23 U.S.C. § 158, observing: "Admittedly, those legislative decisions are not subject to the same equal protection scrutiny as the Louisiana statutes. Nevertheless, unanimous utilization of this approach to the problem is a significant indication that this approach `substantially furthers [the] appropriate state purpose ....'" 95-2189 (on reh'g) at p. 8, 692 So.2d at 341.
Evidence in the Present Case
At the trial of the motion to quash, the State introduced the record of the Manuel case[5] and three additional witnesses, along with supporting documents. Defendants introduced one witness and supporting documents.
The 1997 National Highway Traffic Safety Administration (NHTSA) Report found an over-involvement of persons eighteen to twenty years of age in traffic accidents. Alcohol-related traffic fatality rates, based on the percentage of licensed drivers in each age group, were over twice as high in the disadvantaged group as the rates for persons twenty-one and over. Moreover, more persons in the disadvantaged group died in 1997 in 0.01 to 0.09 percent BAC crashes than did persons in any other contiguous age group. In Louisiana, the statistics generally showed equal or greater representation in fatal crashes and in alcohol-related fatal crashes, especially in low BAC fatal crashes.
Richard Compton, NHTSA senior research psychologist and technical advisor who conducted research and evaluation studies in highway safety, analyzed national studies of alcohol-related crashes based on population, number of licensed drivers, and number of vehicle miles traveled. He opined:
[T]here is no question that 18 to 20 year olds have a higher rate of involvement in alcohol-related crashes. They represent a group that is inexperienced at drinking and inexperienced at driving and the combination has proven to be quite deadly. They get in more crashes when drinking than people at a higher age. That is the basis [sic] justification for having a lower BAC limit.
Compton concluded with the observation that "[a]lcohol leads to elevated crash risk and it does so at a higher rate for those under 21."
Dr. Helmut Schneider, a statistics professor and a consultant for the Louisiana Highway Safety Commission (LHSC), compiled and analyzed traffic records data for the LHSC from 1994 to 1997. He confirmed that drivers between eighteen and twenty years of age are significantly over-represented in alcohol-related fatal crashes and even more so in alcohol-related fatal crashes with low BAC percentages. He verified that the risk of alcohol-related crashes is higher, at the same BAC levels, for those who are younger than twenty-one than for older drivers. Although eighteen-to-twenty-year-old drivers represent only five percent of licensed drivers, they were responsible for twenty-three percent of low level alcohol-related Louisiana fatalities in 1997 and nearly ten percent of all alcohol-related crashes. He further demonstrated, on a chart of low BAC crashes by single age groups, that the high crash rate peaks at age twenty and then declines quickly after that age group.
Dr. Richard Scribner, a physician who qualified as an expert in preventive medicine and prevention of alcohol-related problems, compiled crash data from 1986 through 1993. He concluded from that data that eighteen-to-twenty-year-old drivers presented the highest risk of alcohol-related fatal crashes and injury crashes in Louisiana.[6] He added that the laws raising the minimum drinking age had been the most widely studied preventive intervention in his field of expertise and have constituted one of the most successful projects.
Defendant presented a witness who had compiled the number of DWI arrests prosecuted to convictions in district court in Evangeline Parish from 1989 through 1997, and in Ville Platte City Court from 1994 until 1999. These figures showed that the eighteen-to-twenty-year-old drivers were not the group with the greatest amount of DWI convictions.
Regarding this evidence of DWI convictions in one parish, Compton asserted that the total number of DWI arrests and convictions for a particular age group is not the proper measure for determining the risk presented by that age group of being involved in an alcohol-related crash. According to Compton, DWI arrests merely indicate who the police are catching and thus represent a random sampling of drinking drivers, rather than an actual count of drinking drivers who cause injury or death. Dr. Schneider further stated that the DWI figures (if assumed to be applicable statewide), combined with the statistics used by him, might show "there are not many 17 to 20 year olds driving under the influence of alcohol, but they crash a lot when they have alcohol." He added that crash data represent a count of total crashes, while DWI convictions are a sampling and are useless, in the field of statistics, in determining the risk of traffic crashes presented by the population of drivers who are drinking.
Arbitrariness of Classification
Analogizing this case to Manuel, we conclude that the age classification in Section 98.1 provides for an even closer "classificatory fit"[7] than the statutes in Manuel. The governmental purpose of improving highway safety bears a closer correlation with Section 98.1, which addresses the dual activities of underaged drinking and driving, than with the statutes at issue in Manuel, which addressed only underaged drinking. Moreover, we repeatedly stated in Manuel that the real problem targeted by the statute was "youthful drinking and driving." Consequently, our reasoning in Manuel applies as well to the issue of the arbitrariness of the classification in the present case.
Defendant argues, however, that the statistical evidence relied upon in Manuel and in this case is fatally flawed in at least one significant respect, namely, the national alcohol-related fatal crash statistics are based in large part not on actual BAC testing, but rather on "imputation." As explained by Compton, imputation is a complex statistical procedure used to impute the BAC for crashes in which a BAC is unavailable. In those cases, a variety of factors are considered to arrive at the imputed BAC value, including the time of day, day of week, type of road, age and gender of driver, and type of vehicle being driven. Nationwide crash data, collected over many years, are analyzed to produce this imputation procedure. The procedure has been tested by comparing the imputed values to actual statistics from states in which the BAC was tested in a high percentage of crashes, and the margin of error has been slim. We thus find no merit to defendant's argument that the use of this well-accepted, statistically sound procedure renders improper the expert's reliance on this national statistical data.
Defendant further contends that the trial court correctly concluded the Evangeline Parish DWI data was entitled to more weight than the national statistics submitted by the State. Noteworthy, the very same type of DWI data was presented by the plaintiffs in Manuel. In rejecting the relevance of that local DWI data and criticizing the trial court's apparent reliance on such DWI data from Evangeline Parish, we reasoned that "[n]ot only did the statistics not show the percentage of drivers in the disadvantaged group in the parish, but also there was no consideration that young drinking drivers are frequently not arrested for the first incident." Manuel, 95-2189 at p. 11, 692 So.2d at 342 n. 8 (on reh'g).
Moreover, statistics are relevant primarily to provide corroborative support for broad sociological propositions that are widely accepted by experts. In this case, the proposition is that lowering the BAC limit for drivers under twenty-one to a "zero tolerance" level of 0.02 percent is substantially related to improving overall highway safety by reducing alcohol-related accidents. For the reasons recited herein as well as the reasons recited in Manuel, we conclude that the State has met its burden of establishing that the classification in La.Rev.Stat. 14:98.1 substantially, and not just incidentally, furthers the admittedly appropriate governmental purpose.
Finally, defendant contends that even if it was appropriate for the Legislature to classify eighteen-to-twenty-year olds for separate civil treatment based on promoting highway safety, it was inappropriate to criminalize conduct by members of this age group when identical conduct by persons twenty-one and over is not punishable as a crime.
Significantly, the statutes declared constitutional in Manuel, which prohibited the purchase and public possession of alcohol by persons under twenty-one years of age, provided a criminal penalty for violations thereof. Moreover, the defining of conduct as criminal is properly a legislative function, subject to constitutional restrictions. We discern no constitutional violation in the Legislature's providing a criminal penalty for operating a motor vehicle by a person who is drinking in small amounts and cannot drink legally at all.
The Legislature provided significantly lesser penalties for violation of Section 98.1 than for violation of Section 98. Defendant, of course, does not take issue with the lesser penalties imposed by Section 98.1, but takes issue with a provision that permits a violation of Section 98.1 to remain on a person's record and thus, according to defendant, potentially to serve as an enhancement of a subsequent DWI conviction.
The Legislature purposefully carved the "zero tolerance" law here at issue out of the driving while intoxicated statute, creating an entirely separate statute. In that separate statute, the Legislature did not include any provision for retaining a violation of Section 98.1 on the defendant's records. Indeed, a record retention provision was placed only in La.Rev.Stat. 32:853, which addresses motor vehicle operating records.[8]
More significantly, the DWI statute expressly limits those criminal offenses which can be used to enhance a subsequent DWI offense to the following:
For purposes of determining whether a defendant has a prior conviction for violation of this Section, a conviction under either R.S. 14:32.1, vehicular homicide, R.S. 14:39.1, vehicular negligent injuring, or R.S. 14:39.2, first degree vehicular negligent injuring, or a conviction under the laws of any state or an ordinance of a municipality, town, or similar political subdivision of another state, which prohibits the operation of any motor vehicle, aircraft, watercraft, vessel, or other means of conveyance while intoxicated, while impaired, or while under the influence of alcohol, drugs, or any controlled dangerous substance shall constitute a prior conviction. This determination shall be made by the court as a matter of law.
La.Rev.Stat. 14:98 F(1). Because there is no express reference to Section 98.1 in the enumeration of Louisiana criminal offenses that can be utilized to enhance a subsequent DWI conviction, we conclude that a violation of Section 98.1 may not be used to enhance a subsequent DWI conviction, regardless of how long it is retained in a person's records.[9] This conclusion is buttressed by the Legislature's carving out of Section 98.1 the conduct that would otherwise be a violation of La.Rev.Stat. 14:98 A(1)(b) and providing that such conduct "shall be charged under its provisions rather than under this Section." La.Rev. Stat. 14:98.1 B.
Decree
For these reasons, the judgment of the trial court is reversed, and the defendant's motion to quash is overruled. The case is remanded to the district court for further proceedings.
MARCUS, J., dissents and assigns reasons.
MARCUS, J. (dissenting).
I adhere to the views expressed in my dissent in Manuel v. State, 95-2189 (La.3/8/96), 692 So.2d 320. The "zero tolerance" law at issue here would be acceptable if it applied to all persons and did not discriminate against a group based on age. Accordingly, I respectfully dissent.
NOTES
[*] Kimball, J., not on panel. Rule IV, Part 2, § 3. Facts
[1] La. Const. art. I, § 3 provides:
No person shall be denied the equal protection of the laws. No law shall discriminate against a person because of race or religious ideas, beliefs, or affiliations. No law shall arbitrarily, capriciously, or unreasonably discriminate against a person because of birth, age, sex, culture, physical condition, or political ideas or affiliations. Slavery and involuntary servitude are prohibited, except in the latter case as punishment for crime. (emphasis added)
[2] In Manuel, this court on rehearing overturned the trial court's declaration of unconstitutionality of the series of statutes that prohibit the sale of alcoholic beverages to persons eighteen to twenty years old, as well as the purchase and public possession of alcoholic beverages by such persons. See La. Rev.Stat. 14:93.10-93.14, and La.Rev.Stat. 26:286. We held that such statutes were not violative of the prohibition against arbitrary age discrimination in La. Const. art. I, § 3.
[3] A constitutional challenge to the now-repealed portion of La.Rev.Stat. 14:98 was considered by this court in State v. Smith, 96-1798 (La.10/21/97), 700 So.2d 493. The trial court in Smith had declared that portion of the statute unconstitutional. This court for procedural reasons (defendant's lack of standing to seek post-conviction relief because he had completed the term of probation) declined to entertain that constitutional issue.
[4] The 1997 legislative amendment was apparently enacted in compliance with a federal mandate to enact "zero tolerance" laws setting a BAC limit of 0.02 or less for drivers of motor vehicles who are under the age of twenty-one. See 23 U.S.C. § 410(b)(1)(D). The express purpose of such laws was "to reduce traffic safety problems resulting from individuals driving while under the influence of alcohol." 23 U.S.C. § 410(a)(1). Such "zero tolerance" laws have been enacted by a large number of states, perhaps as many as forty-four. See NHTSA Web Site, The Facts: Zero Tolerance (visited April 10, 2000) .
[5] The opinion in the Manuel case included a digest of the evidence that need not be repeated here. See Manuel v. State, 95-2189 (La.7/2/96), 692 So.2d 320 (on reh'g).
[6] Other statistics showed that twenty-one to twenty-four-year-old drivers were involved in more alcohol-related crashes in 1996 than were the eighteen-to-twenty-year-old drivers. However, Dr. Scribner explained that determination of the risk posed by a particular group requires analysis of statistical data over several years. The trend over a series of years shows that the eighteen-to-twenty-year-old drivers present the highest risk of alcohol-related crashes.
[7] By "classificatory fit" is meant the relationship between a classification and the governmental purpose. This term is generally used in connection with the concepts of under and over inclusiveness. As we noted in Manuel, the statutes at issue there were "overinclusive," in that they prohibited young adults from drinking even when they would not be driving. 95-2189 at p. 14, 692 So.2d at 344 n. 10 (on reh'g). La.Rev.Stat. 14:98.1, on the other hand, addresses the dual problem of drinking and driving while drinking.
[8] The relevant provision is La.Rev.Stat. 32:853 A(1)(d), which states:
The operating record shall not include those arrests or convictions for a first or second violation of R.S. 14:98.1, if at least two years have elapsed since the date of conviction and the offender has not been convicted of an additional offense for violation of R.S. 14:98 or 98.1. If he has been convicted of one or more of these additional offenses during the two-year period, then the earlier offense shall remain on his record for a period of four years from the date of that conviction.
[9] In this regard, the amicus filed in this court points out that the source of La.Rev.Stat. 14:98.1 was the Arkansas statute for underaged driving under the influence. The Arkansas Supreme Court has held that its similarly defined crime of underaged driving under the influence is not a lesser included offense to DWI because it requires an element lacking for DWI, namely being under the age of twenty-one. McElhanon v. State, 329 Ark. 261, 948 S.W.2d 89 (1997).