(dissenting).
I respectfully dissent. In this case, William Frazier claims that MinmStat. § 609.229 (2000) has a disparate impact on people of color and that, as applied to him, the statute violates the Minnesota Constitution by depriving him of equal protection of the law.1 The court holds that the statistical evidence offered by Frazier fails to establish that section 609.229 has a disparate impact on people of color. I conclude, however, that the evidence, in the form of Minnesota Sentencing Guidelines *840Commission data, does show that section 609.229 has a disparate impact both on blacks and people of color generally. Furthermore, because the classification created by section 609.229 is not genuine and relevant to the purpose of the statute, the statute does not pass constitutional muster under the Minnesota rational basis test.
I.
The threshold legal question raised by Frazier’s argument is whether Frazier, as a black man, can argue that section 609.229 has a disparate impact on all people of color, as opposed to just blacks. With the exception of Native Americans, for which the sample size was one, and the group, “other,” which may or may not refer to a racial minority group, the application of section 609.229 similarly affects blacks and Asians, as it concerns convictions under the statute. Although Asians are more likely than blacks or whites to be convicted under section 609.229, this does not mean that a problem does not exist with respect to blacks, who are still three times more likely to be convicted than whites. The statistics show that the disparate impact alleged is a multi-ethnic problem, involving more than a one-to-one correlation between only blacks and whites. See United States v. Bd. of Educ. of City of Chicago, 554 F.Supp. 912, 920-21 (N.D.Ill.1983) (applying similar analysis in grouping all minorities together to determine if schools had been satisfactorily desegregated). As such, it is appropriate in this particular context to group minorities together for the purpose of analyzing Frazier’s disparate impact claim. See Kohn v. Minneapolis Fire Dep’t, 583 N.W.2d 7, 12-14 (Minn.App.) rev. denied (Minn. Oct. 20, 1998) (concluding that district court did not err in grouping all minorities together because number of Hispanics affected was too small to stand alone and there was similar past history of discrimination against other minority groups). In any event, the data presented by Frazier establishes that the statute has a disparate impact not only on people of color generally, but also on blacks. Thus, for purposes of this case, it makes little difference whether the appropriate class is defined as African Americans or people of color.
The aim of section 609.229 is to prevent and punish violent group criminal activity. However, by restricting the definition of “criminal gang” to groups that have a “common name or common identifying sign or symbol,” Minn.Stat. § 609.229, subd. 1(2) (2000), the statute singles out and applies a harsher penalty to only a portion of the people who engage in violent group criminal activity. Although this definition may be race-neutral on its face, its effect has been to target people of color. In the past five months alone, this court has decided at least two criminal cases in which a white defendant was a member of a group that did not fall within the statute’s definition of criminal gang merely because the group lacked a “common name or common identifying sign or symbol.” See State v. Stewart, 643 N.W.2d 281 (Minn.2002); McCollum v. State, 640 N.W.2d 610 (Minn.2002).
In McCollum, the defendant murdered the victim after the victim allegedly informed the police of the location of two of the defendant’s associates. 640 N.W.2d at 612-17. Several of the defendant’s associates were involved in the sale of illegal drugs. Id. at 613. One of them had been selling drugs out of the victim’s house, and another had a white powdery substance in his possession when he was arrested by the police. Id. Before his death, the victim was severely beaten over the course of three days by a group that included the defendant and six other individuals. Id. at 614-15. It appears that the murder in McCollum was designed to encourage *841compliance with the group’s internal rules by driving home the message that group members must protect one another from law enforcement.
In Stewart, the defendant was convicted of murder for his role in the drive-by-shooting of a bicyclist on Mississippi River Boulevard in Saint Paul. 643 N.W.2d at 283-84. Two other men were in the car with the defendant when the defendant pulled the trigger. Id at 285. In the weeks leading up to the shooting, the three men had agreed to commit armed robbery and murder, and had driven along the boulevard on several occasions with a loaded gun. Id. at 285, 288. The men used the term “first blood” to refer to the first member of the group to kill someone. Id. at 285. There was also evidence that the defendant and one of the men had engaged in past criminal activity together by participating in a burglary and agreeing to commit a murder for hire. Id. at 289-91. The fact that neither the defendant in McCol-lum nor the defendant in Stewart was charged with committing a crime for the benefit of a gang underscores the underin-elusive nature of section 609.229 and the effect of including the “common name or common identifying sign or symbol” requirement within the definition of criminal gang.
Before examining the evidence of disparate impact presented by Frazier in this case, it should be made clear that there is an important difference between the equal protection jurisprudence developed by the United States Supreme Court through its interpretations of the United States Constitution and the approach adopted by this court under the Minnesota Constitution. A party who challenges a statute on equal protection grounds under the United States Constitution must prove the existence of discriminatory intent or purpose. McCleskey v. Kemp, 481 U.S. 279, 292, 107 S.Ct. 1756, 95 L.Ed.2d 262 (1987); Village of Arlington Heights v. Metro. Housing Dev. Corp., 429 U.S. 252, 264-65, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977). Discriminatory purpose, the Court has explained, implies that the state decision maker selected or reaffirmed a particular course of action at least in part “because of,” and not merely “in spite of,” its adverse affects upon an identifiable group. McCleskey, 481 U.S. at 298, 107 S.Ct. 1756. Under the federal approach to equal protection, then, a party cannot successfully challenge a statute by merely showing that it has a disproportionate negative impact on the members of a particular racial minority group.2 McCleskey, 481 U.S. at 298-99, 107 S.Ct. 1756.
In State v. Russell, 477 N.W.2d 886 (Minn.1991), this court utilized a more protective approach under the Minnesota Constitution to strike down a facially neutral criminal statute that had a discriminatory impact on blacks. In Russell, five black men challenged the constitutionality of a statute that penalized the possession of crack cocaine more severely than the possession of powder cocaine. Id. at 887. The men claimed that the crack cocaine statute violated the equal protection guarantees of the federal and state constitutions because it had a discriminatory impact on blacks. Id. The men did not trace the statute’s discriminatory impact to any discriminatory intent or purpose on the part of the legislature or those charged *842with implementing the statute. This court analyzed the statute’s distinction between possession of crack cocaine and possession of powder cocaine using the three-pronged Minnesota rational basis test. The test requires:
(1) [t]he distinctions which separate those included within the classification from those excluded must not be manifestly arbitrary or fanciful but must be genuine and substantial, thereby providing a natural and reasonable basis to justify legislation adapted to peculiar conditions and needs;
(2) the classification must be genuine or relevant to the purpose of the law; that is there must be an evident connection between the distinctive peeds peculiar to the class and the prescribed remedy; and
(3) the purpose of the statute must be one that the state can legitimately attempt to achieve.
Russell, 477 N.W.2d at 888. Explaining its rationale for applying the Minnesota rational basis test to the crack cocaine statute, this court stated that “[t]here comes a time when we cannot and must not close our eyes when presented with evidence that certain laws, regardless of the purpose for which they were enacted, discriminate unfairly on the basis of race.” Id. at 888 n. 2.
Under Minnesota law, then, a defendant who challenges the constitutionality of a statute on the basis that it has a disparate impact on the members of a minority racial group is entitled to review of the statute under the Minnesota rational basis test if the defendant shows that the statute falls more harshly on one group than another. Russell, 477 N.W.2d at 888.3 In this case, Frazier has presented data which shows that, over a period of five years, 30.8% of those convicted under section 609.229 are black. The data also shows that people of color account for 89.7% percent of those convicted under the statute. Because the data presented here shows that blacks are three times more likely, and that people of color generally are 8.8 times more likely, to be convicted under section 609.229 than whites, Frazier has succeeded in showing that the effects of the statute fall more harshly on blacks and people of color than whites. I do not see the point in requiring Frazier to undertake further data analy-ses, all for the purpose of showing what can already plainly be seen from the conviction data itself. Indeed, in its oral argument to this court, the state repeatedly conceded that the data from the Minnesota Sentencing Guidelines Commission show that the statute has a disparate impact on people of color. Having made this showing, Frazier is entitled to have this court review the statute under the Minnesota rational basis test.
The court makes much ado over the reliability of Frazier’s data, namely the racial distribution of the population of confirmed gang members in Minnesota and the indicators of gang membership. Aside from the fact that the statistics concerning the racial distribution of confirmed gang members is only marginally relevant to analyzing Frazier’s challenge,4 the court fails to point out that the state does not *843dispute the reliability or validity of this data, which it produced. Moreover, it is illogical to require Frazier to present information that would enable the court to evaluate the reliability of the indicators of gang membership, given that these are the indicators that the state currently uses, regardless of whether such indicators are, in fact, reliable. Unlike the court, I can find no reason to question what the court alternately calls the reliability and validity of this data.
In Russell, this court accepted data similar to Frazier’s as the basis for its conclusion that the crack cocaine statute had a disparate impact on blacks. In Russell, the district court was presented with statistics showing that, of all persons convicted under the crack cocaine statute over a one-year period, 96.6% were black. 477 N.W.2d at 887 n. 1. If conviction statistics covering a one-year period were sufficient to establish disparate impact in Russell, I see no reason why conviction statistics covering a five-year period are insufficient to establish disparate impact in this case. “[T]o dictate that several must suffer discrimination before one could object * * * would be inconsistent with the promise of equal protection for all.” Batson v. Kentucky, 476 U.S. 79, 96, 106 S.Ct. 1712, 90 L.Ed.2d 69 (1986) (citing McCray v. New York, 461 U.S. 961, 965, 103 S.Ct. 2438, 77 L.Ed.2d 1322 (1983) (Marshall, J., dissenting from denial of certiorari)).
In sum, by presenting data indicating that blacks account for 30.8% of those convicted and that people of color generally account for 89.7% of those convicted under section 609.229, Frazier has demonstrated that the statute has a disparate impact on blacks and people of color. Because Frazier has demonstrated that the statute has a disparate impact, the next step is to examine the constitutionality of the statute through the application of the Minnesota rational basis test. As this court stated in Russell, “[i]t is particularly appropriate that we apply our stricter standard of rational basis review in a case * * * where the challenged classification appears to impose a substantially disproportionate burden on the very class of persons whose history inspired the principles of equal protection.” 477 N.W.2d at 889.
II.
On occasion, this court has applied the three prongs of the Minnesota rational basis test in reverse order. Kolton v. County of Anoka, 645 N.W.2d 403, 412 (Minn.2002); Metropolitan Sports Facilities Comm’n v. County of Hennepin, 478 N.W.2d 487, 489 n. 4 (Minn.1991). I believe that it is appropriate to employ this approach here. Thus, the first question is whether the purpose of the statute is one that the state may legitimately attempt to achieve. Metropolitan Sports Facilities Comm’n, 478 N.W.2d at 489 n. 4; Guilliams v. Comm’r of Revenue, 299 N.W.2d 138, 142 (Minn.1980). According to the state, the purpose of section 609.229 is to prevent and enhance the penalty for group criminal activity. The state cites legislative history showing that the legislature enacted the statute to counter the increase in violent criminal street gang activity. The state also notes that the Hennepin County Attorney’s Office was forced to dismiss felony charges in three cases during 1990-1991 after key witnesses were threatened by criminal gang members.
It is beyond dispute that the state may act to provide for the security, benefit, and protection of the people by prohibiting conduct that is detrimental to public safety. See Minn. Const, art. 1, § 1 (“Government is instituted for the security, benefit and protection of the people * ⅜ *.”); Russell, 477 N.W.2d at 891 (stating that *844penalizing and eradicating drug dealers is a legitimate state purpose); State v. Vernon, 283 N.W.2d 516, 519 (Minn.1979) (stating that criminalizing possession of cocaine is a legitimate state purpose). As such, the purpose of section 609.229 is one that the state may legitimately attempt to achieve.
The second question is whether the classification created by the statute is genuine or relevant to its purpose; that is, whether there is an evident connection between the distinctive needs peculiar to the. class and the prescribed remedy. Russell, 477 N.W.2d at 888; Guilliams, 299 N.W.2d at 142. The classification at issue in this case is created by the following definition of “criminal gang” set forth in section 609.229, subdivision 1:
■ As used in this section, “criminal gang” means any ongoing organization, association, or group of three or more persons, whether formal or informal, that:
(1) has, as one of its primary activities, the commission of one or more of the offenses listed in section 609.11, subdivision 9;
(2) has a common name or common identifying sign or symbol; and
(3) includes members who individually or collectively engage in or have engaged in a pattern of criminal activity.
This provision draws a distinction between groups that fall within the definition of criminal gang and groups that would fall within the definition but for the lack of a common name or common identifying sign or symbol. Rephrasing the second prong of the Minnesota rational basis test to fit the particular circumstances of this case, the inquiry is whether the statute’s exclusion of groups that lack a common name or common identifying sign or symbol from the definition of criminal gang is genuine or relevant to its purpose of preventing and penalizing violent group criminal activity.
The state contends that the classification is genuine or relevant because common names and common identifying signs and symbols lie at the heart of what it means to be a criminal gang. Citing a commentator, the state notes that “[vjisual presence is very much a part of the gang ambiance. Gang colors, jackets, or other distinct articles of clothing, hairstyles, and distinctive manners all provide objective positive evidence marking members of criminal street gangs.” There is no evidence before us indicating the presence of a relationship between an individual’s style of dress, hairstyle, or “distinctive manners” and their propensity to engage in violent group criminal activity. Moreover, the legislature has explicitly rejected the idea that physical appearance, rather than conduct, can be used to identify criminal gang members. See Minn.Stat. § 299A.64, subd. 2 (2000) (requiring the criminal gang oversight council to develop a strategy to eliminate the harm caused by criminal gangs and providing that the strategy “must target individuals or groups based on their criminal behavior, not -their physical appearance”). By identifying certain physical characteristics as indicators of criminal gang membership, the state’s argument unfairly stigmatizes and casts suspicion on individuals who possess those characteristics but are not members of criminal gangs.
The state’s suggestion that common names or common identifying signs or symbols are somehow unique to criminal gangs is also unconvincing. There are countless law-abiding business, political, educational, social, religious, sports, and other organizations throughout the state that have common names or common identifying signs or symbols. Thus, it cannot be said that those who fall within the *845challenged classification are more likely to participate in the conduct proscribed by section 609.229 than those who do not.
Furthermore, by requiring that criminal gangs have a common name or common identifying sign or symbol, section 609.229 embraces some individuals who participate in violent group criminal activity but excludes others. The previous discussion of our decisions in McCollum and Stewart demonstrates that section 609.229 is un-derinclusive. The purpose of the statute is not served by punishing Frazier, who sold drugs to a police informant for the benefit of a group that falls within section 609.229’s definition of “criminal gang,” more severely than the defendants in McCollum and Stewart, both of whom committed violent first-degree murder for the benefit of a group that has all of the ingredients of a criminal gang except a common name or common identifying sign or symbol. The groups involved in this case, in McCollum, and in Stewart, are all quite similar in terms of their participation in the kind of conduct that section 609.229 was designed to prevent. Indeed, the group in McCollum participated in the same type of criminal activity — the sale of illegal drugs — as the group to which Frazier belonged. Because Frazier is similarly situated to the defendants in McCollum and Stewart, they should all be treated alike under the law. See State v. Merrill, 450 N.W.2d 318, 321 (Minn.1990) (“The equal protection clause of the Fourteenth Amendment requires that all persons similarly situated be treated alike under the law.”); In re Harhut, 385 N.W.2d 305, 310 (Minn.1986) (“The equal protection clauses of the federal and state constitutions require that all persons similarly situated be treated alike under the law.”). Yet section 609.229 distinguishes between them and imposes an enhanced penalty on just one of them.
The state also argues that groups that have a common name or common identifying sign or symbol are more cohesive than other groups, and that this cohesiveness enables criminal gangs to grow into sophisticated criminal enterprises. As illustrated by McCollum, however, groups involved in criminal activity can employ other methods of securing the allegiance of their members. In that case, the group employed violence and, ultimately, murder as a means of enforcing group cohesiveness. In addition, the state’s assertion that groups that have a common name or common identifying sign or symbol are more cohesive than those who do not is not accompanied by any citation to the legislative history that would indicate that the legislature had the cohesion or sophistication of criminal groups in mind when it created the classification. More fundamentally, there is no evidence in the record to support the proposition that having a common name or common identifying sign or symbol fosters cohesiveness or enhances a group’s level of sophistication. In short, there is no basis in the record on which to conclude that limiting the application of section 609.229 to individuals who commit crime for the benefit of a group that has a common name or common identifying sign or symbol is genuine or relevant to the purpose of preventing group criminal activity.
The fact that section 609.229 takes only a partial step toward the fulfillment of its objective usually would not warrant a conclusion that the law is unconstitutional: Indeed, we generally will not hold a statute unconstitutional “merely because it does not assure complete amelioration of the evil it addresses.” Mack v. City of Minneapolis, 333 N.W.2d 744, 751 (Minn.1983); see Federal Distillers, Inc. v. State, 304 Minn. 28, 43-44, 229 N.W.2d 144, 156 (1975). However, this principle does not apply when, as here, the partial step taken *846by the legislature has the effect of singling out people of color and subjecting them to enhanced criminal penalties. Because Frazier has sustained his burden of establishing beyond a reasonable doubt that section 609.229 has a disparate impact on blacks specifically and people of color generally and, because the classification created by section 609.229, subdivision 1, is not genuine or relevant to the statute’s purpose, section 609.229, as applied to Frazier, is unconstitutional. I would therefore remand the case to the district court for resentencing with the instruction that Frazier’s sentence shall not be enhanced pursuant to Minn.Stat. § 609.229, subd. 2 (2000).
. Frazier also claims that section 609.229 violates equal protection because the penalties it imposes for crimes committed for the benefit of a criminal gang are more severe than those imposed by Minnesota's RICO Act, Minn.Stat. §§ 609.901-.912 (2000), for participation in criminal acts in connection with an enterprise; and that the definition of “criminal gang” found in Minn.Stat. § 609.229, subd. 1 (2000), violates the United States Constitution because it classifies individuals as members of criminal gangs based on conduct that is protected by the First Amendment. Given my conclusions with respect to Frazier’s disparate impact claim, it is unnecessary to address his other claims.
. This is not to say that evidence of disproportionate impact is completely irrelevant under the federal approach. Washington v. Davis, 426 U.S. 229, 241, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976). The Court has stated that "an invidious discriminatory purpose may often be inferred from the totality of the relevant facts, including the fact, if it is true, that the law bears more heavily on one race than another.” Davis, 426 U.S. at 242, 96 S.Ct. 2040; see McCleskey, 481 U.S. at 293, 107 S.Ct. 1756.
. Disparate impact results from practices that, although neutral on their face, fall more harshly on one group than another. Langlie v. Onan Corp., 192 F.3d 1137, 1140 (8th Cir. 1999); E.E.O.C. v. Francis W. Parker School, 41 F.3d 1073, 1076 (7th Cir.1994).
. If the effects of section 609.229 fall more harshly on blacks and people of color than whites, then there is a disparate impact regardless of whether people of color are overrepresented in the population of individuals convicted under the statute compared to their representation in the population of confirmed gang members.