Otha Wheeler, a former employee of Aventis Pharmaceuticals (“Aventis”), was terminated by the company. Wheeler sued the company claiming that she had been discriminated against because of her gender and race. Aventis responded that Wheeler’s violations of company policy justified the termination and moved for summary judgment. The district court1 granted Aventis’s summary-judgment motion2 and dismissed Wheeler’s claims. Wheeler appeals the denial of her race-discrimination claim. We affirm.
I. Facts
Wheeler worked for Aventis for approximately thirteen years. During the course of her employment, she had two notable adverse dealings with one of her co-workers, Linda Driver. First, during a peer-review session,3 Wheeler made unfavorable comments about Driver’s job performance. Wheeler also stated-to Driver-that Driver was not well liked by fellow employees. According to Wheeler, in retaliation for these statements, Driver threatened to report to management that Wheeler frequently grabbed male co-workers’ genitalia.
Subsequently, Driver lodged a complaint with Wheeler’s supervisor, Peggy Lewis, that Wheeler- — on various occasions— grabbed the crotch area of her male coworkers. Peggy Lewis met with an Aven-tis human-resources specialist, John Lewis (no relation), about Driver’s allegation of inappropriate touching. John Lewis instructed Peggy Lewis to immediately begin conducting an investigation. John Lewis also instructed Peggy Lewis to interview only those people directly involved with Driver’s allegations.4 Peggy Lewis interviewed approximately ten employees who reportedly witnessed — or experienced — Wheeler’s alleged grabbing. Peggy Lewis also interviewed employees whose names were given to her by previously-interviewed employees during the course of the investigation. However, Peggy Lewis did not re-interview Driver after her initial complaint about WTieeler’s behavior.
During the investigation, several male employees disclosed to Peggy Lewis that *856Wheeler had fondled them while they were working. Several of the men noted that they made concerted efforts to avoid Wheeler in order to protect themselves from the unwanted touching. Several other interviewees revealed that they witnessed Wheeler groping-or attempting to grope-various male employees. After Peggy Lewis completed the interviews, she again met with John Lewis to report her findings. After confirming Peggy Lewis’s findings, John Lewis met with Wheeler to discuss the accusations made against her. John Lewis explained to Wheeler that several of her co-workers had stated that, on numerous occasions, she had touched her male co-workers’ genitalia.
In response, Wheeler denied the specific allegations, but admitted that she — and her co-workers — engaged in various types of “horseplay,” which included discussing each others’ sex lives and sex toys during “dirty hour,”5 placing “kick me” signs on co-workers’ backs, squirting alcohol or other liquids at each other, and placing rubber snakes or spiders on the assembly line. Based upon its investigation, Aventis placed Wheeler on paid leave.
John Lewis then consulted with his supervisors about the appropriate course of action. Cheryl Flood, the Director of Associate Relations, made the decision to terminate Wheeler’s employment. According to Aventis, Wheeler was terminated because she violated the company’s sexual-harassment policy, which states:
The Company will not tolerate any form of harassment against associates, by its executives, managers, co-workers.... [Sjexual harassment also includes ... repeated offenses, excessive flirtations, advances or propositions, obscene or sexually oriented language or gestures ... and offensive physical contact such as grabbing, patting, pinching or brushing against another’s body. It refers to behavior which is not welcome, which is personally intimidating, hostile or offensive, which debilitates morale, and/or shall therefore interfere with their work effectiveness.
After Wheeler’s termination, the remaining employees in the department were verbally warned that no more horseplay would be tolerated.
Approximately one year after this warning issued, another worker in Wheeler’s former department informed Aventis that a fellow employee, Joe Don Harrell, had acted contrary to its policy. Specifically, Harrell had — at some point in the past-smeared grease on a female co-worker’s buttocks, and grabbed the breasts of two of his female co-workers. In response to this information, Peggy Lewis advised John Lewis of the allegations regarding Harrell. Aventis began an investigation into Harrell’s behavior that followed the same course as the Wheeler investigation. After substantiating the claim against Harrell, Aventis terminated his employment for inappropriate touching in violation of its sexual-harassment policy.
Also, approximately one year after Wheeler’s termination, an allegation was made that Toni Conrad, a white female in Wheeler’s department, would expose her breasts to others upon request. Aventis reprimanded Conrad but did not subject her to any other disciplinary action for the alleged exposure. According to Wheeler, Aventis’s delayed handling of the allegations against Harrell and Conrad reflect its unwillingness to investigate the “horse*857play” claim against her in a non-diserimi-natory manner.
II. Discussion
Our de novo review of summary judgments is well established. A moving party is entitled to summary judgment only if the pleadings, discovery, and affidavits show that there is no genuine issue of material fact and that the moving'party is entitled to judgment as a matter of law. Fed.R.Civ.P. 56(c). To avoid summary judgment, the nonmoving party must make a sufficient showing on every essential element of her case on which she bears the burden of proof. Osborn v. E.F. Hutton & Co., Inc., 853 F.2d 616, 618 (8th Cir.1988). However, in employment discrimination cases, because intent is inevitably the central issue, we apply the standard with caution. Gill v. Reorganized School Dist. R-6, Festus, Mo., 32 F.3d 376, 377 (8th Cir.1994).
Because Wheeler adduced no direct or circumstantial evidence of discrimination, her discrimination claim is subject to the burden-shifting method of proof first described in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-05, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973). Under this method, Wheeler must first establish a prima-facie case of racial discrimination, and if Aventis offers a legitimate, nondiscriminatory explanation for Wfheeler’s discharge, then Wdieeler must show that Aventis’s explanation is merely a pretext for racial discrimination. Id.
A.
AVheeler’s first task is to establish the minimal requirements for a primafacie case of race discrimination. To make a prima-facie showing of discrimination, Wdieeler must prove 1) that she is a member of a protected group; 2) that she was meeting the legitimate expectations of her employer; 3) that she suffered an adverse employment action; and 4) that circumstances exist which give rise to an inference of discrimination. Williams v. Ford Motor Co., 14 F.3d 1305, 1308 (8th Cir.1994). There is no dispute as to the first three elements. Thus, if Wheeler can show circumstances that give rise to an inference of discrimination, she has satisfied the first stage of the McDonnell Douglas analysis.
WTieeler can establish the remaining element by producing facts that similarly-situated employees, who are not black, were treated differently. We then consider, for purposes of evaluating Wdieeler’s prima-facie case, “whether the employees are involved in or accused of the same or similar conduct and are disciplined in different ways.” Id. at 1309. Here, Wheeler was accused of — and ultimately terminated for — inappropriately touching male employees’ sex organs'. WOieeler acknowledged that some touching took place but took issue with its characterization. She ' contended that the conduct was not unwelcome and should not be classified as anything but “horseplay” or having been done in jest. She argued that other employees often engaged in such “horseplay” at work and received little — or no — resulting discipline.
The test to determine if one is “similarly situated” varies at each stage of a McDonnell Douglas analysis. At the prima-facie stage it is “not onerous,” however at the third stage (proving pretext) it is “rigorous.” Ford Motor Co., 14 F.3d at 1308. Considering the low threshold at the prima-facie stage, we hold Wdieeler established each requisite element of her prima-facie case. The undisputed facts showed that several employees engaged in “horseplay” that was — at least arguably— *858comparable to Wheeler’s alleged actions, and Aventis imposed different sanctions.
B.
In turn, Aventis responded to Wheeler’s prima-facie case of discrimination with what it contended was a legitimate, nondiscriminatory explanation — that Wheeler was terminated based on her violation of its sexual-harassment policy. The investigating supervisors believed that Wheeler had physically grabbed — and attempted to grab — -her male co-workers’ genitalia on numerous occasions. We agree that an employee’s termination for violation of a sexual-harassment policy could be a legitimate, non-racially discriminatory basis.
C.
Thus, the inquiry now moves to the question of pretext — whether Wheeler produced sufficient evidence that the proffered reason was not the true reason for Aventis’s decision to terminate. Wheeler attempted to show pretext with evidence of inconsistent disciplinary action for (in her view) comparable conduct of other employees, evidence of racial animus, a “grossly inadequate investigation” into her allegation of “horseplay,” missing documents, and conflicting stories. Thus, the sole issue in this disparate treatment race-discrimination claim is whether any of this evidence creates a genuine issue of material fact in support of her claim that Aventis’s stated reason for terminating her was not true and that Aventis intentionally discriminated against her because of her race in violation of Title VII of the Civil Rights Act and the Missouri Human Rights Act.
Addressing Wheeler’s initial claim of pretext — discriminatory discipline — it must be repeated that the burden for establishing “similarly situated” at the pretext stage is rigorous. For discriminatory discipline claims, employees are similarly situated only when they are involved in or accused of the same offense and are disciplined in different ways. Harvey v. Anheuser-Busch, Inc., 38 F.3d 968, 972 (8th Cir.1994). First, Wheeler argues that her termination for grabbing co-workers’ genitalia was merely a pretext for discrimination because Conrad, a white employee, was not disciplined for a similar incident (exposing her breasts). While instances of disparate treatment can support a claim of pretext, Wheeler must prove that she and Conrad were similarly situated in all relevant respects. Lanear v. Safeway Grocery, 843 F.2d 298, 301 (8th Cir.1988). “ ‘Employees are similarly situated when they are involved in or accused of the same offense and are disciplined in different ways.’ ” Harvey, 38 F.3d at 972 (quoting Boner v. Board of Comm’rs, 674 F.2d 693, 697 (8th Cir.1982)) (emphasis added).
We find that Conrad and Wheeler were not similarly situated because their offenses differed substantially. Although they were both involved in sexual “horseplay,” their alleged actions involved differing levels of misconduct towards others. Taking the facts in the light most favorable to Wheeler, during their alleged “horseplay” Conrad exposed her breasts on request, while Wheeler grabbed male co-workers’ private area in a way that— according to the men interviewed- — -was offensive. Sexually-offensive conduct that involves physical contact is not the same as offensive comments, gestures, or lewd displays. Aventis was not obligated to treat the two behaviors as substantially similar because they involved objectively different conduct. Ward v. Procter & Gamble Paper Products Co., 111 F.3d 558, 560-61 (8th Cir.1997). The record reflects that the only other employee accused of touching co-workers’ genitalia received the same discipline as Wheeler — termination.
*859Next, Wheeler attempts to show pretext by presenting evidence that two of her co-workers expressed racial animus by stating that women and blacks at Aventis were “untouchable.” The implication of the comment was that the declarants reflected bias in their belief that blacks received preferential treatment from the company. However, these comments were not made by decision makers and — -because Wheeler is not alleging a racially-hostile work environment — are thus irrelevant. Aucutt v. Six Flags Over Mid-America, Inc., 85 F.3d 1311, 1315-16 (8th Cir.1996).
Wheeler next claims that Aventis inadequately investigated her complaints about inappropriate behavior on the assembly line. Specifically, she notes that no black employees were interviewed in the course of the investigation into her conduct, and that had Aventis been more diligent it would have uncovered Harrell’s breast-touching conduct a year sooner. To the contrary, Peggy Lewis claimed that she only conducted interviews that she considered relevant and explained that -she limited her interviews to possible victims or witnesses to the alleged touching, and others as names were revealed to her. Wheeler attempts to discredit Peggy Lewis’s rationale by noting that her accuser was not interviewed.
Wheeler’s argument misses the point. Peggy Lewis’s limited investigation was based upon a valid business judgment. We, therefore, are not permitted to review the logic behind the decision. Hudson v. McDonnell Douglas Corp., 63 F.3d 771, 781 (8th Cir.1995). Also, even if Aventis had been faster or more complete in questioning its employees (thereby perhaps discovering Harrell’s touching episode), it nonetheless does not follow that Wheeler was discriminated against on account of her race. At best, it proves that Aventis was derelict in expeditiously discovering sexual harassment in its workplace.
Finally, Wheeler argues that the fact that there was conflicting testimony regarding the grabbing episode (and the fact that Aventis failed to produce any of her peer reviews) might lead a jury to conclude that the touching never occurred. The district court assessed these arguments and found that none of the evidence cited by Wheeler produced a material-disputed fact.6 We agree. Such allegations, even if true, do nothing to establish that the reason stated by Aventis for the discharge was pretextual, Our inquiry is not whether Aventis’s decision was correct or wise but only “whether [the reported incident] was the real reason for [Wheeler’s] termination and not a pretext for [race] discrimination.” Gill, 32 F.3d at 379.
III. Conclusion
Wheeler failed to establish that Aven-tis’s proffered reason was pretextual. Wheeler’s proof did not put in dispute Aventis’s evidence that there was a factual basis for the several reports that Wheeler grabbed male co-workers’ genitalia and that Aventis believed that the reported conduct — -which violated its published sexual-harassment policy — was a proper basis for its decision to discharge Wheeler. Wheeler has not shown that she was treated differently than any other similarly-situated Aventis employee. Consequently, we-affirm the district court’s order granting summary judgment to Aventis.