BACKGROUND

In the early 1990’s, the New York City School Construction Authority (“the SCA”) hired Charleston Colin Albert, Whitfield Chandler, Johnnie B. Harris, James Keller and Everton Mack (collectively “the plaintiffs”), as either First or Second level Project Officers (“PO-I” or “PO-II”). As PO-II’s, the plaintiffs reported to a Senior Project Officer (“SPO”) on the day-to-day activities of public school renovations and repairs.

In 1998, the plaintiffs, all African-American males, filed a civil rights action in the District Court for the Eastern District of New York pursuant to 42 U.S.C. § 1981, 42 U.S.C. § 2000e, et seq. (“Title VII”), the N.Y. Human Rights Law, Exec. Law § 290, et seq., and N.Y.C. Admin. Code 8-101, et seq.

The plaintiffs claim that, for race-related reasons: (1) the SCA failed to promote them from either PO-I’s to PO-II’s or from PO-II’s to SPO’s; (2) there were salary disparities between black and non-black PO-II’s; and (3) the SCA retaliated against them for filing administrative complaints with the SCA’s Equal Employment Opportunity (“EEO”) counselor.

The plaintiffs’ issues were divided into treatment and impact claims and tried concurrently before a jury and the district court. The claims for disparate treatment *309in promotions, disparate treatment in pay, and retaliation, were submitted to a jury and the disparate impact claims were tried by Judge Wolle in a bench trial.

After a twelve-day trial and almost two days of deliberations, the jury returned a verdict for the SCA on all of plaintiffs’ claims. Subsequently, Judge Wolle delivered a verdict for the SCA dismissing plaintiffs’ disparate impact claims. Plaintiffs appeal solely from the jury verdict, claiming erroneous evidentiary rulings and jury instructions.

DISCUSSION

A. Evidentiary Rulings

Before trial, the SCA moved in limine to exclude alleged evidence of discrimination pre-dating the statute of limitations on plaintiffs’ claims. The SCA asked the court to exclude: (1) a “preliminary sketch” (or “report”) prepared by Noel Poyser, the SCA’s EEO officer on alleged statistical disparities in promotion; (2) Poyser’s lay testimony about the report; and (3) employment applications from three white employees to show that the SCA’s salary enhancement guidelines were applied in a discriminatory fashion. The plaintiffs opposed SCA’s motion and asked the court to admit Poyser’s report, his related testimony, and the salary documents as “substantive evidence of discrimination” or as “background evidence.”

The court reserved decision on the motion to exclude Poyser’s report, and revisited the issue several more times before ruling that the report could “be referred to, but not presented [and] anything in it that preceded [the statute of limitations] will not be talked about to the jury.” The court ultimately determined that the report was more prejudicial than probative. Thus, the jury was permitted to consider Poyser’s testimony about the report only “for the limited purpose of proving the timing and background of when the plaintiffs made complaints about racial discrimination and what SCA said and did in response.”

For the reasons stated below, the district court’s evidentiary rulings were not an abuse of discretion.

Federal Rule of Evidence 403 provides that:

Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence.

Rule 403 rulings “are entitled to considerable deference and will not be overturned absent a clear abuse of discretion.” Costantino v. Herzog, 203 F.3d 164, 173 (2d Cir. 2000).

An erroneous evidentiary ruling is reversible only if it affects “substantial rights,” as “when ... a district court excludes a party’s primary evidence in support of a material fact, and failure to prove that fact defeats the party’s claim.” Schering Corp. v. Pfizer Inc., 189 F.3d 218, 224 (2d Cir. 1999).

Poyser’s report — which contained incomplete data on the race of applicants for promotions and lacked adequate comparisons of applicants to promotees based on nondiscriminatory criteria — had limited probative value for plaintiffs’ intentional discrimination claims. (The court admitted, but ultimately rejected, the report at the bench trial on plaintiffs’ disparate impact claims.) Furthermore, four of the fifteen positions and three of the years analyzed in the report concerned a time outside the statute of limitations, and the report’s conclusions were based on an analysis of all fifteen positions, not just the *310eleven within the limitations period. The report’s information outside the limitations period was also cumulative.

Given the low probative value and high potential for confusion and prejudice, the court’s exclusion of the report was not an abuse of discretion.

While evidence of discrimination predating the statute of limitations may sometimes be admissible as background evidence, Fitzgerald v. Henderson, 251 F.3d 345, 365 (2d Cir. 2001), the district court retains discretion to determine when this is admissible, Malarkey v. Texaco, Inc. 983 F.2d 1204, 1211 (2d Cir. 1993).

Regarding the forty-four SPO and PO-II positions challenged by plaintiffs, there was ample background information. The court was well within its discretion to exclude the report given the prejudicial impact and its tendency to confuse the jury, compared to the report’s low probative value.

We do not disagree that evidence of conditions and events prior to the enactment, application, or limitations period of Title VTI may be admissible to show that current salary disparities have been perpetuated. See generally, Bazemore v. Friday, 478 U.S. 385, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986); Sobel v. Yeshiva Univ., 839 F.2d 18 (2d Cir. 1988). However, the court’s exclusion of the salary documents in this case was proper for two reasons. First, the plaintiffs’ questions about the documents were directed to a witness who lacked personal knowledge of the criteria for salary enhancements at the relevant time. (The court also noted that one of the documents was partially illegible.) Second, and more importantly, there is no evidence in the record that the same salary guidelines within the limitations were in effect between 1989 and 1991, when the employment applications were generated.

B. Jury Instructions

The plaintiffs contend that the district court erroneously instructed the jury on several issues, including the standard of proof, the use of statistical evidence, and plaintiffs’ salary and promotion claims. For the reasons stated below, we find no reversible errors in any of the jury instructions.

This Court reviews jury instructions de novo. See, e.g., LNC Invest., Inc. v. First Fidelity Bank, N.A., 173 F.3d 454, 460 (2d Cir. 1999). “A jury instruction is erroneous if it misleads the jury as to the correct legal standard or does not adequately inform the jury on the law.” Id. However, a trial court has considerable discretion in the formulation and style of jury instructions. See Parker v. Sony Pictures Entm’t, Inc., 260 F.3d 100, 106 (2d Cir. 2001).

An erroneous jury instruction is harmless if it is clear that it did not influence the jury’s verdict. LNC Invs., 173 F.3d at 462.

1. Standard of Proof

The district court’s instruction that “preponderance of the evidence is evidence that is more convincing than opposing evidence” was not reversible error.

In truth, this Court has advised on more than one occasion that “ ‘[i]n an ordinary civil suit ... words like ‘convince’ ... should be shunned in a jury charge.’” Ostrowski v. Atl. Mut. Ins. Cos., 968 F.2d 171, 187 (2d Cir. 1992) (endorsing scale-tipping language of model federal jury instructions) (citation omitted). We reaffirm this Court’s caveat, but note that other than the solitary word “convincing,” the charge here bears little resemblance to that criticized in Ostrowski.

*311In Ostrowski, this Court was specifically concerned about use of the word “convincing” in conjunction with “confidently” because “to tell the jury that it must be able to say ‘confidently’ that the evidence is in the plaintiffs favor seems quite likely to require more than a preponderance.” Id. Here, the district court’s lucid instruction contained no “confidence” language, and the phrase “more than the opposing evidence” conveyed the substance of the scale-tipping instruction requested by the plaintiffs below.

Therefore, despite the potential implications of “convincing” in jury instructions, there was no error on these facts, nor any conceivable prejudice to the plaintiffs.

2. The Use of Statistics

Nor was the court’s instruction on the probative value of comparative salary charts prepared by plaintiffs’ counsel’s paralegal reversible error.

The court instructed the jury that for plaintiffs’ statistical evidence “to be probative of intentional discriminatory intent” on the promotion claims, it “must compare the relevant persons filling the position with the qualified population of persons seeking to be hired or promoted ... including the relevant labor market.” As for the salary disparity claim, the court instructed that “the statistical evidence must compare the salary history and related-back provisions of the persons whose salaries are being challenged with the salary and related background history of a plaintiff.”

The district court’s instruction is not materially different from a similar charge affirmed by this Court. See Luciano v. Olsten Corp., 110 F.3d 210, 218 (2d Cir. 1997) (affirming instruction that “statistics and numbers offered must be compared with a number of factors, including among other factors, the relevant labor market, including the qualified applicant pool, the relative numbers of men and women who were both qualified for and interested in management jobs with the [defendant] company”); cf. Smith v. Xerox Corp., 196 F.3d 358, 370 (2d Cir. 1999) (because “a disparate treatment claim looks at how an individual was treated compared to her similarly situated coworkers .... statistical analyses that compare coworkers who competed directly against each other ... are appropriate”). Therefore, the charge was proper.

Even if we accept plaintiffs’ argument that the district court should have instructed the jury that their statistical evidence had probative value absent the aforementioned analysis and comparison, any defect would have been rectified by the court’s further instruction that “[i]n determining the usefulness of the statistical evidence, you must consider all the surrounding facts and circumstances,” including logic, common sense, accuracy, accompanying expert testimony, and whether the statistics reflect nondiseriminatory reasons for different treatment.

3. Discrimination Claims

a. Salary Claims

We find no error in the district court’s salary claim instruction that each plaintiff must show that “[h]e was performing his job at a level which satisfied the employer’s reasonable expectations.”

Plaintiffs claim that under Belfi v. Prendergast, 191 F.3d 129 (2d Cir. 1999), satisfying an employer’s reasonable expectations is an unauthorized element of a salary disparity case. This is incorrect.

We note, as a preliminary matter, that Belfi is not dispositive because there is no “rigid” or “mechanized” formula to articulate the elements of a discrimination claim. *312See Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978). Further, in Belfi itself, one of the Court’s articulations of the salary disparity claim included the element that the plaintiff “was qualified for the job in question,” 191 F.3d at 140, which is functionally equivalent to “satisfying the employer’s reasonable expectations.”

Nor are we limited to Belfi in determining whether the district court’s articulation was an abuse of discretion. In non-salary cases, this Court has included in a plaintiffs initial burden the element of “satisfactory job performance,” Farias v. Instructional Sys., Inc., 259 F.3d 91, 98 (2d Cir. 2001), and “whether an employee’s performance meets his employer’s legitimate expectations,” Meiri v. Dacon, 759 F.2d 989, 995 (2d Cir. 1985) (citation and quotations omitted). These articulations are even closer to the district court’s articulation in this case.

Accordingly, the court’s use of the same charge for the salary and promotion claims to avoid jury confusion was not an abuse of discretion.

b. The Promotion Claims

In contrast to the above evidentiary exclusions and jury instructions, the district court’s instruction on the plaintiffs’ promotion claim was erroneous, but the error was harmless.

The court instructed that a plaintiff must “show that after [he] was rejected for promotion, the position remained open and defendant continued to interview persons with the same qualifications as plaintiff until defendant filled the job with a white person.” The plaintiffs requested that the court substitute “nonAfrican-American” for “white person” on the theory that the SCA may have favored Hispanics. The court overruled their objection reasoning that the “entire thrust of the plaintiffs’ case has been to distinguish between black and white.”

As a general proposition, the instruction was contrary to federal discrimination law. See, e.g., de la Cruz v. N.Y.C. Human Res. Admin., 82 F.3d 16, 20 (2d Cir. 1996) (Puerto Rican established prima facie case of racial discrimination where job was awarded to African-American female).

However, we are convinced that the error had no influence on the jury verdict for the following reasons. First, the district court’s attachment to the verdict sheet of the names and occupants of all the relevant positions, including those held by the five Hispanic employees, mitigated the defective instruction. Second, only two of the forty-four contested positions at issue were awarded to Hispanics. Third, the verdict sheet records the jury’s specific finding that none of the SCA’s employment decisions were racially motivated. Thus, the effect of error, if any, would have been de minimis. Finally, we also decline to remand for a limited retrial based on the plaintiffs’ technical, but unemphasized, challenges at trial to the five positions held by Hispanics. There is no evidence in the record for a rational juror to find that the SCA intentionally discriminated against the four African-American plaintiffs who applied for the two challenged positions held by Hispanics.

CONCLUSION

For the reasons set forth above, the judgment is AFFIRMED.