LATINO OFFICERS ASSOCIATION CITY OF NEW YORK, INC., for itself and its members, Wilfred Maldonado, Anthony Miranda, also known as Ed, Manuel Nunez, Clifford Muniz, Adam Alvarez, Charles Castro, Louis Vega, Thaddeus Gamory, Michael Padilla, Carlos Jimenez, Jose Mercedes, Ruben Malave, Hilda Susu, Daniel Figueroa, and all others similarly situated, Hector Ariza, Manuel Gomez, Christopher Castro, Haydee Cartagena, Manuel Delgado, Parnell Peterson, Hiram Monserrate, individually and on behalf of a class of all others similarly situated, and Fernando Sanchez, also known as Freddy, also known as Selle, Plaintiffs-Appellants, v. The CITY OF NEW YORK, New York City Police Department, Rudolph W. Giuliani, as the Mayor of the City of New York and individually, Howard Safir, as the Commissioner of the Police Department of the City of New York and individually, Patrick Kelleher, as First Deputy Police Commissioner and individually, George Grasso, as Deputy Commissioner and individually, PATRICK Bradley, as Deputy Inspector and individually, Joseph Flynn, as Director and individually, Richard Kubick, as Special Prosecutor and individually, Charles V. Campisi, as Chief and individually, The Police Relief Fund, Inc., The New York City Police Department, Mayor Michael Bloomberg, and Police Commissioner Raymond W. Kelly, of New York City, Defendants-Appellees, 558 F.3d 159

JOSÉ A. CABRANES, Circuit Judge: Plaintiffs appeal from an October 26, 2007 order of the United States District Court for the Southern District of New York (Lewis A. Kaplan, Judge), denying their Motion for Contempt and/or to Compel Compliance with the District Court’s judgment of September 17, 2004. See Latino Officers Ass’n v. City of New York, 519 F.Supp.2d 438 (S.D.N.Y.2007).

In the District Court, plaintiffs submitted statistical data meant to show that within the New York City Police Department (“NYPD”), minority police officers were disciplined at a higher rate, and more severely, than were their white counterparts. Based on this data, plaintiffs argued that defendants were allowing discrimination to occur in violation of their settlement agreement and the District Court’s judgment entered pursuant thereto.

Because plaintiffs fail to show causation associated with the data they presented, and because plaintiffs offer no analysis regarding the statistical significance of their findings, we agree with the District Court’s conclusion that plaintiffs have not shown by clear and convincing evidence that defendants violated the September 17, 2004 judgment. Indeed, if anything, the record demonstrates that defendants have taken affirmative steps to combat employment discrimination. Accordingly, we affirm the order of the District Court.

BACKGROUND In September 1999, the Latino Officers Association and several Hispanic and African American police officers (collectively, “plaintiffs”) commenced an action against the City of New York, the NYPD, and several municipal officials (collectively, “defendants”), alleging systematic discrimination in the disciplinary system of the NYPD, in violation of Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. §§ 1981, 1983, 1985, the First and Fourteenth Amendments of the United States Constitution, the New York State Human Rights Law, the New York City Human Rights Law, and New York common law.

In August 2002, the District Court certified a class consisting of all Latino and African-American individuals who have been, are, or will be employed by the NYPD as uniformed officers, including civilians who perform the same employment functions as uniformed officers, who have been or will be subjected to discrimination on the basis of race, color or national origin in the form of a hostile work environment, disparate disciplinary treatment, and retaliation for the exercise of their rights. Latino Officers Ass’n. v. City of New York, 209 F.R.D. 79, 81, 93-94 (S.D.N.Y.2002).

In December 2003, the parties signed a Stipulation and Order, memorializing a negotiated settlement agreement. At the outset, the settlement provided that “[t]he New York City Police Department ..., as an equal opportunity employer, does not and will not allow discrimination based on actual or perceived race, color, national origin, ethnicity or any other reason prohibited by federal, state or local law.” J.A. at 327 (Stipulation and Order of December 18, 2003).

Under the heading “Affirmative Injunctive Relief,” id. at 328, as the District Court found, the settlement further provided, inter alia, that: the [NYPD] would [1] establish a “Disciplinary Review Unit” (“DRU”) to track and analyze whether minority members of the NYPD were being treated in a discriminatory manner when disciplined, [2] establish an “Advisory Committee” to address employment discrimination and retaliation concerns, [3] develop a “Know Your Rights” guide to the NYPD discipline system, and [4] enhance existing databases and create new databases to capture, and report to plaintiffs on a specified schedule, data thought to be relevant to analyzing whether or not discrimination was continuing in the NYPD discipline system.

Latino Officers Ass’n, 519 F.Supp.2d at 440-41 (footnotes omitted). On September 17, 2004, the District Court entered a judgment and order approving the negotiated settlement and incorporating all of its terms. See Latino Officers Ass’n. v. City of New York, No. 99 Civ. 9568, 2004 WL 2066605 (S.D.N.Y. Sept. 15, 2004). On December 6, 2006, plaintiffs moved for an Order of Contempt and/or to Compel Compliance with the Settlement and Order of September 17, 2004.

Specifically, plaintiffs alleged that defendants had failed to comply with the settlement agreement and the judgment entered pursuant thereto because, inter alia, (1) defendants had not established the Disciplinary Review Unit, as agreed, (2) defendants had not provided plaintiffs with the agreed-upon statistical data in a timely and complete fashion, and (3) discrimination had continued in the NYPD discipline system, as evidenced by statistical data proffered by plaintiffs. See Latino Officers Ass’n, 519 F.Supp.2d at 441.

In an October 26, 2007 Memorandum and Opinion, the District Court found that plaintiffs had failed to establish that defendants should be held in contempt for failing to comply with the settlement agreement, and it denied plaintiffs’ motion.

Specifically, with regard to plaintiffs’ allegations that discrimination had continued in the NYPD discipline system in violation of the settlement’s provision that the NYPD “does not and will not allow discrimination,” J.A. at 327, the District Court concluded that “[pjlaintiffs’ argument fails for one basic reason: defendants did not warrant that racial discrimination never again would occur in the NYPD discipline system.” Latino Officers Ass’n, 519 F.Supp.2d at 447.

The District Court reasoned: While the Agreement recites that the NYPD “will not allow discrimination based on actual or perceived race, color, national origin, ethnicity or any other reason prohibited by federal, state or local law,” it did not create a regime of strict liability. If plaintiffs believe that discrimination continues, notwithstanding the Agreement, they may seek redress through appropriate means, such as the federal and state laws that served as the basis for the underlying class action. The remedy, in the absence of proof that the NYPD is “allow[ing] discrimination,” is not contempt.

Plaintiffs’ showing falls far short of establishing that the NYPD is “allow[ing] discrimination,” even assuming that discriminatory behavior occurs from time to time, as it regrettably does in many parts of our society. Id. at 447-48 (footnote omitted). Accordingly, the District Court denied plaintiffs’ motion. Plaintiffs then filed a timely notice of appeal.

DISCUSSION Before this Court, plaintiffs appeal only the District Court’s denial of their motion with respect to defendants’ alleged failure to comply -with the settlement by “allowing” discrimination to occur within the NYPD discipline system. “Although the parties do not contest our jurisdiction” to entertain this appeal, “we are obliged to ascertain it independently.” Catskill Dev., L.L.C. v. Park Place Entm’t Corp., 547 F.3d 115, 123 (2d Cir.2008).

We conclude that we do have jurisdiction because where, as in this case, “civil contempt proceedings are instituted after the conclusion of the principal action rather than during the pendency of the action, the order disposing of the contempt proceedings is ap-pealable” as a final decision of a district court under 28 U.S.C. § 1291. United States v. O’Rourke, 943 F.2d 180, 186 (2d Cir.1991) (internal quotation marks omitted). This rule encompasses appeals from orders both granting and denying post-judgment civil contempt sanctions.

See Tranzact Technologies, Inc. v. 1Source Worldsite, 406 F.3d 851, 854 (7th Cir.2005); Sanders v. Monsanto Co., 574 F.2d 198, 199 (5th Cir.1978) (“[I]f a motion for civil contempt is denied after the entry of the judgment which was the subject of the contempt, the denial is final and reviewable because no further district court action is necessary to give life to the denial.”). See generally 15B Charles A. Wright & Arthur R. Miller, Federal Practice and Procedure § 3917 (3d ed. 2008) (“Complete disposition of contempt proceedings initiated to enforce a final judgment supports appeal.

Appeal can be taken from an order that denies civil contempt sanctions.” (citing cases)). We therefore proceed to address the merits of the appeal. As before the District Court, plaintiffs here rely on statistical data that show, inter alia, that minority officers were thirty-five percent more likely to receive informal discipline in the form of a “Command Discipline” than were white officers. Appellants’ Br. at 26.

Plaintiffs also argue that their data demonstrate that for the most recent period measured, minority members of the NYPD were “much more apt to be the subject of formal discipline,” including that they were more likely to be investigated, to be found guilty, and to receive harsher penalties when found guilty, than were their white counterparts. Id. Plaintiffs argue that in light of these facts, the District Court abused its discretion by not granting their motion for contempt and to compel compliance. See id. at 17.

For their part, defendants argue that the statistical evidence provided by plaintiffs is misleading in that, for example, some of the statistics measured only the number of disciplinary incidents, instead of people who had been disciplined, thereby distorting the results when a single officer had several disciplinary incidents. They argue that “where there is ample evidence of the defendants’ attempts to diligently comply [with the settlement], the Court did not abuse its discretion in finding that plaintiff[s] had failed to meet [their] burden here.” Ap-pellees’ Br. at 23.

We note that a contempt order is a “potent weapon,” Int’l Longshoremen’s Ass’n v. Phila. Marine Trade Ass’n, 389 U.S. 64, 76, 88 S.Ct. 201, 19 L.Ed.2d 236 (1967), that is inappropriate if “there is a fair ground of doubt as to the wrongfulness of the defendant’s conduct,” Cal. Artificial Stone Paving Co. v. Molitor, 113 U.S. 609, 618, 5 S.Ct. 618, 28 L.Ed. 1106 (1885). “A contempt order is warranted only where the moving party establishes by clear and convincing evidence that the alleged contemnor violated the district court’s edict.” King v. Allied Vision, Ltd., 65 F.3d 1051, 1058 (2d Cir.1995) (emphasis added).

Specifically, the “movant must establish that (1) the order the contemnor failed to comply with is clear and unambiguous, (2) the proof of noncompliance is clear and convincing, and (3) the contem-nor has not diligently attempted to comply in a reasonable manner.” Id. (emphasis added). Finally, when a district court’s ruling on a contempt motion is challenged on appeal, its

interpretation of the terms of the underlying order or judgment is subject to de novo review; its factual findings are accepted unless they are shown to be clearly erroneous; and its ultimate ruling on the contempt motion is reviewed for abuse of discretion. See Dunn v. N.Y. State Dep’t of Labor, 47 F.3d 485, 490 (2d Cir.1995); United States v. Local 1804-1, Int’l Longshoremen’s Ass’n, 44 F.3d 1091, 1095-96 (2d Cir.1995); see also Perez v. Danbury Hosp., 347 F.3d 419, 423 (2d Cir.2003).

Upon review, we conclude that the District Court did not abuse its discretion in determining that plaintiffs failed to prove that defendants should be held in contempt for violating the Settlement and Order. At the outset, we underscore that it is the moving party — in this case the plaintiffs — who bears the burden of establishing the three factors set forth in King v. Allied Vision, Ltd., 65 F.3d at 1058.

Accordingly, the relevant inquiry is not whether defendants adequately rebutted plaintiffs’ evidence, but rather, whether plaintiffs’ evidence alone was adequate to establish that (1) the order was clear and unambiguous, (2) defendants did not, in fact, comply with the order (and the evidence of the noncompliance must be clear and convincing), and (3) defendants did not diligently attempt to comply with the order in a reasonable manner. Id.

Beginning with the second King factor — whether plaintiffs’ proof of noncompliance is clear and convincing — we note that the Supreme Court has cautioned that “statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, they may be rebutted. In short, their usefulness depends on all of the surrounding facts and circumstances.” International Broth. of Teamsters v. United States, 431 U.S. 324, 339, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). See also Radue v.

Kimberly-Clark Corp., 219 F.3d 612, 616 (7th Cir.2000) (“A more basic problem is that statistics can only show a relationship between an employer’s decisions and the affected employees’ traits; they do not show causation.”). Without more information to support causation, and lacking any analysis from the plaintiffs’ expert regarding the statistical significance of the disparities he identified, we are in complete agreement with the District Court that “[pllaintiffs’ showing falls far short of establishing that the NYPD is ‘allow[ing] discrimination,’ ” Latino Officers Assn., 519 F.Supp.2d at 448.

Indeed, if anything, the record shows that defendants have taken substantial steps to eliminate discriminatory practices in the NYPD. The District Court found that, pursuant to the settlement agreement, defendants had (1) established a Disciplinary Review Unit, titled the Employment Practices Unit, to review the NYPD’s disciplinary process and how it affects employment discrimination, (2) developed a “Know Your Rights” guide, which details the NYPD disciplinary process, and (3) produced statistical reports on the NYPD discipline system.

Beyond the terms of the settlement agreement, the record shows that defendants had decided to conduct a review of at least one command each month to assess disciplinary actions taken and use information from their databases to train commanding officers about discipline trends.

In the absence of clear and convincing evidence of defendants’ noncompliance, and in light of the fact that defendants have taken these affirmative steps to curb employment discrimination in the NYPD — including the very steps plaintiffs deemed necessary to combat such discrimination — we conclude that plaintiffs have failed to meet their burden of showing why defendants should be held in contempt. Finally, we note that our holding here does not preclude plaintiffs from pressing their claims through other means.

As the District Court noted, plaintiffs may still bring an action relying on the federal and state laws that served as a basis for the underlying class action. CONCLUSION For the reasons stated above and substantially for the reasons stated by Judge Kaplan in his careful Memorandum and Opinion, we affirm the October 26, 2007 order of the District Court. .

Despite the fact that defendants had created an "Employment Practices Unit,” which was to serve the same purpose as the "Disciplinary Review Unit” as set forth in the settlement agreement, plaintiffs argued that the difference in nomenclature violated the agreement.

In the Memorandum and Opinion of October 26, 2007, the District Court concluded that "[t]he provision [of the settlement agreement] does not clearly and unambiguously require defendants to name the unit in any specific manner.” Latino Officers Ass’n, 519 F.Supp.2d at 444. .The expert's failure to assess the significance of the disparities he identified in the 2005 disciplinary data is troubling in light of the fact that several of those disparities were eliminated or notably reduced in the 2006 data presented to the District Court.

For example, plaintiffs’ expert reported that in 2005, minority officers were 84% more likely (30.8% versus 16.7%.) to receive "charges filed" dispositions of their disciplinary proceedings, considered a severe sanction. In 2006, however, that difference disappeared: 23.9% of white officers received a "charges filed” disposition, as did 23.2% of minority officers. Similarly, according to the expert, in 2005, minority officers received the "relatively more favorable” guilty-by-plea dispositions 30% less often than whites (40.2% versus 57.3%). In 2006, that discrepancy was reduced to 12% (40.9% versus 46.3%).

To be sure, not all of the disparities identified by the plaintiffs’ expert in the 2005 data were reduced or eliminated in the 2006 data. However, in the absence of analysis indicating which of the disparities he identified, if any, are statistically significant or not attributable to other factors, the notable year-to-year variation in the data undermines plaintiffs' argument on appeal that the District Court erred in concluding that plaintiffs’ statistical analysis was insufficient to support their contempt motion. . Defendants concede that the initial reports were not issued in a timely fashion.

However, plaintiffs do not raise on appeal the issue of noncompliance with the specific reporting deadlines set forth in the settlement agreement. . Because plaintiffs have failed to establish the second King factor, see King v. Allied Vision, Ltd., 65 F.3d at 1058, we do not reach the other King factors.