Da Silva Moore v. Publicis Groupe & MSL Group (868 F. Supp. 2d 137, 2012)

OPINION AND ORDER

ANDREW J. PECK, United States Magistrate Judge:

Plaintiffs’ “Motion for Recusal or Disqualification” (Dkt. No. 169) is based not on any claim that the Court has an actual bias, but rather on “an appearance of partiality.” (Dkt. No. 192: Pis. Reply Br. at 1 n. 1: “Plaintiffs have never accused Judge Peck of actual bias or sought to impugn Judge Peck’s integrity. Plaintiffs’ only ground for recusal is that the facts taken together create an appearance of partiality.”) Plaintiffs’ recusal motion is DENIED.

FACTUAL BACKGROUND

The main ground of plaintiffs’ motion is that my support for predictive coding showed bias favoring MSL and coerced plaintiffs into assenting to the concept of predictive coding. The chronology of events in this case puts the lie to plaintiffs’ claim.

This case was referred to me on November 28, 2011. (Dkt. No. 48.) Well before that date, MSL had informed plaintiffs that it “proposes using keywords as well as the analytical tools available in the Axcelerate review platform, which includes predictive coding....” (Dkt. No. 178: Anders Aff. Ex. 2: Anders 10/21/11 Letter to Wipper at 3; see also Anders Aff. Ex. 3: Anders 11/3/11 Letter to Wipper at 4-9.) Plaintiffs requested that MSL fully disclose its proposed predictive coding methodology for plaintiffs’ consideration. (See Dkt. No. 124: Nurhussein 3/19/12 Aff. Ex. E: Wipper 11/9/11 Letter to Anders at 5; see also id. Ex. D: Wipper 10/25/11 Letter to Anders at 2.)

December 2, 2011 Conference

I held my first conference with the parties on December 2, 2011. (Dkt. No. 51: 12/2/11 Conf. Tr.) During that conference, MSL’s counsel stated that an open issue was “plaintiffs reluctance to utilize predictive coding to try to cull down” approximately three million electronic documents from the agreed-upon custodians. (12/2/11 Conf. Tr. at 7-8.) Because of my Search, Forward article that I would call to the parties’ attention at the conference, I stated that: “You must have thought you died and went to Heaven when this was referred to me,” to which MSL’s counsel responded: “Pm just thankful that, you know, we have a person familiar with the predictive coding concept.” (12/2/11 Conf. Tr. at 8-9.) Plaintiffs did not move to recuse me at that time for making this comment. To the contrary, plaintiffs’ counsel clarified that MSL had “over simplified [plaintiffs’] stance on predictive coding,” i.e., that they were not opposed but had “multiple concerns ... on the way in which [MSL] plan to employ predictive coding” and plaintiffs wanted “clarification.” (12/2/11 Conf. Tr. at 21.)

The Court did not rule on any predictive coding issue but offered the parties the following advice:

Now, if you want any more advice, for better or for worse on the ESI plan and whether predictive coding should be used, or anything else ... I will say right now, what should not be a surprise, I wrote an article in the October Law Technology News called Search Forward, ivhich says predictive coding should be used in the appropriate case.

Is this the appropriate case for it ? You all talk about it some more. And if you can’t figure it out, you are going to get back in front of me. Keywords, certainly unless they are well done and tested, are not overly useful. Key words along with predictive coding and other methodology, can be very instructive.

I’m also saying to the defendants who may, from the comment before, have read my article. If you do predictive coding, you are going to have to give your seed set, including the seed documents marked as nonresponsive to the plaintiffs counsel so they can say, well, of course you are not getting any [relevant] documents, you’re not appropriately training the computer.

(12/2/11 Conf. Tr. at 20-21, emphasis added.) The conference adjourned with the parties agreeing to further discuss the ESI protocol. (12/2/11 Conf. Tr. at 34-35.)

My October 2011 Search, Forward Article

As noted, at the December 2, 2011 conference I alerted the parties to my article, Search, Forward: Will manual document review and keyword searches be replaced by computer assisted coding?, appearing in the October 2011 issue of Law Technology News. See Andrew Peck, Search, Forward, L. Tech. News, Oct. 2011, at 25-26, 29. The article reviewed the problems with manual review and keyword searches (if poorly done), and generally described how computer-assisted review, a/k/a predictive coding, worked. See Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, - F.R.D. -,---, 2012 WL 607412 at *2-3 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.) (quoting Search, Forward, L. Tech. News, Oct. 2011, at 29), adopted, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012). I ended the article with the comment that lawyers could view the Search, Forward article “as a sign of judicial approval” of predictive coding, but only for appropriate cases, stating: Andrew Peck, Search, Forward, L. Tech. News, Oct. 2011, at 29. Having had this article and my general opinion about predictive coding brought to the parties’ attention at the December 2 conference, plaintiffs did not move for my recusal.

In my opinion, computer-assisted coding should be used on those cases where it will help “secure the just, speedy, and inexpensive” (Fed.R.Civ.P. 1) determination of cases in our e-discovery world.

January 4, 2012 Conference

Plaintiffs’ counsel wrote asking the Court to postpone the next scheduled conference, stating that “[although plaintiffs are prepared to consider the use of predictive coding as a search method in general, Plaintiffs need more time to evaluate and provide feedback on [MSL’s] draft proposal and its methodology.” (Dkt. No. 178: Anders Aff. Ex. 5: Wipper 12/19/11 Letter to Court, at 2.) Plaintiffs said that after their consultant (DOAR Litigation Consulting) examined MSL’s proposal, the “parties then may engage in dialogue in order to address the issues, if any, that require resolution by the Court.” (Id.) The Court granted the extension (over MSL’s objection), and reminded the parties to read my article: “As to predicative coding, you should read my article, ‘Search, Forward’ in the Oct. 2011 issue of Law Technology News.” (Dkt. No. 58: 12/20/11 Memo Endorsed Order.)

On January 3, 2012, in anticipation of the January 4, 2012 conference, plaintiffs submitted a letter to the Court containing their discovery proposals, including an ESI protocol utilizing predictive coding. (Anders Aff. Ex. 6: Wipper 1/3/12 Letter to Court.) Plaintiffs informed the Court that they “have attempted to work within Defendants’ proposed methodology while honoring their restrictions” but that “Plaintiffs believe that Defendants’ proposed use of predictive coding while arbitrarily imposing a cap [on cost] runs counter to the reasoning behind the use of predictive coding.” (Id. at 2.) Plaintiffs’ proposed protocol stated that “the following is a summary of the Parties’ agreement on the use of Predictive Coding” for search of MSL’s email archive. (Anders Aff. Ex. 6: 1/3/12 Wipper Letter to Court, attached “Plaintiffs Proposed Protocol Relating to the Production of Electronically Stored Information (‘ESI’)” at 12.) Plaintiffs’ Proposed Protocol, in a section entitled “General Overview of Predictive Coding Process,” described that MSL would use Recommind’s Axcelerator software and develop a “seed set” to train the computer, it would be a transparent process, and it would include quality testing. (Pis. Proposed Protocol at 12-20.) In other words, after discussions with its consultant DOAR, plaintiffs were proposing using predictive coding, but with certain differences in the details compared to MSL’s proposal.

At the January 4, 2012 conference, after addressing other discovery issues, the parties turned to the email search issue. The following colloquy occurred:

MS. WIPPER [Plaintiffs’ Counsel]: ... I’d like DOAR to respond and give you an overview, if we may, of our proposal on predictive coding.

THE COURT: All light, though I guess I’d like to know where it differs [from MSL’s].

MS. WIPPER: Well, its actually a direct response to their proposal.

MR. NEALE [CEO of DOAR, Plaintiffs Consultant]: ... We have not taken issue with the use of predictive coding, or frankly, with the confidence levels that they have proposed....

(1/4/12 Conf. Tr. at 51, emphasis added.) The parties spent time conferring with each other while the Court handled another case, and later in the afternoon reported that they had made “a lot of progress.” (1/4/12 Conf. Tr. at 53-55.) In discussing what needed to be done and when the next conference would be, MSL counsel Anders noted he would be on vacation, and the Court suggested that another lawyer such as Ms. Chavey cover in his absence so as not to lose time. (1/4/12 Conf. Tr. at GO-61.) The Court also noted that while he might not be working on the case, Anders could involve his firm’s Florida ediscovery counsel, Ralph Losey, who I stated I knew “very well.” (1/4/12 Conf. Tr. at 61.) Counsel then suggested a date for the next conference, and the Court noted that “That’s LegalTech week,” but that February 2 would work. (1/4/12 Conf. Tr. at 66.) Counsel did not ask what I would be doing at LegalTech.

Legal Tech

I was a speaker at seven panels in connection with LegalTech on January 30, 31 and February 1, 2012. LegalTech is run by ALM (the company that publishes The American Lawyer, Law Technology News and other publications). LegalTech New York 2012 had thirty-nine sponsors, including Recommind, and 186 exhibitors, not to mention hundreds of paying attendees. I did not speak at any panels sponsored by Recommind, nor did I receive any expense reimbursement or compensation from Recommind during LegalTech 2012 nor at any other conference. The panels in which I participated at LegalTech 2012 involved general ediscovery issues (such as cooperation, preservation, proportionality, etc.), effective keyword searching, and predictive coding. My participation in these educational panels (for which attendees received CLE credit), and that of the other panelists, was arranged before this case was referred to me.

Another panelist on two of the panels was Ralph Losey, national ediscovery counsel in the Florida office of Jackson Lewis, MSL’s counsel. (Dkt. No. 178: Anders Aff. Ex. 10: Losey Aff ¶ 8.) Losey’s affidavit makes clear that we have never had any ex parte communication about this lawsuit. Losey stated:

I have never spoken with U.S. Magistrate Judge Andrew J. Peck about this case nor had any other ex parte communication with him about this case.

(Anders Aff. Ex. 10: Losey Aff. ¶ 6.) I confirm that. {See also Dkt. No. 158: 4/2/12 Order.) Those panels (and my preparation for them) involved only the subject of computer-assisted review in general terms and in comparison to other search techniques (like manual review and keywords). In other words, the level of the panel discussion was similar to that of my Search, Forward article. There was absolutely no discussion of the details of the predictive coding protocol in this case, or in what a predictive coding protocol should look like in general. While plaintiffs’ recusal motion makes it sound like I spoke at length about this case on these panels, that simply is not true. Typical of my minor reference during panels to this case (without mentioning the case name, at a time when the case was not yet all over the ediscovery blogs) is this passing reference:

I’ve actually only seen computer-assisted review being utilized in one of my cases. It is still ongoing at the moment. There is more or less agreement that they’re willing to go that route, although [there is] still lots of discussion about exactly what sources are going to be input and other things that are still in front of me — so I’m not going to say much more about it. The only thing I will say is when the District Judge referred the matter to me and I saw from the letters that had been submitted to the District Judge that the defendant was particularly pushing using computer-assisted review, I couldn’t resist and the first thing I said to them when they walked into my courtroom, was saying to the defendants, “Boy, you must have thought you died and went to heaven when this discovery matter got referred to me in light of my article on the subject.”

(Dkt. No. 171: Wittels Aff. Ex. B: Excerpt of a Video Recording of the Judicial Perspectives on Technology-Assisted Review Panel at LegalTech N.Y.2012.)

February 8, 2012 Conference and DOAR’s Press Release

The Court’s next conference in this case was on February 8, 2012. (Dkt. No. 97: 2/8/12 Conf. Tr.) In advance of the conference, plaintiffs and MSL wrote to the Court enclosing their proposed ESI protocols and highlighting their areas of difference. (See Dkt. No. 124: Nurhussein 3/19/12 Aff. Ex. H: Wipper 1/25/12 Letter to Court; Dkt. No. 171: Wittels Aff. Ex. Z: Anders 1/25/12 Letter to Court.) Plaintiffs’ proposed protocol, like MSL’s, used predictive coding. (Nurhussein 3/19/12 Aff. Ex. H: Wipper 1/25/12 Letter to Court at 3-5 & Proposed Protocol at 20-35.) Plaintiffs, however, proposed training the computer with two rounds of iterative review of 16,555 documents reviewed each round, while MSL’s proposal was for up to seven rounds of review of at least 500 documents per round. (See Nurhussein 3/19/12 Aff. Ex. H: Wipper 1/25/12 Letter to Court, Proposed Protocol at 31-34; Wittels Aff. Ex. Z: Anders 1/25/12 Letter to Court at 6-7; 2/8/12 Conf. Tr. at 70-89.) The Court heard argument on these issues from the parties’ counsel and their ediscovery vendors at the February 8, 2012 conference. (See 2/8/12 Conf. Tr.)

The Court’s rulings at the February 8 conference are summarized in my February 24, 2012 opinion, and will not be repeated here. Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, — F.R.D. -, - — -, 2012 WL 607412 at *3-6 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.), adopted, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012). Suffice it to say that the Court was comparing plaintiffs’ predictive coding protocol and defendant MSL’s predictive coding protocol. (2/8/12 Conf. Tr. at 57-89.) See Da Silva Moore v. Publicis Groupe, — F.R.D. at---, 2012 WL 607412 at *5-6. The Court accepted MSL’s predictive coding protocol, with seven iterative review rounds, but with the caveat that if the system was not stabilized at that point, the parties would present quality control verification information to the Court and further rounds would be ordered to stabilize the system. (2/8/12 Conf. Tr. at 76-77, 83-84, 88.) See Da Silva Moore v. Publicis Groupe, — F.R.D. at -, 2012 WL 607412 at *6.

Moreover, plaintiffs’ consultant, DOAR’s CEO Paul Neale, agreed that, in general, computer-assisted review works and works better than most alternatives. (2/8/12 Conf. Tr. at 76.) Neale noted that “it is fair to say we are big proponents of it.” (2/8/12 Conf. Tr. at 76.) Neale added, however, that “[t]his is new technology and it has to be proven out” and that plaintiffs “don’t at this point agree that this is going to work.” (2/8/12 Conf. Tr. at 75.)

A few days after the February 8 conference, plaintiffs ’ consultant DOAR issued a press release about my February 8 oral ruling and touted DOAR’s role in this “ ‘important step in the wider adoption of predictive coding technologies.’ ” (Dkt. No. 178: Anders Aff. Ex. 8: 2/13/12 DOAR Press Release.) It is worth quoting DOAR’s press release in full:

DOAR Experts Advise Plaintiffs in Hearing Before Magistrate Judge Peck on the Use of Predictive Coding

First Federal Judicial Acceptance on Predictive Coding Limits

For Immediate Release

Monday, February 13, 2012 New York, NY — In the first federal case where Predictive Coding use will be adopted by both the court and the parties, DOAR Litigation Consulting CEO, Paul Neale, and Vice President of Discovery Consulting, Gene Klimov, advised Plaintiffs’ counsel on the development of an electronically stored information (“ESI”) protocol that includes the use of predictive coding. DOAR participated in the negotiations with defense counsel and in the drafting of a comprehensive ESI protocol in da Silva Moore, et al. v. Publicis Groupe SA et al, which will be jointly submitted by the parties. Furthermore, DOAR evaluated the defendants’ proposed use of predictive coding and ensured that the process will be transparent enough to allow plaintiffs to verify the reliability of the process. Plaintiffs’ counsel will participate in every stage of the review of the documents that will seed, train and verify the process.

Mr. Neale presented the plaintiffs’ position to the court during a hearing on February 8, 2012 before Magistrate Judge Andrew Peck of the Southern District of New York. The parties plan to submit a final version of the joint protocol on February 17, 2012 after which it is expected that Judge Peck will issue a written opinion.

“While we most often advise producing parties on the use of alternative technologies such as predictive coding, I believe that our support of the requesting party in this case will prove to be an important step in the wider adoption of predictive coding technologies,” Mr. Neale states. “Gene and I are honored to be working with the lawyers at Sanford Wittels & Heisler on such an important precedent-setting case in the area of Electronically Stored Information (ESI) and the use of predictive coding.”

Predictive coding (or computer assisted review) is a rapidly evolving technology that provides parties in litigation with an alternative to the time and cost associated with the traditional, manual review of large volumes of documents. However, while the use of predictive coding is growing, its reliability and defensibility have yet to be fully explored by the courts.

da Silva Moore, et al. v. Publicis Groupe SA et al, 11-CV-1279, is being heard by Judge Peck in the U.S. District Court, Southern District of New York.

(Anders Aff. Ex. 8: DOAR 2/13/12 Press Release.)

The Court notes that even at the May 14, 2012 conference, DOAR’s Neale repeated under oath his company’s approval of predictive coding generally, stating:

Well, your Honor, as you know, we [DOAR] were retained to advise the plaintiffs on evaluation and use and defendants’ use of predictive coding. And again, while we as a company are proponents of it, we have been evaluating the steps and the process and have been trying to advise the plaintiffs and the Court accordingly.

(5/14/12 Conf. Tr. at 51-52, emphasis added.)

After the February 8 conference (and D OAR’s February 13 press release), the parties submitted their “final” ESI Protocol which the Court “so ordered.” (Dkt. No. 92: 2/17/12 ESI Protocol & Order.) Plaintiffs included a paragraph in the Protocol objecting to it “in its entirety,” noting that “Plaintiffs [had] submitted their own proposed ESI protocol [including predictive coding] to the Court, but it was largely rejected.” (2/17/12 ESI Protocol & Order ¶ J. 1 at 22.)

Plaintiffs’ Objections to the Court’s February 8,2012 Rulings

Plaintiffs filed their objections to my February 8 rulings on February 22, 2012. (Dkt. No. 93: Pis. Rule 72(a) Objections; see Dkt. No. 94: Nurhussein Aff.; Dkt. No. 95: Neale Aff.) Plaintiffs’ consultant, DOAR CEO Neale, reiterated that: “As stated during the February 8, 2012 hearing and cited in Judge Peck’s opinion, I am a proponent of the use of predictive coding, when it can be validated as reliable.” (Dkt. No. 125: Neale 3/19/12 Aff. ¶9.) Plaintiffs also conceded that they were willing to consider the use of predictive coding, “if it was established as reliable.” (Dkt. No. 123: Pis. Rule 72 Objections Reply Br. at 5, 10.) Plaintiffs distinguished between generalities and details: “As Plaintiffs repeatedly warned, although the use of predictive coding may be appropriate under certain circumstances, the devil is in the details.” (Id. at 1, emphasis added.) Plaintiffs’ objections referred to my Search, Forward article and my panels at LegalTech, noting that I “primarily discussed, as a general matter, the propriety of computer-assisted document review (no doubt meaning some variations of predictive coding), but the real question was whether MSL’s specific protocol would adequately address its Rule 26 obligations.” (Id. at 1-3, 7-8.)

The Court’s February 24, 2012 Decision is Affirmed by Judge Carter

The Court issued its formal opinion on the ESI protocol on February 24, 2012. Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, — F.R.D. -, 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.). Judge Carter overruled plaintiffs’ objections and adopted my decision on April 26, 2012. Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012).

On March 9, 2012, the Court held a further discovery conference. (Dkt. No. 209: 3/9/12 Conf. Tr.) Plaintiffs did not raise recusal at all at the conference. (See generally 3/9/12 Conf. Tr.)

Plaintiffs’ Recusal Request and Motion

It was not until March 28, 2012 that plaintiffs wrote a letter asking me to recuse myself from the case. Because MSL wrote that it would want a chance to respond, I entered an order asking plaintiffs whether they wanted me to rule based on letter-briefs or whether they wanted to file a formal motion, and added the following comment:

The Court notes that my favorable view of computer assisted review technology in general was well known to plaintiffs before I made any ruling in this case, and I have never endorsed Recommind’s methodology or technology, nor received any reimbursement from Recommind for appearing at any conference that (apparently) they and other vendors sponsored, such as Legal Tech. I have had no discussions with Mr. Losey about this case, nor was I aware that he is working on the case. It appears that after plaintiffs’ counsel and vendor represented to me that they agreed to the use of predictive coding, plaintiffs now claim that my public statements approving generally of computer assisted review make me biased. If plaintiffs were to prevail, it would serve to discourage judges (and for that matter attorneys) from speaking on educational panels about ediscovery (or any other subject for that matter). The Court suspects this will fall on deaf ears, but I strongly suggest that plaintiffs rethink their “scorched earth” approach to this litigation.

(Dkt. No. 158: 4/2/12 Order.)

On April 13, 2012, plaintiffs filed their recusal motion. (Dkt. No. 169: Motion; see also Dkt. No. 170: Pis. Recusal Br.; Dkt. No. 171: Wittels Aff.)

Subsequent Events in the Case

On April 20, 2012, plaintiffs asked Judge Carter to hold off on ruling on their objections to my February 8 and February 24 rulings about predictive coding until the recusal motion was decided; on April 25, 2012, Judge Carter rejected their request. (Dkt. No. 174: 4/25/12 Memo Endorsed Order.)

Also on April 25,1 held a further discovery conference. (Dkt. No. 180:4/25/12 Conf. Tr.) At the start of the conference, plaintiffs asked me to stay discovery until Judge Carter ruled on their pending motions to further amend their complaint and for FLSA collective action certification. (4/25/12 Conf. Tr. at 2-3.) Only after I denied that application did plaintiffs ask that I take no further action in the case until after I (and presumably Judge Carter) ruled on plaintiffs’ recusal motion. (4/25/12 Conf. Tr. at 13.) I denied that request. (4/25/12 Conf. Tr. at 14.) The order in which plaintiffs raised those two applications is indicative of their apparent strategy of seeing whether they prevail on matters and when they do not, only then raising recusal.

On May 14, 2012, I sua sponte reconsidered plaintiffs’ request and stayed MSL’s review and production of ESI pending Judge Carter’s ruling on plaintiffs’ motion to amend the complaint and for collective action certification; MSL originally objected but then did not oppose the stay, which I entered in order to avoid the expense of redoing discovery if collective action certification was granted. (Dkt. No. 202: 5/14/12 Conf. Tr. at 71-84.) See Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, 2012 WL 1698980 at *1 (S.D.N.Y. May 14, 2012) (Peck, M.J.).

ANALYSIS

I. RECUSAL LEGAL STANDARD

Under 28 U.S.C. § 455(a), “[a]ny justice, judge, or magistrate judge of the United States shall disqualify himself in any proceeding in which his impartiality might reasonably be questioned.” 28 U.S.C. § 455(a). “The purpose of § 455(a) ‘is to promote confidence in the judiciary by avoiding even the appearance of impropriety whenever possible.’ ” Green v. N.Y.C. Health & Hosps. Corp., 343 Fed. Appx. 712, 713 (2d Cir.2009) (quoting Lilje berg v. Health Servs. Acquisition Corp., 486 U.S. 847, 865, 108 S.Ct. 2194, 2205, 100 L.Ed.2d 855 (1988)).

“In determining whether Section 455(a) requires recusal, the appropriate standard is objective reasonableness— whether ‘an objective, disinterested observer fully informed of the underlying facts, [would] entertain significant doubt that justice would be done absent recusal.’ ” United States v. Carlton, 534 F.3d 97, 100 (2d Cir.), cert, denied, 555 U.S. 1038, 129 S.Ct. 613, 172 L.Ed.2d 468 (2008); accord, e.g., Green v. N.Y.C. Health & Hosps. Corp., 343 Fed.Appx. at 713-14; In re Basciano, 542 F.3d 950, 956 (2d Cir.2008), cert, denied, 555 U.S. 1177, 129 S.Ct. 1401, 173 L.Ed.2d 596 (2009); United States v. Wecht, 484 F.3d 194, 213 (3d Cir.2007) (“ ‘The test for recusal under § 455(a) is whether a reasonable person, with knowledge of all the facts, would conclude that the judge’s impartiality might reasonably be questioned.’ ”); United States v. Lauersen, 348 F.3d 329, 334 (2d Cir.2003) (“Disqualification under section 455(a) requires a showing that would cause ‘an objective, disinterested observer fully informed of the underlying facts [to] entertain significant doubt that justice would be done absent recusal.’ ”), cert, denied, 541 U.S. 1044, 124 S.Ct. 2190, 158 L.Ed.2d 735 (2004). Thus,

the existence of the appearance of impropriety is to be determined “not by considering what a straw poll of the only partly informed man-in-the-street would show[,] but by examining the record facts and the law, and then deciding whether a reasonable person knowing and understanding all the relevant facts would recuse the judge.”

United States v. Bayless, 201 F.3d 116, 126-27 (2d Cir.), cert, denied, 529 U.S. 1061, 120 S.Ct. 1571, 146 L.Ed.2d 474 (2000). “The Court of Appeals has cautioned that when answering this question [i.e., the objective test], ‘the grounds asserted in a recusal motion must be scrutinized with care, and judges should not recuse themselves solely because a party claims an appearance of partiality.’ ” Barnett v. United States, 2012 WL 1003594 at *1.

To establish a basis for recusal, “[m]ovants must overcome a presumption of impartiality, and the burden for doing so is substantial.” Metro. Opera Ass’n, Inc. v. Local 100, Hotel Emps. Int’l Union, 332 F.Supp.2d at 670 (quotations omitted); accord, e.g., United States v. Denton, 434 F.3d 1104, 1111 (8th Cir.2006) (“A judge is presumed to be impartial, and ‘the party seeking disqualification bears the substantial burden of proving otherwise.’ ”).

“Discretion is confided in the district judge in the first instance to determine whether to disqualify himself. The reasons for this are plain. The judge presiding over a case is in the best position to appreciate the implications of those matters alleged in a recusal motion.” In re Drexel Burnham Lambert Inc., 861 F.2d at 1312 (citation omitted); accord, e.g., In re Basciano, 542 F.3d at 956; In re Certain Underwriter, 294 F.3d 297, 302 (2d Cir.2002); United States v. Roldan-Zapata, 916 F.2d 795, 802 (2d Cir.1990), cert, denied, 499 U.S. 940, 111 S.Ct. 1397, 113 L.Ed.2d 453 (1991). “[A] judge has an affirmative duty ... not to disqualify himself unnecessarily, particularly ‘where the request for disqualification was not made at the threshold of the litigation and the judge has acquired a valuable background of experience.’ ” Nat’l Auto Brokers Corp. v. Gen. Motors Corp., 572 F.2d 953, 958 (2d Cir.1978), cert, denied, 439 U.S. 1072, 99 S.Ct. 844, 59 L.Ed.2d 38 (1979); accord, e.g., LoCascio v. United States, 473 F.3d 493, 498 (2d Cir.), cert, denied, 552 U.S. 1010, 128 S.Ct. 554, 169 L.Ed.2d 374 (2007); United States v. Yonkers Bd. of Educ., 946 F.2d 180,183 (2d Cir.1991). “A judge is as much obliged not to recuse himself when it is not called for as he is obliged to when it is.” In re Drexel Burnham Lambert Inc., 861 F.2d at 1312; accord, e.g., In re Literary Works in Elec. Databases Copyright Litig., 509 F.3d 136, 140 (2d Cir.2007); Metro. Opera Ass’n, Inc. v. Local 100, Hotel Emps. Int’l Union, 332 F.Supp.2d at 670; In re Certain Underwriter, 294 F.3d at 302.

“In deciding whether to recuse himself, the trial judge must carefully weigh the policy of promoting public confidence in the judiciary against the possibility that those questioning his impartiality might be seeking to avoid the adverse consequences of his presiding over their case.” In re Drexel Burnham Lambert Inc., 861 F.2d at 1312; accord, e.g., In re Basciano, 542 F.3d at 956; In re Certain Underwriter, 294 F.3d at 302. This is because “[l]itigants are entitled to an unbiased judge; not to a judge of their choosing.” In re Drexel Burnham Lambert Inc., 861 F.2d at 1312; accord, e.g., Mulligan v. Loschiavo, 173 Fed.Appx. 26, 28 (2d Cir.2006). Moreover,

the public interest mandates that judges not be intimidated out of an abundance of caution into granting disqualification motions: “A trial judge must be free to make rulings on the merits without the apprehension that if he makes a disproportionate number in favor of one litigant, he may create the [appearance] of bias,” and ‘“[a] timid judge, like a biased judge, is intrinsically a lawless judge.’ ”

Metro. Opera Ass’n, Inc. v. Local 100, Hotel Emps. Int’l Union, 332 F.Supp.2d at 670.

II. DA SILVA MOORE’S RECUSAL MOTION IS UNTIMELY

A. Legal Standard

Although 28 U.S.C. § 455 does not explicitly address timeliness, such a requirement “has been read into this section,” addressing two underlying concerns. Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d 326, 333 (2d Cir.1987); accord, e.g., Taylor v. Vt. Dep’t of Educ., 313 F.3d 768, 794 (2d Cir.2002) (Sotomayor, C.J.); United States v. Brinkworth, 68 F.3d 633, 639 (2d Cir.1995); Polizzi v. United States, 926 F.2d 1311, 1321 (2d Cir.1991). “First, judicial resources should not be wasted; and, second, a movant may not hold back and wait, hedging its bets against the eventual outcome.” Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d at 334; accord, e.g., LoCascio v. United States, 473 F.3d 493, 497 (2d Cir.2007), cert, denied, 552 U.S. 1010, 128 S.Ct. 554, 169 L.Ed.2d 374 (2007); United States v. Brinkworth, 68 F.3d at 639; Gil Enters., Inc. v. Delvy, 79 F.3d 241, 247 (2d Cir.1996); In re IBM Corp., 45 F.3d 641, 643 (2d Cir.1995) (A “prompt application avoids the risk that a party is holding back a recusal application as a fail-back position in the event of adverse rulings on pending matters”); United States v. Yonkers Bd. of Educ., 946 F.2d 180, 183 (2d Cir.1991)

Recusal motions must be made “at the earliest possible moment after obtaining knowledge of facts demonstrating the basis for such a claim.” Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d at 333; accord, e.g., Weisshaus v. Fagan, 456 Fed. Appx. 32, 34 (2d Cir.2012); United States v. Amico, 486 F.3d 764, 773 (2d Cir.2007). Courts have held that:

For purposes of timeliness, the applicant is charged with knowledge of all facts “known or knowable, if true, with due diligence from the public record or otherwise.” Any other rule would allow a member of a law firm aware of facts that might lead to judicial disqualification to sit on the information, wait to see which way the wind appears to be blowing with the judge, and then to come forward in an effort to get rid of the judge if a colleague responsible for a case begins to perceive that the judge is unreceptive to the client’s position or even simply wants a delay.

Universal City Studios, Inc. v. Reimerdes, 104 F.Supp.2d at 349-50; accord, e.g., United States v. Daley, 564 F.2d 645, 651 (2d Cir.1977) (motion for recusal untimely because, inter alia, the facts upon which it was based “as a matter of public record, were at all times ascertainable by counsel”), cert, denied, 435 U.S. 933, 98 S.Ct. 1508, 55 L.Ed.2d 530 (1978).

In deciding whether a recusal motion is timely, a court looks to a number of factors, including whether: “(1) the movant has participated in a substantial manner in trial or pre-trial proceedings; (2) granting the motion would represent a waste of judicial resources; (3) the motion was made after the entry of judgment; and (4) the movant can demonstrate good cause for delay.” Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d at 334 (citations omitted); accord, e.g., Weisshaus v. Fagan, 456 Fed.Appx. at 34; United States v. Amico, 486 F.3d at 773; Taylor v. Vt. Dep’t of Educ., 313 F.3d at 794-95; United States v. Brinkworth, 68 F.3d at 639.

B. Application

Plaintiffs’ seek my recusal due to my advocacy of predicative coding, relationship with Losey and speaking engagements at LegalTech. (See generally Dkt. No. 170: Pis. Br.; Dkt. No. 192: Pis. Reply Br.) Plaintiffs and defendants, however, have been aware of my view on predicative coding since at least December 2, 2011, and my relationship with Losey and engagement with LegalTech since January 4, 2012. (See page 143 above.)

During the first conference on December 2, 2011, I made the parties aware of my knowledge of predictive coding and recommended that parties read my Search, Forward article. (See pages 140-41 above.) Specifically, I said:

Now, if you want any more advice, for better or for worse on the ESI plan and whether predictive coding should be used, or anything else ... I will say right now, what should not be a surprise, I wrote an article in the October Law Technology News called Search Forward, which says predictive coding should be used in the appropriate case.

(See page 141 above, emphasis added.) Indicating that I am familiar with the ediscovery industry and know several people involved, when discussing whether a special master would be useful, I said, ‘You know, I know enough people in the industry that I can recommend some, or you all can get your vendors to recommend somebody....” (See page 141 n. 1 above.) I also made the comment, ‘You must have thought you died and went to Heaven when this was referred to me” (see page 140 above), and MSL’s counsel responded that counsel was “just thankful that ... we have a person familiar with the predictive coding concept” (see page 140 above).

On December 20, 2011, I granted plaintiffs’ request for an adjournment of the upcoming discovery conference, and in doing so reminded the parties to read my article, stating, “As to predictive coding, you should read my article, ‘Search, Forward’ in the Oct. 2011 issue of Law Technology News.” (See page 142 above.)

During the January 4, 2012 conference, while discussing substitute counsel for MSL counsel Anders during his vacation, I made the parties aware that I knew Ralph Losey, a partner in MSL’s counsel’s ediscovery practice group. (See page 143 above.) The following colloquy occurred:

THE COURT: I know every lawyer thinks they’re indispensable and I’m not pulling the “Jackson Lewis is a big firm and you’re all fungible,” but is there not another person who may be less email savvy or computer savvy than you, such as Ms. Chavey, for example, who can follow up, along with the folks from Recommind and plaintiffs’ counsel, and not lose an entire week because you’re on vacation?

MS. CHAVEY: Of course, your Honor.

THE COURT: And I happen to know, it may not be on this case, if it’s a true e-discovery dispute, I happen to know your Florida e-discovery counsel very well—

MR. ANDERS: He knows a little bit.

THE COURT: You can bring Mr. Losey into the mix if need be.

MR. ANDERS: OK, understood.

(Dkt. No. 71: 1/4/12 Conf. Tr. at 61, emphasis added; see page 143 above.) Also at the January 4 conference while setting the date for the next conference, I implicitly made it known that I would be unavailable due to my speaking engagements at LegalTech. (See page 143 above.) The following colloquy occurred:

THE COURT: OK, next, date for our next court conference, what’s your pleasure?

MS. CHAVEY: Your Honor, what about February 2nd?

THE COURT: That’s LegalTech week. Yes, by Thursday that’s OK. February 2nd at 9:30.

(1/4/12 Conf. Tr. at 66; see page 143 above.) Plaintiffs’ ediscovery consultant, DOAR, is very familiar with the Legal Tech conferences. (See page 143 n. 6 above.)

Plaintiffs argue that their motion should “not be denied as untimely [because]: (1) the ease remains in its early stages; (2) granting the motion would not represent a waste of judicial resources; (3) the motion was not made after entry of judgment; and (4) Plaintiffs have acted promptly upon discovering the relevant facts.” (Pis. Reply Br. at 6.)

With respect to the first factor, plaintiffs have been active participants in pretrial proceedings since before my first conference on December 2, 2011. While plaintiffs are correct that the case is in its early stages, discovery has been ongoing since at least October 2011 (see Dkt. No. 44: 10/12/11 Order), if not before. At conferences and through written communications, plaintiffs have enlisted both myself and the District Judges with respect to the ediscovery protocol, adjournments, issues arising during depositions, stipulations and other pretrial issues. (See, e.g., Dkt. No. 88: 2/8/12 Conf. Tr.; Dkt. No. 91: 2/2/12 Telephone Conf. Tr.; Dkt. No. 92: 2/17/12 ESI Protocol & Order; Dkt. No. 108: 3/8/12 Stipulation & Order; Dkt. No. 109: Am. Joint Scheduling Order; Dkt. No. 118: 3/19/12 Stipulation & Order; Dkt. No. 128: 3/28/12 Memo Endorsed Order; Dkt. No. 209: 3/9/12 Conf. Tr.; Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, — F.R.D. -, 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.), adopted, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012).)

With respect to the second factor, this case was referred to me on November 28, 2011 for general pretrial supervision. (See page 140 above.) Since that time, I have familiarized myself with this case and have expended considerable time and attention in responding to the parties’ discovery issues and disputes. While plaintiffs assert that “granting the motion would not represent a waste of judicial resources” (Pis. Reply Br. at 6), another magistrate judge would have to spend ample time to familiarize himself or herself with this complex case. Moreover, plaintiffs have stated that if recusal is granted, they will ask Judge Carter to review and overturn my decisions and orders in the case. (Pis. Br. at 25.)

With respect to the third factor, while it is true that, as plaintiffs assert, “the motion was not made after entry of judgment” (Pis. Reply Br. at 6), it is also true that plaintiffs waited to seek my recusal until after I adopted MSL’s predictive coding protocol (see pages 147-48 above). Plaintiffs themselves assert that “[predictive coding without proper safeguards will impact the merits of this case.” (Pis. Reply Br. at 7.) Courts in the Second Circuit have required “a prompt application [to avoid] the risk that a party is holding back a recusal application as a fallback position in the event of adverse rulings on pending matters.” In re IBM Corp., 45 F.3d 641, 643 (2d Cir.1995). It appears that plaintiffs are improperly using the recusal motion as “fall-back position” to an unfavorable ruling. See, e.g., Weisshaus v. Fagan, 456 Fed.Appx. 32, 34 (2d Cir.2012) (Motion held untimely. “Although there was no dispositive ruling as to [defendant] at the time [plaintiff] brought her recusal motion, the district court aptly noted that the motion came on the heels of its direction that [plaintiff] submit to a deposition, thus strongly suggesting that the motion was a mere fallback position in response to an adverse ruling.”); Silver v. Kuehbeck, 217 Fed. Appx. 18, 23-24 (2d Cir.2007) (Motion held untimely. “[A]fter the April 18, 2005 settlement conference, from which [plaintiff] was ‘left with no choice but to assume that what transpired ... may have colored’ the district court’s perception, [plaintiff] continued to litígate the case for two more months, even appearing before the Court for oral arguments on the motion to dismiss on May 23 without requesting recusal. It thus appears that [plaintiff] held back his ‘recusal application as a fail-back position in the event of adverse rulings on pending matters.’”); LoCascio v. United States, 473 F.3d 493, 497 (2d Cir.2007) (recusal motion untimely where plaintiff “made no mention of the above remark until after the District Court had denied his motion to amend and after it had denied his § 2255 petition.”), cert, denied, 552 U.S. 1010, 128 S.Ct. 554, 169 L.Ed.2d 374 (2007); Gil Enters., Inc. v. Delvy, 79 F.3d 241, 247 (2d Cir.1996) (“In the face of defeat, [plaintiff] now seeks to do just that which this Court warned against in In re IBM and use its post-hoc recusal motion as a ‘fall-back position.’ Since [plaintiff's objection was raised well later than ‘the earliest possible moment,’ and because [plaintiff] has failed to demonstrate any bias on the part of the district court beyond [the Judge]’s own expressed frustration, we reject this ground for relief’); Armenian Assembly of Am., Inc. v. Cafesjian, 783 F.Supp.2d 78, 88 (D.D.C.2011) (“[I]t appears that the filing of the [recusal] motion was motivated by the fact that Plaintiffs received a largely adverse decision from the Court rather than by the sudden discovery of an alleged bias.”); Katzman v. Victoria’s Secret Catalogue, 939 F.Supp. 274, 278 (S.D.N.Y.1996) (Plaintiff “and her counsel now seek to do just that which the Second Circuit and other courts have warned against; namely, use a post-judgment recusal motion to try to get a second bite at the apple.”), affd, 113 F.3d 1229 (2d Cir.1997).

With respect to the fourth factor, plaintiffs assert that they “have acted promptly upon discovering the relevant facts” but provide no explanation for their delay. (See Pis. Reply Br. at 6.) Despite plaintiffs’ knowledge as of December 2, 2011 of my views on predicative coding, and by January 4, 2012 as to my relationship with Losey and my speaking at LegalTech (see pages 140 41, 143 above), plaintiffs did not request my recusal until March 28, 2012 by letter (see pages 147-48 above) and did not file their formal recusal motion until April 13, 2012 (see page 148 above). The movant “is charged with knowledge of all facts ‘known or knowable, if true, with due diligence from the public record or otherwise.’” Universal City Studios, Inc. v. Reimerdes, 104 F.Supp.2d 334, 349 (S.D.N.Y.2000); accord, e.g., Armenian Assembly of Am., Inc. v. Cafesjian, 783 F.Supp.2d at 87 (timeliness rule “‘has been applied when the facts upon which the [recusal] motion relies are public knowledge, even if the movant does not know them.’ ”). Plaintiffs here had the requisite knowledge no later than January 4, 2012, but the recusal request did not come until nearly three months later. (See 147-A8 above.) Courts have found shorter delays to be untimely. See, e.g., Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d 326, 334 (2d Cir.1987) (motion untimely where party waited two months after events giving rise to charge of bias or prejudice before making its recusal motion, despite fact that other of the “factors do not support a finding of untimeliness”); Six W. Retail Acquisition, Inc. v. Sony Theatre Mgmt. Corp., 97 Civ. 5499, 2003 WL 282187 at *1, *4 (S.D.N.Y. Feb. 7, 2003) (two-month delay untimely), affd, 124 Fed. Appx. 73 (2d Cir.), cert, denied, 546 U.S. 1016, 126 S.Ct. 660, 163 L.Ed.2d 526 (2005); Katzman v. Victoria’s Secret Catalogue, 939 F.Supp. at 278 (two-month delay untimely); Lamborn v. Dittmer, 726 F.Supp. 510, 515 (S.D.N.Y.1989) (“Recusal motions are often denied on the basis of untimeliness where there has been only a short delay.” (citing, inter alia, In re Mar tirir-Trigona, 573 F.Supp. 1237, 1244-45 (D.Conn.1983) (motion untimely based on twelve-day delay), appeal dismissed, 770 F.2d 157 (2d Cir.1985), cert, denied, 475 U.S. 1058, 106 S.Ct. 1285, 89 L.Ed.2d 592 (1986))). I have made no efforts to hide my views, relationships or affiliations. If plaintiffs truly believed that any of these issues, individually or collectively, created a bias or the appearance of partiality, they should have promptly moved for my recusal.

Accordingly, plaintiffs’ recusal motion is untimely. In any event, it also is merit-less, as will be discussed in the following sections.

III. DA SILVA MOORE’S RECUSAL MOTION IS MERITLESS

Plaintiffs assert that my recusal is required due to: (1) my “public comments concerning the case” (Dkt. No. 170: Pis. Br. at 12-14; Dkt. No. 192: Pis. Reply Br. at 8-9), (2) my “participation on pro-predictive coding panels with defense counsel Ralph Losey while presiding over the parties’ dispute on predictive coding” (Pis. Br. at 14-15; Pis. Reply Br. at 7-8), (3) my “numerous speaking engagements in favor of predictive coding, which were at least indirectly sponsored and funded by Recommind and other e-discovery vendors” (Pis. Br. at 15-17), and (4) my “failure to disclose [my] activities enhances the appearance of impropriety” (Pis. Br. at 19-21).

A. Speaking Engagements

With respect to my speaking engagements on the subject of computer-assisted review, I only spoke generally about computer-assisted review in comparison to other search techniques. {See pages 143-44 above.) The fact that my interest in and knowledge about predictive coding in general overlaps with issues in this case is not a basis for recusal. See, e.g., Hoatson v. N.Y. Archdiocese, 280 Fed.Appx. 88, 90 (2d Cir.2008) (In a case involving the Catholic Church, fact that the district judge was “member of the Guild of Catholic Lawyers of the Archdiocese of New York,” had received an award from that organization and occasionally attended its meetings discussing Catholic education and Catholic commitment to social justice was not a basis for appearance of impropriety recusal.); Hu v. Am. Bar Ass’n, 334 Fed.Appx. 17, 19 (7th Cir.2009) (A “judge’s membership in a bar association, or his receipt of reimbursement for participating in bar-association activities, does not create the type of relationship that would cause us to doubt his ability to preside impartially over a case in which the bar association is a party.”); Lunde v. Helms, 29 F.3d 367, 370-71 (8th Cir.1994) (no basis for recusal where judge was an alumnus of the university defendant, had made financial contributions to the university and had participated in the university’s educational programs), cert, denied, 513 U.S. 1155, 115 S.Ct. 1111, 130 L.Ed.2d 1076 (1995); Wu v. Thomas, 996 F.2d 271, 275 (11th Cir.1993) (recusal not required where judge made past contributions to the university defendant and held position as unsalaried adjunct professor), cert denied, 511 U.S. 1033, 114 S.Ct. 1543, 128 L.Ed.2d 195 (1994); Sierra Club v. Simkins Indus., Inc., 847 F.2d 1109, 1117-18 (4th Cir.1988) (judge’s Sierra Club membership before appointment to the bench did not require recusal from case where the Sierra Club was a party), cert, denied, 491 U.S. 904, 109 S.Ct. 3185, 105 L.Ed.2d 694 (1989); United States v. Alabama, 828 F.2d 1532, 1543-44 (11th Cir.1987) (“all judges come to the bench with a background of experiences, associations and viewpoints.... A judge is not required to recuse himself merely because he holds and has expressed certain views on a general subject.” Judge’s background as civil rights lawyer and state legislator did not require disqualification in desegregation case, but other factors did.), cert, denied, 487 U.S. 1210, 108 S.Ct. 2857, 101 L.Ed.2d 894 (1988); Shaw v. Martin, 733 F.2d 304, 316 (4th Cir.) (“One who has voted as a legislator in favor of a statute permitting the death penalty in a proper case cannot thereafter be presumed disqualified _ to hear capital cases as a judge or predisposed to give a death sentence in any particular case.”), cert, denied, 469 U.S. 873, 105 S.Ct. 230, 83 L.Ed.2d 159 (1984).

While speaking on ediscovery panels in January 2012, I mentioned that I had a case in front of me using computer-assisted review, but I did not mention the parties or counsel involved. (See pages 144-45 & n. 11 above.) The only arguably identifiable statement was the “died and went to Heaven” comment. (See pages 144-45 & n. 11 above.) This comment, however, was originally made in open court at the December 2, 2011 conference and was available in the public transcript. (See pages 140-41 above.) Consequently, this is not a recusable statement. See, e.g., United States v. Pitera, 5 F.3d 624, 626-27 (2d Cir.1993) (upholding district court’s denial of recusal motion by defendant in narcotics prosecution, where judge had lectured to the DEA Task Force including advice on steps to take to increase convictions, but also spoke at a PLI program for criminal defense lawyers, and the “record discloses that the Judge commendably lectures to a variety of trial practice seminars”), cert, denied, 510 U.S. 1131, 114 S.Ct. 1103,127 L.Ed.2d 415 (1994); United States v. Yonkers Bd. of Educ., 946 F.2d 180, 184-85 (2d Cir.1991) (finding “no impropriety” in the District Judge’s public comments in the media about a pending case because the judge “only restated what he had been saying in open court for the past few years and did not discuss the details of remedy implementation”); Wilborn v. Wells Fargo Bank, N.A. (In re Wilborn), 401 B.R. 848, 863 (Bankr. S.D.Tex.2009) (“Here, the undersigned judge’s disagreement with certain legal arguments previously rejected by this Court and others, expressed at the Dallas Seminar, are not sufficient to warrant recusal.”); Metro. Opera Ass’n, Inc. v. Local 100, Hotel Emps. Int’l Union, 332 F.Supp.2d 667, 674-75 (S.D.N.Y.2004) (“[T]he summary of factual findings from the Opinion, used as a springboard in the [CLE] Presentation to discuss ‘best practices’ in electronic discovery to avoid such findings, is insufficient to require recusal, even when a motion to reconsider was pending.”).

The Metropolitan Opera case is particularly instructive. There, Judge Preska issued a “lengthy opinion” granting judgment for plaintiff and awarding it attorneys’ fees based on defendants’ discovery abuses. See Metro. Opera Ass’n, Inc. v. Local 100, Hotel Emps. Int’l Union, 332 F.Supp.2d at 669. While a motion for reconsideration was pending, Judge Preska gave a presentation at a BNA CLE program “on the topic of electronic discovery and how technological advances might affect discovery obligations.” Id. at 669-70. “The Presentation began with a summary of selected discovery failures set out in the Opinion, ... and proceeded to recommend ... steps to take to avoid these and other pitfalls in electronic discovery, including steps that echoed findings in the Opinion.” Id. Judge Preska denied defendants’ § 455(a) disqualification motion. Id. at 676. Movants “argue[d] that a few of the phrases in the course of the Presentation are indicative of [the judge’s] ‘personal involvement’ or ‘emotional commitment’ to one side of the dispute,” including reference to a colloquial comment that the [defendant] “ ‘junked’ ” some of its computers. Id. at 672-73. Judge Preska responded: “Whether or not my subjective pedagogical intent was successfully conveyed, the remarks complained of could not be interpreted by the objective disinterested observer as conveying the level of personal involvement required for recusal.” Id. at 673. Judge Preska concluded:

As noted above, the Presentation began with a summary of selected findings in the Opinion relevant to the seminar topic. Movants do not suggest that the summary inaccurately reported what the Opinion stated but object because certain of the findings recited detail “failings” of defendants and their counsel and because some of those findings are challenged in the motion for reconsideration. First, the finding that certain actions (or inactions) constituted failings by defendants and their counsel was the very basis of the Opinion....

Second, as noted above, the audience was informed that a reconsideration motion was pending. Movants nowhere explain why a reasonable observer would assume from this summary of findings from the Opinion, recited for the stated purpose of advising counsel how to avoid similar failures in the future, that the Court could not decide the pending motion for reconsideration in an impartial manner any more than he or she would make that assumption from merely reading the Opinion. Accordingly, the summary of factual findings from the Opinion, used as a springboard in the Presentation to discuss “best practices” in electronic discovery to avoid such findings, is insufficient to require recusal, even when a motion to reconsider was pending.

Having “carefully weighted] the policy of promoting public confidence in the judiciary against the possibility that those questioning [my] impartiality might be seeking to avoid the adverse consequences of [my] presiding over their case.” I find that movants have not carried their substantial burden of showing that a reasonable observer, with knowledge and understanding of the relevant facts, would “entertain significant doubt that justice would be done absent recusal,” based upon the Presentation at the BNA seminar. Accordingly, movants’ motion to disqualify is denied.

Id. at 674-75, 676 (citations omitted).

My comments were nowhere near the comments in Metropolitan Opera, which were not a basis for recusal. I did not mention this case by name. (See pages 148-44 above.) While I briefly mentioned that I had a case in front of me where a party proposed using predictive coding (and noted that, on the record, I had made the “died and went to heaven” comment), the case was not yet the subject of publicity (which only resulted after the February 13, 2012 press release by plaintiffs’ consultant DOAR) and the objective reasonable obsexwer would not have known that those brief comments referred to this case. (See pages 144, 145-47 above.) Moreover, the comments were minor and fleeting, involved only facts on the public record and, like Metropolitan Opera, were for educational purposes.

To the extent plaintiffs are complaining about my general discussion at these CLE presentations about the use of predictive coding in general, those comments would not cause a reasonable objective observer to believe I was biased in this case. I did not say anything about predictive coding at these LegalTech and other CLE panels that I had not already said in my Search, Forward article, i.e., that lawyers should consider using predictive coding in appropriate cases. My position was the same as plaintiffs’ consultant, DOAR CEO Neale. (See pages 145-46 above.) Both plaintiffs and defendants were proposing using predictive coding in this case. (See page 142 above.) I did not determine which party’s predictive coding protocol was appropriate in this case until the February 8, 2012 conference, after the panels about which plaintiffs complain.

In objecting to my February 8 ruling, plaintiffs informed Judge Carter that while predictive coding may be appropriate under certain circumstances, “the devil is in the details.” (See page 147 above.) I did not discus the “details” of a predictive coding protocol {e.g., number of iterations needed to train the computer, how many “seed” documents would be used, appropriate sample size) at any of the Legal-Tech or other CLE panels (or in my Search, Forward article), nor did any of the other panelists. Thus, a reasonable objective observer would not think that my comments at these educational panels gives the appearance of bias for MSL or against plaintiffs. My participation in the panels, like Judge Preska’s CLE presentation in Metropolitan Opera, does not require recusal.

B. Counsel Losey & Recommind’s Participation in LegalTech

While I participated on two panels with defense counsel Losey, we never had any ex parte communication regarding this lawsuit. {See pages 143-44 above.) My preparation for and participation in ediscovery panels involved only ediscovery generally and the general subject of computer-assisted review. {See pages 143-44 above.) Losey’s affidavit makes clear that we have never spoken about this case, and I confirm that. {See pages 143-44 above.) During the panel discussions (and preparation sessions), there was absolutely no discussion of the details of the predictive coding protocol involved in this case or with regard to what a predicative coding protocol should look like in any case. {See pages 143-44 above.) Plaintiffs’ assertion that speaking on an educational panel with counsel creates an appearance of impropriety is undermined by Canon 4 of the Judicial Code of Conduct, which encourages judges to participate in such activities. See Code of Conduct for United States Judges, Canon 4 (“A judge may engage in extrajudicial activities, including law-related pursuits and ... educational ... activities, and may speak, write, lecture, and teach on both law-related and nonlegal subjects.”), available at http:// www.uscourts.gov/Viewer.aspx?doc=/ uscourts/RulesAndPolicies/conduct/Vol02 A-Ch02.pdf (last visited June 14, 2012).

The cases make clear that participation on an educational panel with counsel is not a basis for recusal, for sound policy reasons. See, e.g., Leja v. Schmidt Mfg., Inc., No. Civ. 01-5042, 2010 WL 2571850 at *2 (D.N.J. June 22, 2010) (“In my own case, during the course of 31 years on the bench, I have developed numerous personal friendships with members of the Bar and have participated in many charitable, legal and public service organizations in which lawyers, law firms and other judges have participated. This is probably the experience of most judges. To permit such associations to become grounds for recusal would either push judges towards a hermit like existence or open the floodgates to recusal motions.”); In re Wolverine Proctor & Schwartz, LLC, 397 B.R. 179, 183 (Bankr.D.Mass.2008) (“[T]he Court rejects [creditor’s] assumption that [the judge’s] service on the Financial Literacy Committee with [counsel], and numerous other volunteer lawyers and fellow judges, gives rise to a disqualifying connection and establishes reasonable grounds for doubting this Court’s impartiality.”); In re Healy, No. 04-28375-D-13L, 2006 WL 3751617 at *4 (Bankr.E.D.Cal. Dec. 18, 2006) (“It is common knowledge, of course, that judges regularly appear on panels and at presentations for members of the bar, and that such events are regularly advertised in various publications that might be viewed by both the public and the bar. But it is not reasonable to conclude that the participation of a judge with members of the bar who appear before the judge’s court would create a predisposition, or an appearance of a predisposition, to favor the members of the bar who participate over those who do not. The Debtor’s assertion of an appearance of impropriety is undermined by Canon 4 of the Code of Conduct, which not only permits judges to participate in such activities, but encourages judges to do so.”); Moran v. Clarke, 213 F.Supp.2d 1067, 1073 (E.D.Mo.2002) (“A judge’s involvement with other attorneys in bar association activities is not a basis for recusal. Indeed, the commentary to Canon 4 ... encourages judges to ‘contribute to the improvement of the law, the legal system, and the administration of justice.... [T]he judge is encouraged to do so, either independently or through a bar association, judicial conference, or other organization dedicated to the improvement of the law.’ A judge should not be required to withdraw from all social relationships and live in seclusion.’ ”); Bailey v.