Justice Breyer, with whom Justice Stevens, Justice Souter, and Justice Ginsburg join, concurring in part and dissenting in part.

I concur in Parts I and II and dissent from Parts III and IV of the Court’s opinion. In Parts III and IV, the Court holds unlawful an agency conclusion on the ground that it is “not supported by substantial evidence.” Ante, at 19; see 29 U. S. C. §160(e); 5 U. S. C. §706(2)(E). That question was not presented to us in the petition for certiorari. In deciding it, the Court has departed from the half-century old legal standard governing this type of review. See Universal Camera Corp. v. NLRB, 340 U. S. 474, 490–491 (1951). It has rewritten a Board rule without adequate justification. It has ignored certain evidentiary presumptions developed by the National Labor Relations Board (Board) to provide guidance in the application of this rule. And it has failed to give the kind of leeway to the Board’s factfinding authority that the Court’s precedents mandate. See, e.g., Beth Israel Hospital v. NLRB, 437 U. S. 483, 504 (1978).

To decide whether an agency’s conclusion is supported by substantial evidence, a reviewing court must identify the conclusion and then examine and weigh the evidence. As this Court said in 1951, “[w]hether on the record as a whole there is substantial evidence to support agency findings is a question which Congress has placed in the keeping of the Courts of Appeals.” Universal Camera, supra, at 491. The Court held that it would “intervene only in what ought to be the rare instance when the standard appears to have been misapprehended or grossly misapplied.” Ibid.; see Beth Israel Hospital, supra, at 507 (“ ‘misapprehended or grossly misapplied’ ”); Golden State Bottling Co. v. NLRB, 414 U. S. 168, 173 (1973) (“ ‘misapprehended or grossly misapplied’ ”). Consequently, if the majority is to overturn a Court of Appeals’ “substantial evidence” decision, it must identify the agency’s conclusion, examine the evidence, and then determine whether the evidence is so obviously inadequate to support the conclusion that the reviewing court must have seriously misunderstood the nature of its legal duty.

The majority opinion begins by properly stating the Board’s conclusion, namely that the employer, Allentown, did not demonstrate that it

“held a reasonable doubt, based on objective considerations, that the Union continued to enjoy the support of a majority of the bargaining unit employees.” Ante, at 6 (emphasis added).

The opinion, however, then omits the words I have italicized and transforms this conclusion, rephrasing it as:

“Allentown lacked a genuine, reasonable uncertainty about whether Local 724 enjoyed the continuing support of a majority of unit employees.” Ante, at 8.

Key words of a technical sort that the Board has used in hundreds of opinions written over several decades to express what the Administrative Law Judge (ALJ) here called “objective reasonable doubt” have suddenly disappeared, leaving in their place what looks like an ordinary jury standard that might reflect, not an agency’s specialized knowledge of the workplace, but a court’s common understanding of human psychology. The only authority cited for the transformation, the dictionary, in fact offers no support, for the majority has looked up the wrong word, namely “doubt,” instead of the right word, “objective.” In any event, the majority’s interpretation departs from settled principles permitting agencies broad leeway to interpret their own rules, see, e.g., Thomas Jefferson Univ. v. Shalala, 512 U. S. 504, 512 (1994) (courts “must give substantial deference to an agency’s interpretation of its own regulations”); Bowles v. Seminole Rock & Sand Co., 325 U. S. 410, 413–414 (1945) (same), which may be established through rulemaking or adjudication, see NLRB v. Bell Aerospace Co., 416 U. S. 267, 294 (1974); SEC v. Chenery Corp., 332 U. S. 194, 202 (1947) (same).

To illustrate the problem with the majority’s analysis, I must describe the factual background, the evidence, and the ALJ’s findings, in some detail. In December 1990, three managers at Mack Trucks (and several other investors) bought Mack. All of the 45 employees in the Union’s bargaining unit were dismissed. The new owners changed the company’s name to Allentown and then interviewed and rehired 32 of the 45 recently dismissed workers, putting them back to work at jobs similar to those they previously held. The Union, which had represented those employees for 17 years, sought continued recognition; Allentown refused it; the Board’s general counsel brought unfair labor practice charges; and the ALJ found that Allentown was a “successor” corporation to Mack, 316 N. L. R. B. 1199, 1204 (1995), a finding that was affirmed by the Board, id., at 1199, and was not challenged in the Court of Appeals. Because Allentown was found to be a “successor” employer, the Union was entitled to a rebuttable presumption of majority status. See Fall River Dyeing & Finishing Corp. v. NLRB, 482 U. S. 27, 41 (1987). Absent some extraordinary circumstance, when a union enjoys a rebuttable presumption of majority status, the employer is obligated recognize the union unless 30% of the union’s employees petition the Board for a decertification election, (and the union loses), Texas Petrochemicals Corp., 296 N. L. R. B. 1057, 1062 (1989), enf ’d as modified, 923 F. 2d 398 (CA5 1991); see 29 U. S. C. §159(c)(1)(A)(ii); 29 CFR §101.18(a) (1997), or the employer shows that “either (1) the union did not in fact enjoy majority support, or (2) the employer had a good-faith doubt, founded on a sufficient objective basis, of the union’s majority support,” see NLRB v. Curtin Matheson Scientific, Inc., 494 U. S. 775, 778 (1990) (emphasis deleted; internal quotation marks and citations omitted).

Allentown took the last-mentioned of these options. According to the ALJ, it sought to show that it had an “objective” good-faith doubt primarily by presenting the testimony of Allentown managers, who, in turn, reported statements made to them by 14 employees. The ALJ set aside the statements of 5 of those employees as insignificant for various reasons—for example because the employees were not among the rehired 32, because their statements were equivocal, or because they made the statements at a time too long before the transition. 316 N. L. R. B., at 1206–1207. The majority does not take issue with the ALJ’s reasoning with respect to these employees. The ALJ then found that statements made by six, and possibly seven, employees (22% of the 32) helped Allentown show an “objective” reasonable doubt. Id., at 1207. The majority does not quarrel with this conclusion. The majority does, however, take issue with the ALJ’s decision not to count in Allentown’s favor three further statements, made by employees Marsh, Bloch, and Mohr. Id., at 1206–1207. The majority says that these statements required the ALJ and the Board to find for Allentown. I cannot agree.

Consider Marsh’s statement. Marsh said, as the majority opinion notes, that “ ‘he was not being represented for the $35 he was paying.’ ” Ante, at 9; 316 N. L. R. B., at 1207. The majority says that the ALJ was wrong not to count this statement in the employer’s favor. Ante, at 9–10. But the majority fails to mention that Marsh made this statement to an Allentown manager while the manager was interviewing Marsh to determine whether he would, or would not, be one of the 32 employees whom Allentown would re-employ. The ALJ, when evaluating all the employee statements, wrote that statements made to the Allentown managers during the job interviews were “somewhat tainted as it is likely that a job applicant will say whatever he believes the prospective employer wants to hear.” 316 N. L. R. B., at 1206. In so stating, the ALJ was reiterating the Board’s own normative general finding that employers should not “rely in asserting a good-faith doubt” upon “[s]tatements made by employees during the course of an interview with a prospective employer.” Middleboro Fire Apparatus, Inc., 234 N. L. R. B. 888, 894, enf ’d, 590 F. 2d 4 (CA5 1978). The Board also has found that “ ‘[e]mployee statements of dissatisfaction with a union are not deemed the equivalent of withdrawal of support for the union.’ ” Torch Operating Co., 322 N. L. R. B. 939, 943 (1997) (quoting Briggs Plumbingware, Inc. v. NLRB, 877 F.2d 1282, 1288 (CA6 1989)); see also Destileria Serralles, Inc., 289 N. L. R. B. 51 (1988), 882 F. 2d 19 (CA1 1989). Either of these general Board findings (presumably known to employers advised by the labor bar), applied by the ALJ in this particular case, provides more than adequate support for the ALJ’s conclusion that the employer could not properly rely upon Marsh’s statement as help in creating an “objective” employer doubt.

I do not see how, on the record before us, one could plausibly argue that these relevant general findings of the Board fall outside the Board’s lawfully delegated authority. The Board in effect has said that an employee statement made during a job interview with an employer who has expressed an interest in a nonunionized work force will often tell us precisely nothing about that employee’s true feelings. That Board conclusion represents an exercise of the kind of discretionary authority that Congress placed squarely within the Board’s administrative and fact-finding powers and responsibilities. See Radio Officers v. NLRB, 347 U. S. 17, 49–50 (1954). Nor is it procedurally improper for an agency, rather like a common law court, (and drawing upon its accumulated expertise and exercising its administrative responsibilities) to use adjudicatory proceedings to develop rules of thumb about the likely weight assigned to different kinds of evidence. Cf. Bell Aerospace, 416 U. S., at 294; Chenery, 332 U. S., at 202.

Consider next Bloch’s statement, made during his job interview with Worth, that those on the night shift (five or six employees) “did not want the Union.” 316 N. L. R. B., at 1207. The ALJ thought this statement failed to provide support, both for reasons that the majority mentions (“ ‘Bloch did not testify and thus could not explain how he formed his opinion about the views of his fellow employees’ ”), Ante, at 10; 316 N. L. R. B., at 1207, and for reasons that the majority does not mention (“no showing that [the other employees] made independent representations about their union sympathies to [Allentown] and they did not testify in this proceeding”). Ibid.

The majority says that “reason demands” that Bloch’s statement “be given considerable weight.” Ante, at 10. But why? The Board, drawing upon both reason and experience, has said it will “view with suspicion and caution” one employee’s statements “purporting to represent the views of other employees.” Wallkill Valley General Hospital, 288 N. L. R. B. 103, 109 (1988), enf ’d as modified, 866 F. 2d 632 (CA3 1989); see also Louisiana-Pacific Corp., 283 N. L. R. B. 1079, 1080, n. 6 (1987); Bryan Memorial Hospital, 279 N. L. R. B. 222, 225 (1986), enf ’d 814 F.2d 1259 (CA8 1987). Indeed, the Board specifically has stated that this type of evidence does not qualify as “objective” within the meaning of the “objective reasonable doubt” standard. Wallkill Valley General Hospital, supra, at 109–110 (finding that statement by one employee that other employees opposed the union “cannot be found to provide objective considerations” because statement was a “bare assertion,” was “subjective,” and “lacking in demonstrable foundation”; statement by another employee about the views of others was similarly “insufficiently reliable and definite to contribute to a finding of objective considerations”) (emphases added).

How is it unreasonable for the Board to provide this kind of guidance, about what kinds of evidence are more likely, and what kinds are less likely, to support an “objective reasonable doubt” (thereby helping an employer understand just when he may refuse to bargain with an established employee representative, in the absence of an employee-generated union decertification petition)? Why is it unreasonable for an ALJ to disregard a highly general conclusory statement such as Bloch’s, a statement that names no names, is unsupported by any other concrete testimony, and was made during a job interview by an interviewer who foresees a nonunionized workforce? To put the matter more directly, how can the majority substitute its own judgment for that of the Board and the ALJ in respect to such detailed workplace-related matters, particularly on the basis of this record, where the question of whether we should set aside this kind of Board rule has not even been argued?

Finally, consider the Allentown manager’s statement that Mohr told him that “if a vote was taken, the Union would lose.” 316 N. L. R. B., at 1207. Since, at least from the perspective of the ALJ and the Board, the treatment of this statement presented a closer question, I shall set forth the ALJ’s discussion of the matter in full.

The ALJ wrote,

“Should Respondent be allowed to rely on Mohr’s opinion? As opposed to Bloch who offered the opinion that the night shift employees did not support the Union, Mohr, as union steward, was arguably in a position to know the sentiments of the service employees in the bargaining unit in this regard. However, there is no evidence with respect to how he gained this knowledge, or whether he was speaking about a large majority of the service employees being dissatisfied with the Union or a small majority. Moreover, he was referring to the existing service employee members of the Mack bargaining unit composed of 32 employees, whereas the Respondent hired only 23 of these men. Certainly the composition of the complement of employees hired would bear on whether this group did or did not support the Union. He also was not in a position to speak for the 11 parts employees of Mack or the 7 parts employees hired by Respondent. Mohr himself did not indicate personal dissatisfaction with the Union.” Id., at 1208.

The ALJ concluded:

“Given the almost off-the-cuff nature of [Mohr’s] statement and the Board’s historical treatment of unverified assertions by an employee about other employees’ sentiments, I do not find that Mohr’s statements provides [sic] sufficient basis, even when considered with the other employee statements relied upon, to meet the Board’s objective reasonable doubt standard for withdrawal of recognition or for polling employees.” Ibid.

One can find reflected in the majority opinion some of the reasons the ALJ gave for discounting the significance of Mohr’s statement. The majority says of the ALJ’s first reason (namely that “there is no evidence with respect to how” Mohr “gained this knowledge”) that this reason is “irrelevan[t].” Ante, at 10. But why so? The lack of any specifics provides some support for the possibility that Mohr was overstating a conclusion, say, in a job-preserving effort to curry favor with Mack’s new managers. More importantly, since the absence of detail or support brings Mohr’s statement well within the Board’s pre-existing cautionary evidentiary principle (about employee statements regarding the views of other employees), it diminishes the reasonableness of any employer reliance.

The majority discusses a further reason, namely that Mohr was referring to a group of 32 employees of whom Allentown hired only 23, and “the composition of the complement of employees hired would bear on whether this group did or did not support the Union.” 316 N. L. R. B., at 1208. The majority considers this reason “wholly irrational,” because, in its view, the Board cannot “rationally” assume that

“the work force of a successor company has the same disposition regarding the union as did the work force of the predecessor company, if the majority of the new work force came from the old one,” ante, at 11.

while adopting an opposite assumption

“for purposes of determining what evidence tends to establish a reasonable doubt regarding union support,” ibid.

The irrationality of these assumptions, however, is not obvious. The primary objective of the National Labor Relations Act is to secure labor peace. Fall River Dyeing & Finishing Corp. v. NLRB, 482 U. S., at 38. To preserve the status quo ante may help to preserve labor peace; the first presumption may help to do so by assuming (in the absence of contrary evidence) that workers wish to preserve that status quo, see id., at 38–40; the second, by requiring detailed evidence before dislodging the status quo, may help to do the same. Regardless, no one has argued that these presumptions are contradictory or illogical.

The majority fails to mention the ALJ’s third reason for discounting Mohr’s statement, namely, that Mohr did not indicate “whether he was speaking about a large majority of the service employees being dissatisfied with the Union or a small majority.” 316 N. L. R. B., at 1208. It fails to mention the ALJ’s belief that the statement was “almost off-the-cuff.” Ibid. It fails to mention the ALJ’s reference to the “Board’s historical treatment of unverified assertions by an employee about other employees’ sentiments” (which, by itself, would justify a considerable discount). Ibid. And, most importantly, it leaves out the ALJ’s conclusion. The ALJ did not conclude that Mohr’s statement lacked evidentiary significance. Rather, the ALJ concluded that the statement did not provide “sufficient basis, even when considered with other employee statements relied upon, to meet the Board’s objective reasonable doubt standard.” Ibid. (emphasis added).

Given this evidence, and the ALJ’s reasoning, the Court of Appeals found the Board’s conclusion adequately supported. That conclusion is well within the Board’s authority to make findings and to reach conclusions on the basis of record evidence, which authority Congress has granted, and this Court’s many precedents have confirmed. See, e.g., Beth Israel Hospital v. NLRB, 437 U. S., at 504.

In sum, the majority has failed to focus upon the ALJ’s actual conclusions, it has failed to consider all the evidence before the ALJ, it has transformed the actual legal standard that the Board has long administered without regard to the Board’s own interpretive precedents, and it has ignored the guidance that the Board’s own administrative interpretations have sought to provide to the bar, to employers, to unions, and to its own administrative staff. The majority’s opinion will, I fear, weaken the system for judicial review of administrative action that this Court’s precedents have carefully constructed over several decades.

For these reasons, I dissent.