Does the National Labor Relations Board's finding that a successor employer lacked a good-faith reasonable doubt about the incumbent union's majority support, which would justify an employer poll of employees, find substantial support in the record as a whole?

Mack Trucks, Inc., operated an Allentown, Pennsylvania branch whose service and parts employees were represented by Local Lodge 724 of the International Association of Machinists. In May 1990, Mack notified its Allentown managers that it intended to sell the branch. Several managers formed Allentown Mack Sales & Service, Inc., which purchased the assets on December 20, 1990, and operated the branch as an independent dealership. Allentown hired 32 of the original 45 Mack employees.

Before and shortly after the sale, employees made statements to Allentown's prospective owners suggesting waning union support. Eight employees, in job interviews, made statements indicating they personally no longer supported the union. Ron Mohr, a union shop steward and bargaining committee member, told a manager that employees did not want a union and that, with a new company, the union would lose a vote. Kermit Bloch told a manager that the entire night shift did not want the union.

On January 2, 1991, Local 724 requested recognition and bargaining. Allentown refused on January 25, claiming a good-faith doubt about union support, and announced a secret-ballot poll under Board guidelines. A priest supervised the February 8, 1991 poll, which the union lost 19 to 13. The union filed an unfair-labor-practice charge.

The union filed an unfair-labor-practice charge with the National Labor Relations Board. An Administrative Law Judge concluded that Allentown was a successor employer to Mack Trucks and inherited a presumption of continuing majority union support. The ALJ held that the poll satisfied the procedural standards of Struksnes Construction Co. but violated sections 8(a)(1) and 8(a)(5) of the National Labor Relations Act because Allentown lacked an objective reasonable doubt about the union's majority status. The Board adopted the ALJ's findings and ordered Allentown to recognize and bargain with Local 724. Allentown petitioned the Court of Appeals for the District of Columbia Circuit, challenging both the facial rationality of the Board's polling standard and its application to the facts. The court enforced the Board's order over a dissent. 83 F.3d 1483 (1996). The Supreme Court granted certiorari, 520 U.S. 1103 (1997).

(Scalia, J.)

No. The Board's "reasonable doubt" standard for employer polls is facially rational and consistent with the Act, but the Board's finding that Allentown lacked a reasonable doubt is not supported by substantial evidence on the record as a whole.

On the facial challenge (Part II, a five-Justice majority), courts defer to Board requirements that are rational and consistent with the Act. Fall River Dyeing & Finishing Corp. v. NLRB, 482 U.S. 27 (1987). Using one unitary standard for polling, RM elections, and withdrawals of recognition is puzzling but not arbitrary or capricious. The Board may rationally limit polling because it is disruptive and unsettling, and polls retain value: they let an employer avoid antagonizing employees through abrupt withdrawal, they protect employers who want conclusive evidence before withdrawing, and a poll is less likely to draw a charge, especially if the union wins. The standard may also be lower than for RM elections because a union that loses a formal election is barred for a year, while a union that loses a poll may still request a Board election.

On substantial evidence (Part III, a five-Justice majority), "doubt" means uncertainty, not disbelief of the opposite. The question is whether a reasonable factfinder could find that Allentown lacked a genuine, reasonable uncertainty about majority support. It could not. Seven employees' statements supported doubt about their own support, and Marsh's statement of dissatisfaction engendered uncertainty. Bloch's and Mohr's statements about the workforce as a whole carried substantial weight absent reason to think they lacked a basis or were lying. The Board's rationale for discounting Mohr's statement, that he referred to Mack's former workforce, was irrational because the Board presumes continuity between predecessor and successor workforces when imposing the bargaining duty. If the union had majority support, the remaining 25 employees would have had to support it by about 17 to 8.

Part IV (a five-Justice majority) holds that the Board must apply in fact the standards it announces, such as good-faith reasonable doubt and preponderance of the evidence. Reasoned decisionmaking under the APA forbids applying a standard different from the one announced. Reviewing courts take announced standards to mean what they say, and no deference attaches to an agency's eccentric view of what a reasonable factfinder should demand. Prior Board practice of discounting such evidence does not alter the review standard. The judgment is reversed and remanded with instructions to deny enforcement.

Concurring in part and dissenting in part (Rehnquist, C.J., joined by O'Connor, Kennedy, and Thomas, JJ.) The Board's polling standard is neither rational nor consistent with the Act, so Part II is wrong, though Parts I, III, and IV are correct. The standard is so rigorous that it excludes most circumstantial and much direct evidence, leaving polling available only where it has almost no value, and it requires information the employer is barred from gathering by the most effective method. The majority answers that polls still have value, for example by letting an employer avoid the antagonism of an abrupt withdrawal of recognition, and that the unitary standard is not so irrational as to be arbitrary. The Act does not address polling, and the Board's authority rests on section 8(a)(1), yet a poll with Struksnes safeguards does not interfere with, restrain, or coerce employees. Polling promotes employee choice and the employer's legitimate interest in not bargaining with a nonmajority union. The Board also irrationally equates polls, RM elections, and withdrawals of recognition, which have different consequences. Under Gissel Packing and related First Amendment cases, employer solicitation of views is protected absent coercion, so the Board's interpretation raises serious constitutional concerns and receives no deference.

Concurring in part and dissenting in part (Breyer, J., joined by Stevens, Souter, and Ginsburg, JJ.) Parts I and II are correct, but Parts III and IV are wrong. The substantial-evidence question was not presented in the petition for certiorari, and the Court departs from the Universal Camera standard, under which courts intervene only where the standard was misapprehended or grossly misapplied. The Court recasts the Board's "objective reasonable doubt" standard as ordinary "uncertainty," dropping the word "objective" and ignoring settled deference to agencies' interpretation of their own rules. The ALJ reasonably discounted Marsh's statement as made in a job interview and as dissatisfaction rather than opposition, Bloch's statement as conclusory and unsupported, and Mohr's statement as unverified and off-the-cuff. The Board may use adjudication to develop rules of thumb about the weight of evidence, and the Court's ruling weakens the system of judicial review of agencies. The majority answers that the question presented fairly covers whether the Board erred in finding the evidence insufficient, that "objective" concerns the source of the considerations, not their reliability, and that the Board may raise evidentiary standards only by explicit rule or announcement.