National Labor Relations Board v. Magnesium Casting Company, Inc.

BREYER, Circuit Judge. This case arises out of a hotly contested union effort to organize the maintenance and production employees of Magnesium Casting Company, Inc. (the “Company”). The National Labor Relations Board found that the Company committed numerous unfair labor practices during the course of the election campaign. 1 Since the Company had shown “pervasive and adamant hostility to the representation of its employees” by the union, and since 90 of 178 employees in the bargaining unit had signed union authorization cards, the Board ordered the Company to recognize and bargain with Local 262 of the United Electrical Radio and Machine Workers of America (the “Union”) without an election. See NLRB v. Gissell Packing Co., 395 U.S. 575, 89 S.Ct. 1918, 23 L.Ed.2d 547 (1969). The Board here petitions for enforcement of its order, which the Company resists on two grounds. First, the Company objects to the Board’s unfair labor practice findings regarding the discharge of five employees. Second, the Company claims that several of the authorization cards were invalid. We find all of the challenged Board findings supported by substantial evidence, and we therefore enforce its order. Magnesium Casting Company, located in Hyde Park, Massachusetts, makes metal castings and electrical fixtures. It has about 180 maintenance and production employees, a significant number of whom are Spanish, Greek or Puerto Rican and neither speak nor read English. The union drive underlying this case began in the summer of 1977. It was initiated by an employee in the Company’s die-casting department. That department operates 24 hours a day in three eight-hour shifts. A foreman, an assistant foreman, and about ten other employees work on each shift. The department included many of the Union’s most active supporters, and its operations and procedures figure prominently in this case. The lengthy record of this case provides two important items of background information about the Company’s work practices against which the specific claims at issue must be judged. First, the Company has strict rules against absenteeism and tardiness, but those rules are often honored in the breach. The Company deals with the somewhat erratic work habits of its employees pragmatically. It routinely issues warnings; it occasionally suspends violators; sometimes it discharges them; but then usually it hires them back at some later time. This is not to say that the Company is lax; but rather, given the nature of the work or work force, attendance problems are a fact of life dealt with by the Company through rules that are firm in principle but applied flexibly in practice. Second, the Company operates an incentive compensation system which, although simple in principle, appears on this record to be complex, confusing and subject to widespread (but largely tolerated) abuse and misunderstanding in practice. Essentially, the system seeks to pay workers more if they make more castings. Given the practical impossibility of counting every casting, however, incentive pay is based on the number of cycles a worker causes his machine to complete rather than on the number of castings he actually produces. Machine cycles are recorded on meters linked to each machine but located at a foreman’s desk. When working under the incentive plan, a die-caster’s pay depends on the meter count; when not working under the plan, pay depends simply on time spent on the job. *15 The complications and confusion stem in part from the' ability of employees to make the meters record completed cycles that do not actually occur. Yet whether a worker has in fact “cheated” in this fashion is rarely easy to ascertain. “Metered” and “actual” production often differ simply because castings turn out defective and are thrown back and remelted before being counted in an inspection. Even when an employee has “cheated” by inflating a “meter” count, it is unclear that he has violated a norm of the workplace taken seriously by the Company. Employees are almost always warned of impending inspections before they occur, and violations almost never lead to disciplinary action, let alone dismissal. Finally', as it appears to be common knowledge within the Company that some machines work better than others, “fast” employees assigned to “slow” machines sometimes suspect that this has been done deliberately to hold down their wages. 2 Then, in their view, a little cheating is “fair game.” Difficulties with the incentive compensation system also result from the fact that, on any given machine on any given shift, an employee is not always on or always off the incentive system. If the machine works properly, it seems, an employee is supposed to stay on the system, work hard, and be paid according to his productivity. But if the machine breaks down, the employee is supposed to get off the system and be paid by time rather than by productivity. Switching from one pay system to the other involves punching a certain set of cards in a certain sequence. The precise mechanics of this process, however, seem confusing. Neither the AU nor counsel for either side in this case — let alone the die-cast operators themselves — appears fully to have understood them. The Company claims that by punching cards in a certain way, an employee can increase his compensation. The testimony of the Company’s own witness, however, revealed that attempts to cheat by mispunching cards are easily detectible and unlikely to succeed. In any event, the mispunching of cards is a common occurrence among die-cast operators, and employees are virtually never discharged, if even disciplined at all, for that reason. With this background revealed by the record, we have reviewed the careful and detailed opinion of the administrative law judge — which was adopted by the Board. In checking the opinion against the record, we have found it accurate and its conclusions supported by substantial evidence. NLRB v. Amber Delivery Service, 651 F.2d 57 (1st Cir. 1981). II We turn first to the challenged finding of anti-union discriminatory behavior in violation of § 8(a)(3) of the Act. 3 The behavior at issue consists of the Company’s discharging five employees: Walter Escribano, Leonard Lamkin, Sherman Brown, James Eason, and John Manoloulis. We have held that the standard in such cases is whether the discharges were motivated by anti-union animus rather than legitimate business concerns. In other words, to find that a discharge violated the Act, the Board must reasonably conclude that the discharge would not have taken place in the absence of the anti-union motive. See *16 NLRB v. Eastern Smelting and Refining Corp., 598 F.2d 666, 671 (1st Cir. 1979). In so-called “mixed motive” cases, we have more recently held that, once the Board makes a prima facie showing of significant improper motivation, the employer must produce evidence that it had a legitimate reason, sufficient in itself, for the discharge. NLRB v. Wright Line, A Division of Wright Line, Inc., 662 F.2d 899, 904-05 (1st Cir. 1981). “The imposition of this limited burden, however, does not shift to the employer the burden of proving that an unfair labor practice has not occurred.” Id. at 905. That burden always remains with the Board. Id. 4 Although Wright Line was decided after the Board’s decision here, we see no need for a remand of this case. The ALJ, and hence the Board, did not decide this case upon the niceties of burdens of proof. Rather, in the case of each discharge, the ALJ found evidence of anti-union motive and then went on to find that the employer’s allegedly legitimate reason for the discharge was a pretext. The evidence is set out at length in • the ALJ’s opinion; we discuss it briefly here to indicate why we feel the ALJ’s findings are adequately supported. Walter Escribano. Escribano was fired in October 1977, allegedly for inflating his production count. The Company claims that it did not know of Escribano’s Union activities when it fired him, but there is significant evidence to the contrary. Escribano joined the Union’s organizational effort in August 1977. By October he had attended several union meetings, signed an authorization card, distributed union leaflets at the plant, and was soliciting support for the union almost every day. He testified that on one occasion he told another employee about a union meeting in the presence of a foreman. This latter fact distinguishes this case from NLRB v. South Shore Hospital, 571 F.2d 677 (1st Cir. 1978), where there was no evidence that management had ever directly observed the em~ ployee soliciting union support. See id. at 684; see also id. at 683 (discussing NLRB v. Malone Knitting, 358 F.2d 880 (1st Cir. 1966)). And, the former facts, showing significant union involvement, distinguish this case from NLRB v. Joseph Antell, Inc., 358 F.2d 880 (1st Cir. 1966), where the employee’s only union activity consisted of attendance at a single union meeting held off the employer’s premises at night. Finally, at least as corroboration it is reasonable to suppose that an employer would know who the union activists are in an overt, strongly resisted union campaign. This is so particularly when they solicit at the plant, as Escribano did, during coffee breaks and before their work shifts almost every day. There is also sufficient evidence to support the ALJ’s conclusion that the “cheating” charge was a pretext for firing Escribano. For one thing, the evidence that Escribano actually cheated is weak. A hand count showed fewer castings than registered on his meter, but Escribano and another employee testified (and the foreman’s information sheet showed) that the casting machine was not working properly and produced numerous defective castings that had to be thrown back. The Company’s knowledge of this type of problem, and the ease with which it could have checked the machine, cast doubt on its belief that Escribano had in fact cheated. Cf. Raytheon Co. v. NLRB, 326 F.2d 471, 475 (1st Cir. 1964). For another thing, the Company’s flexible approach to these problems in the past makes it unlikely it would have fired Escribano in the absence of an anti-union motive. No one had been fired for inflating production counts — a fairly common occurrence — in the ten years preceding Escribano’s discharge. Moreover, Escribano was one of the firm’s top machine operators and had only one disciplinary mark on his record (a warning for absenteeism) in three years of uninterrupted employment. Further, *17 when Escribano asked a Company vice president why he had been fired, the vice president said that the Company was “up to here” with him. The AU reasonably interpreted this as a reference to Escribano’s union activity, for (aside from the single warning for absenteeism and the alleged instance of cheating) what else had he done so to offend the Company? The strongest evidence for the Company is that on the same day it fired Escribano, it fired another employee, who was without union sympathies, also allegedly for cheating. Yet weighing this fact against the type of cheating involved, the firm’s past practices, Escribano’s good past record, Escribano’s union activism, and the vice president’s remark, we find the ALJ’s conclusion (based on his hearing the witnesses) is adequately supported. Leonard Lamkin. Lamkin was discharged on November 8, 1977. The evidence of anti-union animus includes the following: Lamkin began the Union’s organizing effort in June 1977. He actively and visibly promoted the Union. One week after the first meeting of employees with a union representative in August, a meeting that Lamkin organized, the Company warned him about his absences and denied him a promotion. One week after a meeting to protest the firing of Escribano, a meeting that Lamkin helped organize and at which he spoke out, the Company fired him. The Company claims Lamkin was fired because of his absences and because he incorrectly claimed he was sick. Indeed, he failed to show up at work quite a few times during the months before he was fired; other die-easting employees with comparable records of absenteeism had been fired in the past; and Lamkin himself produced medical records showing no discernible cause for his stomach pains. The ALJ, however, rejected the Company’s claim as a pretext. Our scrutiny of the record has revealed powerful reasons to sustain that conclusion. First, the Company did not consistently fire employees with spotty attendance records. The Company normally gave such employees many written warnings, almost always including one “final” warning, and very often further warnings after the “final” warning before discharge. Moreover, the Company usually treated personal problems, such as stomach pains, as mitigating circumstances, and it deliberately sought to retain competent machine operators, such as Lamkin, despite absences. It seemed to treat Lamkin, in the ALJ’s view, more harshly than other comparable workers. Second, the procedural circumstances of Lamkin’s discharge were unusual. Before firing him, the Company did not give Lam-kin a “final” warning, nor did it consult Lamkin’s foreman. Rather, a Company vice president announced Lamkin’s discharge at a meeting of production employees about a week after the Escribano protest meeting, and a superintendent later that day sent Lamkin a telex discharging him. Third, it seems as though the Company attempted to “mousetrap” Lamkin. The day he was fired, Lamkin came to the Company to find out why. He explained to Company officials — his superintendent and a vice president — that his absences were due to stomach pains. According to uncontradicted testimony, 5 the vice president then told Lamkin the Company did not want sick people at the plant but that he would be rehired if he could prove he was not sick. The next week, Lamkin produced records of previous physical exams showing he had complained to doctors about stomach pains but that the doctors could discern no physical cause. The vice president declared the records showed Lamkin had no excuse for his absences, told him he was fired, and then refused to talk with him further. *18 Given Lamkin’s union activities, past Company discharge practices, the timing of the discharge, the lack of final warning, lack of foreman consultation, and the apparent attempt to get Lamkin to certify, in effect, that his own absences were without justification, we believe the ALJ’s conclusion of “pretext” was adequately supported. Sherman Brown/James Eason. It is undisputed that Brown and Eason, who were discharged in July 1978, were active and visible Union supporters. The only issue is whether the Company’s alleged reason for the discharge — unexcused absences — was a pretext. The Company correctly points out that Brown and Eason left work a day and one-half before the summer vacation began and failed to report to work the day after it ended. The ALJ, however, reasonably found that Brown and Eason were longtime, competent employees with whom the Company was typically lenient. He also reasonably found that the prevacation absences occurred under unusual circumstances and were not taken seriously by the Company. The 1978 summer vacation for the entire plant was initially scheduled to begin on Wednesday. One week before the scheduled beginning, the Company told its die-casting employees that they would have to stay until Friday evening. Several employees, including Brown, complained and said they had already made plans. According to testimony reasonably credited by the ALJ, Brown’s foreman told Brown he could begin his vacation as planned, and, on Thursday, the assistant foreman told Eason he could leave early as well. The entire first shift, including Brown and Eason, punched out at noon Thursday, three hours early. Brown, Eason and six other die-casters also did not report to work on Friday. The Company took no action against the other six — and its explanations of the differences are flimsy. The Company argues that of the eight die-casters absent on the Friday before vacation, only Brown and Eason also failed to show up for work the day after vacation. But the ALJ found that when Brown returned to work and learned he had been fired, the die-casting superintendent told him that “maybe all this wouldn’t have been happening” if “you people” had not tried to get the Union into the plant, and that Brown should “come back to see” him after all the “bullshit” was over. Given this statement, and the other evidence of Company practices, the ALJ’s finding that these discharges did not flow from legitimate business motives is reasonable. John Manoloulis. The Company discharged Manoloulis in April 1978, allegedly for cheating by mispunching cards. There is ample evidence to make out a prima facie case of anti-union motive. James Skafidas, an assistant foreman, hired Manoloulis to work in the die-casting department in February. On April 24 the Union distributed leaflets written in Greek, the only language Manoloulis understands. Skafidas then asked the Company’s Greek employees if they were going to sign the cards. They were uncertain. Skafidas later asked Manoloulis in private what he intended to do. Manoloulis said he would sign unless the Company reversed a prior decision denying him a dollar per hour raise. According to Manoloulis, Skafidas then indicated that the Company would not give him a raise, made an obscene suggestion about what to do with the union card, said the Company would close its doors before allowing a union, and told Manoloulis that if the Union succeeded, he would personally kick Manoloulis out. In context, the ALJ, who credited Manoloulis, could reasonably find that Skafidas’ statement reflected more than ire based on Manoloulis’ effort to get a raise. The next day Skafidas fired Manoloulis. There is also adequate evidence to support the ALJ’s finding that the Company’s alleged basis for firing Manoloulis — cheating by mispunching cards — was a pretext. First, the punching system was complicated. Manoloulis’ foreman testified that he had found Manoloulis mispunching cards on several occasions. The foreman, however, did not know why Manoloulis persisted in mispunching cards, and repeatedly refused to characterize the mispunching as “cheating.” *19 Second, on the day in question Skafidas admitted that Manoloulis had been assigned to an “average pay machine.” Manoloulis thus could not have affected his pay, even in theory, by mispunching the cards. 6 Third, mispunching occurred frequently, yet no other employee had ever been fired as a result. While the Company has produced a complicated series of inferences that would warrant a finding of a legitimate business' motive, the evidence tilts decidedly in favor of the ALJ’s conclusion to the contrary. ‘ill We turn now to consider the Board’s issuance of a Gissel order requiring the Company to recognize and bargain with the Union without an election. The Company challenges that order on the ground that .it is based upon an erroneous finding that a majority of employees — 90 out of 178 — authorized the Union to represent them. The Company first claims that three discharged employees — Escribano, Lamkin, and Manoloulis — should not have been counted. But it concedes that if the discharges were improper, as we have held they were, the ALJ’s decision to count the union cards of these employees was appropriate. Second, the Company claims that the ALJ should not have excluded from the bargaining unit (for purposes of determining majority status) certain employees who had not started working as of the eligibility date. 7 It is well settled “that an individual must be employed and working on the established eligibility date in order to be eligible to vote” in a union election. NLRB v. Dalton Sheet Metal Co., 472 F.2d 257, 258 (5th Cir. 1973) (citations omitted) (emphasis in original). Conditioning the right to vote on presence at the workplace ensures that the employee is exposed to the arguments for and against unionization. For the same reason, it makes sense to. require presence at the workplace as of the date a majority status determination is made. Such a determination serves as a rough substitute for a union election — a sort of second-best mechanism for assessing the extent of union support among a company’s employees. Employees who have not submitted union cards by the critical date are, in some sense, treated as though they had voted against the union. The negative inference is plausible in the case of working employees, for such employees have presumably been exposed to the arguments for and against unionization. A nonworking employee, however, is less likely to have been exposed to these arguments, and his failure to submit a union card might well reflect nothing more than the lack of a chance to consider the matter. Accordingly, we see nothing wrong with excluding hired but nonworking employees from a majority status count as the Board did here. See also Western Drug, 231 N.L.R.B. 890, 891 (1977), enf. denied on other grounds, 600 F.2d 1324 (9th Cir. 1979). 8 *20 Third, the Company specifically challenges twelve union authorization cards as invalid. We deal with each challenge in turn. a. The Company objects to seven cards — those of Geovanney Olaverria, Kleantis Manoloulis, George Anastasiadis, Charles Trivoskovis, Ovadis Santana, Miguel Olaverria, and Nino Piccirilli — on the ground that the signers 9 allegedly did not believe they were authorizing the Union to represent them. They believed, the Company claims, they were only asking for an election. The Company’s claim is initially belied by the fact that the card itself states, I hereby request and accept membership in the above named union, and authorize it to represent me. . . . And, those employees who do not read English testified either that the card was translated for them or they understood its purport. We recognize, of course, that if the express terms of the card were “deliberately and clearly cancelled by a union adherent with words calculated to direct the signer to disregard and forget the language above his signature,” NLRB v. Gissel Packing Co., 395 U.S. at 606, 89 S.Ct. at 1936, the card would be invalid. For the card to be invalid, we also recognize that the employee need not have been “expressly told in haec verba that the ‘sole’ or ‘only’ purpose of the cards is to obtain an election.” Id. at 608 n.27, 89 S.Ct. at 1937 (quoting with approval the Board’s own statement of the so-called Cumberland Shoe doctrine in Levi Strauss & Co., 172 N.L.R.B. No. 57, 68 L.R.R.M. 1338, 1341-42 n.7 (1968)). The “circumstances surrounding the card solicitation,” however, must “add up to an assurance to the card signer that his card will be used for no purpose other than to help get an election.” Id. (same). The ALJ reasonably concluded that the language of the card had not been so cancelled here. The only evidence of cancellation in respect to Geovanney Olaverria, Kleantis Manoloulis, George Anastasiadis, and Charles Trivoskovis consists of their testimony that they had a general understanding that signing a card would lead to an election. None claimed to have been told, and there is no other evidence showing any were told, that the cards would be used only to secure an election. In fact, Olaverria said that a union solicitor told her she would receive more benefits by joining the union. Anastasiadis indicated he was aware that signing the card would make him a union member. Manoloulis and Trivoskovis, who, along with Anastasiadis, do not read English, each testified that trusted relatives translated the card. Cf. A.J. Krajewski Manufacturing Co. v. NLRB, 413 F.2d 673, 677 (1st Cir. 1969) (as to employees not fluent in English and thus unable to read union cards, “the issue is whether the purpose of the cards was adequately communicated to the signers”). Ovadis Santana testified that the union representative who solicited her card told her that “no one was going to find out, and that there would be an election.” In Gissel itself, the Supreme Court upheld the validity of cards signed by employees who had been told just this. See NLRB v. Gissel Packing Co., 395 U.S. at 584 n.5, 608-09, 89 S.Ct. at 1924, 1937. Santana also testified that she did not read the card and signed it to rid herself of union solicitors. The failure to read a card, however, does not invalidate it, see Tipton Electric Co. v. NLRB, 621 F.2d 890, 896 (8th Cir. 1980), and “an employee’s thoughts (or afterthoughts) as to why he signed a union card, and what he thought that card meant, cannot negative the overt action of having signed a card designating a union as bargaining agent.” Joy Silk Mills, Inc. v. NLRB, 185 F.2d 732, *21 743 (D.C.Cir.1950), cert. denied, 341 U.S. 914, 71 S.Ct. 734, 95 L.Ed. 1350 (1951), quoted with approval in Tipton Electric Co. v. NLRB, 621 F.2d at 898. See also NLRB v. WKRG-TV, Inc., 470 F.2d 1302, 1318 (5th Cir. 1973). Miguel Olaverria gave testimony very similar to that of Santana. In addition to claiming he did not read the card and that he signed it to get rid of the solicitor, he said the solicitor told him that “it would not imply anything to sign the card, because we were going to have a free election.” The solicitor denied saying this, but, regardless, in context it does not amount to an assurance that the card will be used only to secure an election. Olaverria acknowledged that before signing a card he had discussed the Union with solicitors on numerous occasions in the past, that he had told the solicitors of rumors he had heard at the Company about employees who were for or against the Union, and that a solicitor had told him that the purpose of getting the Union into the Company was to secure more benefits for the employees. Only one of the seven — Nino Piccirilli— actually claimed a union solicitor told him that his card “was for the purpose of [an] election only.” The solicitor denied saying this and was credited by the ALJ. Such credibility determinations, of course, are for the Board rather than for us to make, and they stand unless beyond “the bounds of reason.” NLRB v. New England Lithographic Co., 589 F.2d 29 (1st Cir. 1978). Those bounds were not exceeded here. It is true, as the Company notes, that Piccirilli signed but did not otherwise fill out (and claims he did not read), the union card bearing his name, and that the Board failed to examine an available witness who might have been able to rebut Piccirilli’s version of his conversation with the solicitor. That Piccirilli may not have read the union card before signing it is not significant under the appropriate legal standard. See Tipton Electric Co. v. NLRB, 621 F.2d at 896. That he did not fill out the card except for signing it may reflect haste or laziness on his part but does not show that the solicitor assured him that his card would be used only to secure an election. And, finally, although the witness to the Piecirilli-solicitor conversation testified as to other matters, neither the Board nor the Company examined him on the contents of that conversation. The failure to examine the witness on that subject leaves a gap in the evidence, but the gap does not support Piccirilli’s version of the conversation any more than it does the solicitor’s. b. The Company objects to the cards of four employees — Miguel Ortiz, Saturnino Vega, Ramon Martinez and Maria Sanchez — who did not personally sign the cards bearing their names. The fact that they did not sign themselves proves little, for none of the four can read and only one of them writes well enough to sign his own name. Whether they authorized others to sign was the subject of conflicting testimony. Ortiz denied giving anyone permission to sign a card for him. A union solicitor testified that Ortiz, on the way home from work, asked the solicitor to sign a card for him, that this was done in the presence of a friend of Ortiz, and that Ortiz said the matter would stay among the three of them and the Company would not find out. The ALJ credited the solicitor over Ortiz because of inconsistencies in Ortiz’ testimony and his apparent fear of Company reprisals. The ALJ’s determination is within “the bounds of reason.” NLRB v. New England Lithographic Co., 589 F.2d at 29. Vega also denied giving anyone permission to sign a card for him. A union solicitor testified that he and another solicitor visited Vega in his home, talked about the union, and explained the card. He added that a woman present at the meeting, apparently a friend of Vega’s, then filled out and signed the card for Vega and gave it to the solicitor. The Company’s challenge rests on the fact that the solicitor initially testified that Vega signed the card but later testified that the woman signed the card when it became apparent that Vega had not done so. The record, however, does not require accepting the Company’s claim *22 that the “mysterious woman” was “inventad]” by the solicitor. The dispute over Vega’s card involves at most simply a matter of conflicting testimony. In context the ALJ’s rejection of the Company’s claim was reasonable. A union solicitor also testified without contradiction that he and a Company employee visited Martinez at his home, explained the union, and gave him a card. The solicitor added that Martinez gave the card to his wife to fill out for him, that his wife did so in his presence and gave it to the solicitor. The Company claims that neither Martinez nor his wife signed the card bearing his name. This claim is based on discrepancies between the information and signature on the union card and on a tax withholding form that Martinez apparently signed. In view of Martinez’ virtual illiteracy, it is hard to say what the discrepancies show. The inferences the Company would have us draw are farfetched, and were reasonably rejected by the ALJ. Sanchez also denied authorizing anyone to sign a card for her. A Company employee named Rafaela Ruiz, however, testified that she gave Sanchez a union card one morning and that Sanchez returned the card signed later that day. Ruiz testified that Sanchez called her before she (Sanchez) was to testify and said, “I know I gave you the card and the signature was of somebody else, but I gave consent.” The ALJ credited Ruiz over Sanchez because of inconsistencies in Sanchez’ testimony and her apparent fear of getting into trouble with the Company. We have examined the inconsistencies in light of the Company’s arguments and conclude that the ALJ’s judgment was reasonable. c. Finally, the Company challenges a card signed by James LaCour. LaCour first testified that his wife had signed the card. Then on cross-examination he admitted he had signed it but added that his wife had printed his name at the bottom of the card and had given the card to the Union without his permission. A union solicitor testified that he received the card from LaCour himself. The ALJ’s decision to credit the solicitor and not LaCour on this point was reasonable. IV It is obvious from the discussion that the ALJ’s decision depended largely upon credibility judgments and upon judgments about the normal expectations of the workplace and the atmosphere in which work at Magnesium Casting Company was conducted. In sustaining the Board, we were impressed by the apparent care, thoroughness, and evenhandedness with which the ALJ wrote his opinion. We find no evidence of the prejudice or the bias on the part of the ALJ, or the conspiracy on the part of the Union, that acceptance of the Company’s arguments would imply. Affirmed and enforced. Footnotes [1] Specifically, the Board found the Company had threatened plant closure and other reprisals, had coercively interrogated employees, had engaged in surveillance and had created the impression of surveillance, had directed employees not to sign union cards, and had maintained an unlawfully broad anti-solicitation and distribution rule, all in violation of § 8(a)(1) of the National Labor Relations Act (the “Act”), 29 U.S.C. § 158(a)(1). The Board also found the Company had discharged five employees, suspended seven and made working conditions more onerous for one — all with the aim of discouraging union activity and all in violation of either § 8(a)(3) or § 8(a)(1) of the Act, 29 U.S.C. §§ 158(a)(3) and (a)(1). [2] According to the testimony of one of the discharged employees, Walter Escribano, “incentive” pay is based on weekly rather than daily figures. Thus, for the hours spent on the incentive plan in a given week, an employee is paid a wage that exceeds the standard wage by the same percentage that the employee’s hourly productivity — averaged over the hours worked on the system that week — exceeds the standard productivity. Escribano suggested that an employee whose productivity was high during the first part of a week was often assigned to a defective or slow machine during the latter part of the week so that his productivity on average, and hence his pay, would be just about average. [3] National Labor Relations Act § 8(a)(3), 29 U.S.C. § 158(a)(3), provides that: (a) It shall be an unfair labor practice for an employer (3) by discrimination in regard to hire or tenure of employment or any term or condition of employment to encourage or discourage membership in any labor organization. [4] Cf. NLRB v. Amber Delivery Service, 651 F.2d 57, 69 (1st Cir. 1981) (employer must merely “come forward with enough evidence to convince the trier of fact that, under the circumstances, there is no longer a preponderance of evidence establishing a violation”); NLRB v. Cable Vision, 660 F.2d 1, 8 (1st Cir. 1981) (same). [5] The testimony was given by Lamkin. The vice president also testified, but not about his conversation with Lamkin. The superintendent testified as well, however, and he claimed to have asked Lamkin to produce records showing he was sick on the days he had been absent. Nevertheless, the superintendent does not in terms deny that the vice president said what Lamkin claimed he said. [6] The Company suggests that neither Skafidas nor - Manoloulis realized that Manoloulis had been assigned to an average rather than an incentive job, implying that Manoloulis had had a motive to cheat and that Skafidas could reasonably have feared his attempt to do so. The suggestion is implausible and, in any event, not borne out by the record. Skafidas’ testimony, if anything, shows that he knew Manoloulis was assigned to an average pay machine the day he was fired. Manoloulis’ testimony is simply silent on this point. [7] Each of the three had been hired sometime before the eligibility date and each began work the day after that date. There is no evidence in the record that the Union was aware of these employees or had an opportunity to solicit their support before the eligibility date. [8] The Company points out that the employee in Western Drug did not begin working until more than two weeks after the critical date, while here the employees began work the next day. Since the critical factor is exposure to the relevant unionization arguments, this difference is not significant. Similarly, we find no unlawful inconsistency, see Sunbeam Television Corp. v. FCC, 243 F.2d 26 (D.C.Cir.1957), arising out of the fact that the Board in Riviera Manor Nursing Home, 200 N.L.R.B. 333 (1972), included in a “majority status” group employees who had not yet begun working while the Board here did not include such employees. In Riviera Manor, unlike this case, the employees at issue were known to both union and company and had been exposed to the relevant arguments on both sides of the unionization debate. The fact that the Board decided to include employees *20 where there was exposure to argument does not suggest that the Board must include them where there was not. [9] Geovanney Olaverria’s card was actually signed by her husband, Miguel, rather than by her personally. The Company does not challenge the card on that ground, however, for it is undisputed that Geovanney expressly directed her husband to sign the card for her. We therefore treat the card as though she had signed it herself.