Filed 8/18/21 The Regents of the U. of Cal. v. Pub. Employment Relations Bd. CA2/8

NOT TO BE PUBLISHED IN THE OFFICIAL REPORTS

California Rules of Court, rule 8.1115(a), prohibits courts and parties from citing or relying on opinions not certified for publication or ordered published, except as specified by rule 8.1115(b). This opinion has not been certified for publication or ordered published for purposes of rule 8.1115.

IN THE COURT OF APPEAL OF THE STATE OF CALIFORNIA

SECOND APPELLATE DISTRICT

DIVISION EIGHT

THE REGENTS OF THE B305934

UNIVERSITY OF CALIFORNIA,

(PERB Dec. No. 2704-H, Petitioner, Case Nos. LA-CE-1291-H &

LA-CE-1292-H)

v.

PUBLIC EMPLOYMENT

RELATIONS BOARD,

Respondent;

MANUEL SALDIVAR et al.,

Real Parties in Interest.

ORIGINAL PROCEEDING; petition for writ of extraordinary relief from a decision of the Public Employment Relations Board. Petition denied. Paul, Plevin, Sullivan & Connaughton, Sandra L. McDonough and Karyn R. Moore for Petitioner. J. Felix de la Torre, Wendi L. Ross, Laura Z. Davis and Matthew B. Seipel, for Respondent. Mahoney Law Group, Kevin Mahoney and John A. Young, for Real Parties in Interest. ____________________ A three-member panel of the Public Employment Relations Board (the Board) unanimously found the Regents of the University of California (the University) unlawfully fired two employees because of their union activities. The University attacks the Board’s treatment of an administrative law judge’s credibility findings and says the Board’s decision lacks substantial evidence. The University petitioned for writ of extraordinary relief. We deny the petition. Undesignated statutory citations are to the Government Code. I This case involves the University’s termination of Manuel Saldivar and Victor Flores, two employees of the plumbing shop at the University of California Los Angeles (UCLA). Saldivar, a plumber, and Flores, a facilities worker, had worked for the University for about a decade each. In November 2016, Saldivar and Flores participated in union activities, including a demonstration and a strike. Later that month their supervisor, Tim Moore, found an anomaly in their timekeeping records. Moore elevated the issue to his superiors without informally meeting with Saldivar or Flores. The University immediately placed Saldivar and Flores on investigatory leave and later terminated them. Saldivar and Flores brought unfair practice charges against the University alleging retaliation. The University says it investigated and fired Saldivar and Flores not because of their

2 union activities, but because the employees knowingly engaged in timecard fraud. An administrative law judge heard the case and wrote a proposed decision in favor of the University. Saldivar and Flores filed exceptions, which triggered review by the Board. The Board disagreed with the administrative law judge and reversed the proposed decision. The University petitioned for writ of extraordinary relief asking us to overturn the Board’s decision. The Board filed a respondent’s brief, which Saldivar and Flores joined. Now we delve more deeply into the facts of Saldivar’s and Flores’s union activities, the timekeeping issue, the University’s investigation, and the decisions of the administrative law judge and Board. A Saldivar and Flores were active members of their union. Saldivar participated in 2016 contract negotiations with the University. He had been a liaison between the University’s bargaining team and union members in the plumbing shop for four years. On November 4, 2016, Saldivar and Flores joined a union demonstration on campus during their lunch hour. The demonstration was public and outdoors. Demonstrators held signs advocating for fair wages. The union created a flyer to advertise a strike on November 16, 2016. The flyer prominently featured a photo of Saldivar, Flores, and two other picketers from the November 4 demonstration. Saldivar and others posted the flyer throughout campus. Saldivar placed copies in the plumbing shop on tables,

3 on a corkboard by the entrance of Moore’s office, and in a tool room near where management sits. Saldivar testified he saw Moore take the flyer off the corkboard. Moore told him the union could not post anything advertising the strike in the shop. Flores saw the flyer posted at the start of a morning shift and it was gone later that day. Flores did not see who took it down but he heard Moore was responsible. Moore admitted there had been, “a Union pre-strike [flyer] or something regarding the strike and Union business that was placed on the overtime board, the University bulletin board and the safety board and various places throughout the shop, on my office door. And I took those down and put them on the table in the shop.” Moore said this was a different flyer, not the one with Saldivar and Flores’s picture from the demonstration. Moore could not recall specifically what was on the flyer he took down. The University did not offer another flyer in evidence. The union went forward with the November 16 strike. Saldivar, Flores, and all of the plumbers participated. According to a longtime employee of UCLA, this was the first time in at least 28 years skilled trades workers participated in a strike. B We turn to the timekeeping procedures. Plumbers and facilities workers record time electronically. They clock in by swiping their badges on a wall clock or on a “wand,” then entering the work order number for a particular job and swiping their badge a second time. The plumbing shop serves the campus as well as certain off-campus housing and medical centers. People from different departments can report plumbing issues by calling the “Trouble Desk.”

4 Shop employees work standard 40-hour weeks and sometimes have scheduled overtime. Another scenario is “call back.” In brief, call back is when employees are called in to work outside their usual work time. The University’s time fraud claims are based on Saldivar’s and Flores’s use of call back. The contract covering plumbers and facility workers provides a minimum of four hours’ pay for a call back. The contract defines “call back” as “those instances when an employee is ordered back to work without prior notice after completing a shift and leaving the premises or those instances when prior notice is given but the work begins at least three hours after the completion of the regular work schedule.” The call back policy is logical. Havoc-raising clogs and leaks do not follow a nine-to-five schedule. As anyone with a burst pipe will attest, on-demand plumbing is essential. A quick plumbing fix does not necessarily diminish the fix’s value. On call and unscheduled work is inconvenient for workers, though. It can mean a rearranged personal life, disrupted sleep, or fatigue from extra work after having completed a lengthy shift. The call back policy compensates for the inconvenience and creates incentives for this work by guaranteeing substantial compensation. The shop has a written “Call Back Procedure.” The written policy says, “Once you arrive on campus swipe in. Call Trouble Desk @ 68352 or go by the trouble desk, let them know you are here and get the work order number. Use the call back mode on the wand or wall clock when you swipe in on the work order number.” Other steps of the written procedure give specific instructions about how to handle the call like, “DO NOT TURN

5

OFF WATER WITHOUT CONSULTING BUILDING MANAGER

. . . . Redirect water using plastic to safe drainage location if necessary,” call the locksmith if there may be water damage in locked rooms, and phone “Advanced Sewer Technology” if there is an especially unwieldy stoppage. The final step of the procedure directs employees to call the Trouble Desk to say they are finished and will be leaving campus, “unless there are any other calls that need to be attended to.” This step envisions the possibility of a second call back job but does not offer directions about whether employees should log a successive job as call back. Saldivar and Flores testified the shop supervisors had not trained or counseled them regarding call back timekeeping. In practice the process is flexible. Employees communicate informally about call back opportunities over cell phones. Moore has authorized two call backs in one day before. Managers have authorized call back pay when employees left work and had a reason to return, even when the shop did not “order” the employee back and there was not a three-hour gap. C We turn to the University’s investigation. Moore reviews timekeeping entries. It is fairly common for employees to have errors or anomalies with their timekeeping. Examples include swiping the wrong work number or forgetting to swipe in or out. Moore’s normal practice is to resolve errors or anomalies by discussing the issue with the employee and, if necessary, by reviewing procedures with the employee. Moore and Saldivar had met in Moore’s office several times to edit time entries. According to Moore, these were “normal edits” that “any journeyman” would have. Moore had also

6 corrected timekeeping problems for Flores. Before November 2016, the University had never disciplined Saldivar or Flores for timekeeping issues. According to Moore, the weekend after Thanksgiving 2016 had been busy. There was confusion on Saturday, November 26, 2016, and the shop had been “inundated” with calls. The following Monday, November 28, 2016, Moore reviewed timekeeping entries and noticed Saldivar and Flores each had two call backs close in time to one another on November 26. Moore did not contact Saldivar or Flores to discuss the issue. Instead, Moore tried to report the anomaly to Todd Conover, Moore’s senior, who manages the plumbing shop. Conover was not available, so Moore reported the anomaly to the next person up the chain of command, one Doug Grode. Moore told Grode the names of the employees—Saldivar and Flores—and Grode told Moore not to edit their records. Saldivar testified they could have resolved the call back issue if Moore had talked to him and had told him what to do with successive jobs. “[A]t that time, that would have been it.” The next day, November 29, 2016, the University placed Saldivar and Flores on investigatory leave. The University ultimately reviewed 12 months of Saldivar’s timekeeping records. It uncovered one issue. The issue was from December 2015. It is unclear whether the University reviewed 12 months of Flores’s timekeeping records but it identified one call back similar to the November 26, 2016 call back on November 25, 2016. We outline the events of November 25 and 26, 2016, and December 2015.

7 The day after Thanksgiving, on November 25, 2016, Flores worked an overtime shift from 4:00 a.m. to 12:30 p.m. As he was leaving the shop after clocking out, Saldivar was entering. Saldivar asked Flores to help with a flood. Flores agreed and returned to the shop. Saldivar told Flores to use the call back button and Flores complied. The next day, November 26, 2016, Flores began work at 4:00 a.m. He completed his work, returned to the shop, called Moore at 12:20 p.m., and left the shop at 12:30 p.m. Saldivar was on call that day. The Trouble Desk called Saldivar around noon about a lack of hot water in an apartment building. Around 12:40 p.m. or 12:45 p.m., as Flores was driving away, Saldivar called Flores to request help with this apartment job. Flores agreed to help, turned his car around, and returned to the plumbing shop. Around 1:00 p.m., Saldivar and Flores clocked in at the shop. Saldivar told Flores to use the call back button again and Flores complied. Saldivar used the call back button, too. The University does not challenge Saldivar’s use of call back in this instance. While Saldivar and Flores were on the apartment job, a Trouble Desk dispatcher called Saldivar to tell him two other jobs were pending. After finishing the apartment job, Saldivar called the Trouble Desk and learned another plumber, Marc Thompson, had agreed to take one of the two jobs. When he arrived at the plumbing shop, Saldivar saw Thompson outside and asked Thompson to take both jobs. Thompson agreed because Thompson was with another plumber. Saldivar and Flores went to the plumbing shop and swiped out. Thompson then called Saldivar to ask Saldivar to take one of the jobs after all. Saldivar agreed. He asked Flores to help.

8 They swiped into the job using the call back button, completed the job, and punched out. Saldivar testified he believed he needed to press the call back button for successive jobs. Flores followed instructions from Saldivar to do the same. The remaining call back issue came from the University’s 12-month review of Saldivar’s records. On the evening of Saturday, December 5, 2015, Saldivar was on call and the Trouble Desk called him about a stoppage or clog at an apartment. He clocked in around 11:00 p.m. using call back mode. Saldivar attempted to respond to the apartment but the building address, apartment number, and contact number were incorrect. Saldivar unsuccessfully searched for the location. He contacted the Trouble Desk dispatcher, who had not made contact with the caller yet. After an hour and a half, he returned to the shop. Saldivar called the Trouble Desk dispatcher, who by then had made contact with the reporting party. Saldivar and the dispatcher agreed Saldivar would take care of the call the next morning and Saldivar clocked out and began to drive home. Past midnight, while Saldivar was driving home, he grew concerned the plumbing problem could escalate and affect other units. He called the Trouble Desk dispatcher to say he would return. A transcript of this call shows Saldivar said, “I’m gonna go over there and, and swipe back into the number and take care of it right now.” The dispatcher responded, “as a matter of fact . . . I was about to put the work order in and all of that stuff in your, in your, uh pager right now. So, I am writing it up right now.” Saldivar then said he was going to call the reporting party and, “I’m gonna figure out a way to charge them another 4 hours

9 ‘cause this, I mean, for me to go back, I mean, I’m just gonna say that I, I had already left and that they called you and . . . you called me back.” Saldivar swiped in and pressed call back again. The University interviewed Saldivar and Flores as part of the investigation. It did not record statements by Saldivar and Flores. Evidence about these statements comes from notes of two identified people, neither of whom testified, and other unidentified University representatives. The University did not interview Thompson or the Trouble Desk dispatcher from the December call before it decided to terminate Saldivar and Flores. On February 24, 2017, the University issued notices of intent to dismiss Saldivar and Flores. The University concluded Saldivar’s December 2015 and November 26, 2016 conduct and Flores’s November 25 and 26, 2016 conduct constituted conscious actions to engage in timecard fraud. On April 4, 2017, the University issued dismissal letters. D In June 2017, Saldivar and Flores filed unfair practice charges with the Board against the University. They alleged the University violated the Higher Education Employer-Employee Relations Act (§ 3560 et seq.) (“the Act”) when it terminated them. The administrative law judge, Eric Cu, oversaw an evidentiary hearing for the case in September 2018. He issued a proposed decision and order that would have dismissed Saldivar’s and Flores’s complaints for failure to state a prima facie case. Before determining whether Saldivar and Flores made a prima facie case of retaliation, Cu’s proposed decision resolved an

10 estoppel issue. Saldivar and Flores had a post-deprivation hearing with the University that was unrelated to the Board. The University had moved to dismiss Saldivar’s and Flores’s unfair practice charges due to the results of those separate proceedings. Cu admitted the post-deprivation hearing report in the evidentiary record for the limited purpose of addressing the motion to dismiss. Cu’s proposed decision denied the University’s motion to dismiss in part because the post-deprivation hearing process did not fully address the issue of retaliation. Cu found Saldivar and Flores made a prima facie case for some elements of retaliation. Their participation in the November 4, 2016 lunchtime demonstration and Saldivar’s role on the bargaining team constituted protected activity. The proposed decision found Saldivar and Flores did not establish the University knew about the employees’ involvement in the lunchtime demonstration. Cu found Saldivar made a prima facie case regarding the University’s knowledge he participated in the union’s bargaining team. The proposed decision analyzed the remaining elements of retaliation for Saldivar, only. It found the University’s investigation and dismissal of Saldivar were adverse employment acts. The proposed decision turned to the final element of retaliation, unlawful motivation. Closeness in time between Saldivar’s participation in negotiations and the University’s adverse actions lent Saldivar “some support,” but timing alone was insufficient. The proposed decision acknowledged, “[m]ultiple witnesses explained that it is fairly common for employees to make errors” in their timekeeping, but found the University did not deviate from its usual policies. The call back entries did “not appear to fall into the category of obvious errors

11 that Moore would typically contact an employee about.” Cu similarly did not find unlawful motive from other factors, including cursory investigation. Saldivar and Flores filed exceptions to the proposed decision. On April 14, 2020, a unanimous three-member panel of the Board issued a 46-page decision holding the University’s terminations violated the Act. The Board’s decision deviated from Cu’s on two central points. First, the Board found Moore knew Saldivar and Flores participated in the lunchtime demonstration. Second, the Board found evidence of unlawful motive. The Board discussed Cu’s proposed decision and explained why it disagreed. It offered five pages of analysis about Moore’s and Saldivar’s credibility. The Board cited a time Moore changed his testimony about the date on which Moore had talked to Saldivar and Flores to clear up prior anomalies in their timekeeping records. Moore said this happened on a date that was after the strike but later admitted it was on a date before the strike. The Board believed Moore may have been purposefully misleading about the dates. The Board also found Moore was evasive about his knowledge of union activity in the plumbing shop. Moore said he did not know how many plumbers participated in the strike. When counsel pressed him, he said a majority of plumbers participated. Based on testimony of another University representative, the Board found all plumbers participated in the strike. The Board found it strained credulity for Moore not to have known this fact. The Board found Moore’s explanation about the flyer lacked credibility because he said he took down a flyer, but not the one

12 with Saldivar and Flores’s photo. He did not describe this other flyer and the University did not offer it as evidence. The Board found Saldivar credible and explained why it was not persuaded by the University’s contrary arguments. For example, the University did not record Saldivar’s statements during the investigation. In one instance, the University claimed interview notes proved Saldivar changed his story, but the notes contained an incomplete quote lacking key information about the issue. In another instance, the Board disagreed with the University about whether two of Saldivar’s statements were actually inconsistent. These statements were about Saldivar’s conversation with Thompson on November 26, 2016. At the hearing, Saldivar explained, “Word for word, I don’t recall, but it was my understanding” that Thompson would cover the two calls. This was not meaningfully different from an earlier statement Saldivar had made. The Board found the record as a whole did not support Moore’s claim he had not seen the flyer featuring Saldivar and Flores at the demonstration. The Board found the University had an unlawful motive based on four factors: (1) the timing of the adverse actions, (2) the University’s departure from established procedures, (3) the University’s inadequate investigation, and (4) the disproportionate penalty. The Board found Moore broke his customary practice by going to higher management rather than discussing the call back issue with Saldivar and Flores. Moore’s break with practice was notable given the high call activity and confusion in the shop on November 26. The potential for error or confusion about correct

13 procedures on this date was high, but Moore assumed the call back issue was misconduct that required investigation. The Board found the University failed to investigate several critical facts, which also tended to show unlawful motive. The Board found the University’s investigation did not sufficiently consider innocuous explanations of Saldivar’s and Flores’s use of call back. Saldivar and Flores may have honestly and reasonably believed the policy required them to press call back when they clocked in again after clocking out. The Board found the call back policy ambiguous. The written instructions do not define “premises.” Leaving the premises could mean leaving the plumbing shop building, leaving the entire campus, or something else. One could read the policy to require employees to press call back every time they clock in on a job when they are on-call. Superiors had not instructed Saldivar or Flores on this particular issue. The Board noted the University offered little evidence other plumbers shared the University’s interpretation of the call back rules. Saldivar and Flores’s interpretation may have been the norm. The Board also found the University did not interview important witnesses. The University’s claim that Saldivar was dishonest involved whether or not Saldivar knew he had a second job before he initially clocked out on November 26. Thompson, the plumber with whom Saldivar discussed the two jobs, could have shed light on this but the University did not interview Thompson before deciding to terminate Saldivar. Nor did the University interview the Trouble Line dispatcher about the December 2015 event. The Board found the dispatcher’s statements could have meant the dispatcher was

14 about to put in an immediate order for Saldivar to return. Saldivar could have honestly believed he had left the premises within the meaning of the call back policy and the dispatcher had been about to order him to return. Failing to talk to the dispatcher was “glaring,” especially because the dispatcher could have been implicated in what the University viewed as a fraudulent plan. As for Flores, he used call back at the instruction of Saldivar, the plumber with whom he was working. It was not clear Flores intended to misrepresent anything. The University assumed Saldivar and Flores were dishonest without a full investigation where other reasonable interpretations existed. The Board also found the penalty of termination was disproportionate. The Board found these factors together demonstrated unlawful motive. It also found the University did not establish its affirmative defense that it would have taken the same actions absent the employees’ protected activity. The Board found it highly unlikely the University would have initiated the investigation if Moore had used his customary practice of contacting employees to discuss timekeeping issues. Moore’s unlawful motive triggered the investigation, including the 12-month audit. The University skipped progressive discipline and moved to terminate them without interviewing all witnesses and without fully investigating whether Saldivar and Flores knowingly violated the policy. The Board found the University did not prove it would have taken the same action against Saldivar and Flores absent their protected activity.

15 II The Board was entitled to make its own credibility determinations. Substantial evidence supports the Board’s decision the University terminated Flores and Saldivar for their protected activities. A To inform our analysis, we outline the role of the Board and its relationship to administrative law judges. The Board is a quasi-judicial administrative agency that adjudicates unfair labor practice charges under the Act and other state public sector labor relations laws. (See Coachella Valley Mosquito & Vector Control Dist. v. California Public Employment Relations Bd. (2005) 35 Cal.4th 1072, 1084–1086.) By statute, the Board has five members whom the Governor appoints with the Senate’s advice and consent. (§ 3541, subd. (a).) Members serve five-year terms. (Ibid.) The Board has a specialized and focused task to protect employees’ organizational and collective bargaining rights. (See Banning Teachers Assn. v. Public Employment Relations Bd. (1988) 44 Cal.3d 799, 804.) We presume it is equipped to deal with this specialized field of knowledge and its findings within the field carry the authority of expertise that we must respect. Our relationship to the Board is generally one of deference. (Ibid.) Administrative law judges are delegates of the Board. (See § 3563, subd. (j) [Board may delegate its powers to any person the Board appoints].) Administrative law judges may serve as hearing officers. (Cal. Code Regs., tit. 8, § 32168, subd. (a).) Among other enumerated powers and duties, these individuals

16 regulate the course and conduct of hearings, take evidence, and render and serve “proposed decision[s].” (Id., § 32170.) A party may challenge a proposed decision by filing a statement of exceptions with the Board. (Cal. Code Regs., tit. 8, § 32300.) Board decisions on the merits require at least two members of the Board. (§ 3563, subd. (j).) Based on a review of recent decisions, the Board has used two-, three-, and four- member panels. (E.g., City of Santa Maria (2020) PERB Dec. No. 2736-M [45 PERC ¶ 17] [two members], City of Culver City (2020) PERB Dec. No. 2731-M [45 PERC ¶ 6] [three members], Alliance Judy Ivie Burton Technology Academy High (2020) PERB Dec. No. 2719 [44 PERC ¶ 185] [four members].) We have not found statutes or regulations governing the process by which the Board decides which members, or how many members, resolve a case. Neither side raises an issue of impartiality regarding the Board’s processes. B The Board’s credibility findings were proper. This case presents a two-layered question about deference to credibility findings. One layer involves our deference to the Board. The second layer involves the Board’s deference to the administrative law judge. A statute governs the standard of review for the Board’s factfinding. The Board’s factual findings, which include credibility findings, are conclusive if substantial evidence on the whole record supports them. (See § 3564, subd. (c).) We do not reweigh the evidence. (Boling v. Public Employment Relations Bd. (2018) 5 Cal.5th 898, 912.) We must accept the Board’s reasonable inferences. (Id. at p. 913.) Even if contrary findings may seem equally or even more reasonable than the Board’s

17 findings, we uphold the Board’s decision if substantial evidence supports it. (Regents of University of California v. Public Employment Relations Bd. (1986) 41 Cal.3d 601, 617.) The Board, on the other hand, need not defer to the administrative law judge. If a party files exceptions, “[t]he Board itself may: [¶] (1) Issue a decision based upon the record of hearing, or [¶] (2) Affirm, modify or reverse the proposed decision . . . or take such other action as it considers proper.” (Cal. Code Regs., tit. 8, § 32320, subd. (a).) Thus the Board’s discretion is broad. It independently reviews the record, may reweigh the evidence, and may draw its own factual conclusions. (City of Palo Alto v. Public Employment Relations Bd. (2016) 5 Cal.App.5th 1271, 1288.) An administrative law judge’s proposed decision is entitled to some weight but its factual findings, even credibility findings, do not bind the Board. (See McPherson v. Public Employment Relations Bd. (1987) 189 Cal.App.3d 293, 304 (McPherson).) Demeanor-based credibility findings do not conclusively bind the board. (Trustees of Cal. State University v. Public Employment Relations Bd. (1992) 6 Cal.App.4th 1107, 1129 (Cal. State).) That is because the Board, not the administrative law judge, is the ultimate factfinder. (Ibid.) In the analogous setting of the Agricultural Labor Relations Board, the Supreme Court has explained that a board’s disagreement with an administrative law judge does not change the substantial evidence standard. (Harry Carian Sales v. Agricultural Labor Relations Bd. (1985) 39 Cal.3d 209, 243.) If the board identifies evidence supporting its inferences and if the evidence is substantial when measured against the administrative law judge’s contrary findings and opposing

18 evidence, a reviewing court must uphold the board’s findings. (Ibid.) If a board rejects the administrative law judge’s credibility determinations, it should set forth express findings explaining why it reached a contrary conclusion. (See Bam, Inc. v. Board of Police Comrs. (1992) 7 Cal.App.4th 1343, 1346–1349 [board failed to explain reason for rejecting its examiner’s findings, so Court of Appeal remanded for board to vacate a suspension and make findings]; Respers v. University of Cal. Retirement System (1985) 171 Cal.App.3d 864, 873 [board must make findings to articulate to plaintiff and to the courts its reasons for rejecting administrative law judge’s decision].) The University offers three Board decisions to support its position the Board must defer to credibility findings but these decisions tend to show the opposite. In two of the decisions, the Board rejected the administrative law judges’ credibility findings, so these cases demonstrate deviation from the administrative law judge, not deference. (County of Riverside (2018) PERB Dec. No. 2591-M [43 PERC ¶ 66], and State of California (Department of Corrections & Rehabilitation) (2012) PERB Dec. No. 2285-S [37 PERC ¶ 72] (Department of Corrections & Rehabilitation).) Additionally, the three decisions uniformly qualified their statements about deference to credibility findings. The Board defers, if the administrative law judge grounded its credibility determinations in impressions from firsthand observations of testimony. (Department of Corrections & Rehabilitation, supra, PERB Dec. No. 2285-S at p. 10; County of Riverside, supra, PERB Dec. No. 2591-M at p. 6.) It defers, unless the “record warrants overturning.” (County of Riverside, supra, at p. 6.) The Board defers, “absent evidence to support overturning such

19 determinations.” (Los Angeles Unified School District (2014) PERB Dec. No. 2390, p. 12 [39 PERC ¶ 34].) So the Board defers, but only to findings based on observation and not if there is evidence not to defer. These cases prove the Board’s precedents do not compel significant deference to credibility findings. There is cause to question the value of the firsthand observation credibility findings the Board’s precedents favor. The value of demeanor evidence is enshrined in Board precedents, case law, and our evidence code, but commentators have persuasively challenged the assumption factfinders can successfully discern truth or falsity from demeanor. In a 1991 article, Olin Guy Wellborn reviewed empirical evidence about demeanor and found, “ordinary people cannot make effective use of demeanor in deciding whether to believe a witness. On the contrary, there is some evidence that the observation of demeanor diminishes rather than enhances the accuracy of credibility judgments.” (Wellborn, Demeanor (1991) 76 Cornell L.Rev. 1075, 1075 (Wellborn); see also Blumenthal, A Wipe of the Hands, A Lick of the Lips: The Validity of Demeanor Evidence in Assessing Witness Credibility (1993) 72 Neb. L.Rev. 1157, 1200 [concluding demeanor findings based on visual observations are generally of little use in determining witness credibility].) Wellborn explained, “To the extent that people can detect lying or erroneous beliefs in another, they do so primarily by paying close attention to the content of what the other says, not by observing facial expression, posture, tone of voice, or other nonverbal behavior.” (Wellborn, supra, 76 Cornell L.Rev. at pp. 1104–1105.)

20 Prominent jurists have expressed skepticism about demeanor assessments too. “I am convinced . . . much that is thought and said about the trier of fact as a lie detector is myth or folklore.” (Penasquitos Village, Inc. v. N. L. R. B. (9th Cir. 1977) 565 F.2d 1074, 1084 (conc. & dis. opn. of Duniway, J.).) Judge Richard Posner explained, “Judges fool themselves if they think they can infer sincerity from rhetoric and demeanor.” (U.S. v. Wells (7th Cir. 1998) 154 F.3d 412, 414.) The present case does not require us to decide whether it is proper to defer to demeanor findings. We recount the proposed decision’s references to credibility below. Few, if any, of the statements about credibility were express “observational” or demeanor findings, so we do not pursue this issue further. In sum, the Board can diverge from the administrative law judge’s credibility findings if record evidence supports this decision. We defer if substantial evidence supports the Board’s findings. Under the circumstances of this case, the Board was entitled to make its own credibility findings. First we address whether the University properly raised this issue. The University’s argument is based on its assertion the administrative law judge made credibility determinations “stated in detail.” The University did not offer appropriate record support for this contention in its opening brief. We consider its belated citations but find these do not establish the Board had a duty to defer. The University offered insufficient record support about the credibility determinations. Its opening brief purported to recount the administrative law judge’s detailed findings with an 11-line quote. The quote comes not from the administrative law judge’s

21 proposed opinion but from the University’s own closing brief after the hearing. The University brushes off this issue in its reply brief as an “incorrect citation” it “mistakenly cited . . . as various drafts were edited.” Mistake or not, the opening brief failed to support its assertion the administrative law judge made detailed credibility findings with appropriate references to the record (Cal. Rules of Court, rule 8.204(a)(1)(C)), and we deem this assertion waived. (Nwosu v. Uba (2004) 122 Cal.App.4th 1229, 1246.) In its reply brief, the University cites five main credibility findings. The Board had no chance to brief this newly supported argument. We may disregard arguments raised for the first time in the reply brief. (WorldMark, The Club v. Wyndham Resort Development Corp. (2010) 187 Cal.App.4th 1017, 1030, fn. 7.) The University’s citations do not change our conclusion, however, and we address them to explain our reasoning. We take the findings in turn and explain why they did not compel the Board to defer. In a footnote, Cu said Saldivar’s testimony had inconsistencies and contradictions while Moore “testified relatively consistently, and was forthright about what he knew and what he could not recall.” This finding was about the content of the testimony and it was not an express observational finding that might require particular deference under Board precedents. As we have recounted, the Board extensively analyzed Moore’s and Saldivar’s credibility and supported its findings with record citations. The Board’s findings were proper. The proposed decision said Conover, the shop manager, testified “persuasively” on the issue of not needing to interview other employees before dismissing Saldivar. The proposed decision did not elaborate about what made the testimony

22 persuasive. A person’s persuasiveness could be based on observed characteristics, but the proposed decision did not say this. The Board offered a thorough explanation grounded in the hearing transcript, which showed Conover altered his explanation of this issue several times. The Board was entitled to discredit Conover on this basis. The other credibility findings were largely immaterial because the Board’s decision was consistent with them. Cu credited Moore’s testimony about the timing of Moore’s review of the employees’ timecards. The Board agreed on this point, but found Moore deviated from his usual practices on other grounds. Cu credited Moore’s testimony about why the University did not start an investigation against two other employees. The Board did not rely on this issue to support its finding of unlawful motive. Cu credited testimony about the University pursuing dismissal when other employees made intentional misrepresentations on their timecards. The Board found a lack of evidence Saldivar and Flores made intentional misrepresentations on their timecards so this credibility determination was likewise immaterial to the Board’s decision. In the instances where the Board disagreed with the proposed decision’s credibility findings, these were not express observational findings and the Board offered extensive and appropriate reasons to deviate. The University offers three other unavailing reasons why the Board had a duty to defer to the administrative law judge’s credibility determinations. First, the University tries to fight the deferential standard of review. It offers a novel argument: “The fact that there was testimony contrary to the ALJ’s finding further supports the

23 conclusion that the ALJ’s credibility determinations deserve deference as the ALJ rendered a decision after weighing conflicting testimony and applying credibility factors.” The University’s view does not square with the regulatory, statutory, appellate, and Board authorities we have cited. If the University were correct, on a conflicting record, the Board would be powerless to conclude differently from the administrative law judge. But the Board may reverse a proposed decision and issue its own decision based upon hearing records. (Cal. Code Regs., tit. 8, § 32320, subd. (a).) The Board’s factual findings are conclusive if substantial evidence on the whole record supports them. (§ 3564, subd. (c).) Second, the University says results from a separate post- deprivation hearing are notable but the present appeal is not about those proceedings. The administrative law judge added the post-deprivation hearing report to the evidentiary record for the limited purpose of addressing the University’s motion to dismiss. The proposed decision denied the University’s motion in part because the post-deprivation hearing process did not fully address the issue of retaliation. The University did not file any exceptions to the proposed decision. A party waives exceptions the party does not “specifically urge[].” (Cal. Code Regs., tit. 8, § 32300, subd. (c); cf. Rivcom Corp. v. Agricultural Labor Relations Bd. (1983) 34 Cal.3d 743, 756, fn. 6 (Rivcom) [finding waiver due to failure to raise issue in exception].) The Board found no cause to defer to the post-deprivation ruling on any issue. The post-deprivation proceedings are not pertinent to our analysis. Third, the University distorts a statement from the Board’s opinion that, in context, supports the Board. The full statement

24 was, “Saldivar and Moore both showed efforts—common among witnesses in a fraught trial—to bend events to their pre- determined viewpoint. Neither is inherently more believable than the other as to every point. However, at least on the issue of Moore’s knowledge that Saldivar and Flores participated in the lunchtime demonstration, we cannot affirm the ALJ’s finding. The record as a whole does not support Moore’s claim that he did not see a flyer posted in the plumbing shop and throughout the campus, or his integrally related claim that he instead saw and took down an unspecified, different strike flyer that he claims happened to be posted at the same time.” The Board did not credit Moore on this issue, which “somewhat eroded” Moore’s credibility and was one of several facts relevant to the Board’s motive analysis. The University seizes upon the Board’s statement that neither Moore nor Saldivar was “inherently more believable than the other.” In context, these words do not show Moore was credible nor do they discredit the Board’s decisionmaking. Rather, they show the Board’s reasoning was clear-eyed and measured. The Board’s credibility determinations were proper. C Substantial evidence supports the Board’s decision. We review the relevant law. The Legislature passed the Act in 1978 to permit higher education employees full participation in determining their conditions of employment. (§ 3560, subd. (e).) The Act gives higher education employees the right to form, join, and participate in unions. (§ 3565.) It also protects employees from retaliation for participating in protected union activity. (§ 3571, subd. (a).)

25 The elements of retaliation are: (1) an employee engaged in protected activity that (2) the employer knew about, and (3) the employer took an adverse action against the employee (4) because of the protected activity. (Novato Unified School District (1982) PERB Dec. No. 210-E, pp. 4–8 [6 PERC ¶ 13114] (Novato); Cal. State, supra, 6 Cal.App.4th at p. 1123.) If the employee makes a prima facie case, the employer may raise an affirmative defense that it would have taken the adverse action regardless of the employee’s protected activity. The employer has the burden to prove the defense by a preponderance of the evidence. (Novato, supra, PERB Dec. No. at p. 14; McPherson, supra, 189 Cal.App.3d at p. 304.) The University and the Board cite Novato and recite the four prima facie elements we have listed. The Board has called the fourth element “causal connection,” “nexus,” and “unlawful motive.” (E.g., Rio School District (2015) PERB Dec. No. 2449-E, pp. 29–31 [40 PERC ¶ 45].) We will call the fourth element “unlawful motive” because the proposed decision and Board decision used this term. The University separately analyzes “causal link,” which it says requires knowledge, and “retaliatory motive.” Causal link and retaliatory motive both refer to the fourth element and we consider them together. We turn to the four prima facie elements. First, Saldivar’s and Flores’s participation in the November 4, 2016 lunchtime demonstration and Saldivar’s participation in union negotiations were protected activities. The University does not dispute this. Substantial evidence supports the second element, knowledge. Cu and the Board found the University knew about Saldivar’s participation in union negotiations. The University

26 did not challenge the proposed decision’s findings as to this issue and we do not discuss it further. The Board properly inferred the University, and Moore in particular, knew about Saldivar’s and Flores’s participation in the lunch demonstration. The demonstration took place during lunch on a work day. It was on campus, outside, and public. Soon after this demonstration, union supporters, including Saldivar, distributed a flyer around campus. Saldivar placed many copies in the plumbing shop. The flyer had a picture of Saldivar and Flores at the demonstration. Saldivar said he saw Moore take down the flyer. The Board reasonably inferred Moore took down the flyer with Saldivar and Flores’s photo. The Board properly discredited Moore’s claim he took down a different union flyer in a similar time period. The Board also reasonably inferred a supervisor like Moore would have known about the demonstration and would have known the flyer’s photo depicted the demonstration. The public nature of the demonstration, the close timing between the demonstration and the flyer’s presence on campus, and the photo’s depiction of people holding signs as one would expect at a demonstration help support this conclusion. The University challenges the knowledge component because the flyer does not say the photo was from the November 4 demonstration but the Board could infer this from the circumstantial information we have recounted. The University offers other inferences, such as the possibility the union “staged” the photo. This inference seems doubtful. The proposed decision did not offer it. In any event, we defer to the Board’s reasonable inferences and we find support for the Board’s inference Moore saw the flyer and knew it showed Saldivar and

27 Flores at the demonstration. The Board properly found knowledge as to the lunch demonstration. Substantial evidence supports the third element, adverse action. Investigation and termination are adverse actions. There is substantial evidence of the fourth element, unlawful motive. This element is often less straightforward than the first three. It is rare to have direct evidence of motivation. (Cal. State, supra, 6 Cal.App.4th at p. 1124.) When motive is a central issue, the Board usually must rely heavily upon circumstantial evidence and inferences. (Ibid.) The Supreme Court discussed the issue of deference to a different board’s motive findings based on inferences and circumstantial evidence in Rivcom, supra, 34 Cal.3d at page 757. The court explained, “As a rule, motive must be inferred from all the circumstances, and the [Agricultural Labor Relations] Board’s expertise entitles it to considerable deference in deciding that question.” (Ibid.) The Board correctly found four bases for unlawful motive. First is timing. The University began investigating the timecard fraud just weeks after Saldivar and Flores participated in the demonstration. Saldivar was also involved in union negotiations in 2016. This close timing is probative given Saldivar’s and Flores’s long tenure with the University and lack of prior discipline for timekeeping issues. The timing tends to show unlawful motive. Second is departure from typical procedures. Moore immediately reported the timecard anomalies to his superiors without informally discussing them with Saldivar or Flores. It is reasonable to infer Moore, absent improper motive, would have tried to resolve the issues his usual way by talking to Saldivar and Flores first.

28 The circumstances of November 26, 2016, and the ambiguity of the call back policy suggest Moore’s actions were a departure from usual practice. Moore agreed November 26, 2016, was very busy. It would have been reasonable for him to think the call backs were related to this confusion and for him to discuss the issue with Saldivar and Flores first. Additionally, there was evidence Moore had flexibly permitted call backs in the past and the policy was unclear. This suggested the anomaly could have been innocuous and a product of misinterpretation, mistake, or confusion. It was reasonable for the Board to infer the call back issue was one Moore would have typically handled together with the employees. The miscategorization of time could be akin to other types of issues Moore edited and discussed with employees, like using the wrong work order number or failing to clock in. According to Moore, after Moore reported the call back issue to Grode, Grode advised Moore not to edit Saldivar’s and Flores’s timesheets. Grode would not have needed to offer this instruction unless it would have been a usual practice for Moore to edit this type of anomaly. The Board properly found Moore deviated from his usual procedure. Cu inferred the call back issue was different from other anomalies Moore had edited in the past. The Board was entitled to infer differently from Cu on this issue. The nature of the investigation also tended to show unlawful motive. Thompson and the Trouble Line dispatcher could have provided more information about what happened before the successive call backs. This was important given the ambiguity of the call back procedures and the potential Saldivar and Flores honestly believed successive call backs were

29 appropriate under the particular circumstances. More information about the events that came before the successive call backs could have helped the University understand Saldivar’s and Flores’s intent and knowledge. Lastly, there was evidence of disproportionate penalty. Saldivar and Flores were longtime employees and the University had not disciplined either employee for timekeeping issues in the past. As a facilities worker, Flores followed the instructions of the plumber, Saldivar, with whom he was working. Getting directions from another does not absolve a person of wrongdoing. In this case, however, the circumstances support a belief Flores was not knowingly breaking a rule. As to both Flores and Saldivar, the ambiguity and flexibility of the call back practices supports the Board’s finding they did not make knowing misrepresentations. The Board properly found termination to be a harsh penalty under these circumstances. These four factors are sufficient to prove unlawful motivation. The University offers several arguments in favor of its position but these do not require reversal. The University says there is no substantial evidence of unlawful motive because it did not terminate other employees who attended the lunchtime demonstration, nor did it terminate two other employees in the flyer’s photo. This information is probative, but it does not equal “no substantial evidence.” The fact the University began investigating half, not all four, of the employees featured on the flyer’s photo weeks after the demonstration proves little. Saldivar was differently situated from other members of the lunchtime demonstration because he also participated in the

30 Union’s bargaining team. The University claims Saldivar was not a key bargaining team member, “at least from the University’s perspective.” This self-serving position carries little weight. It is reasonable to infer the University was influenced by Saldivar’s combined union activities. Saldivar’s and Flores’s timekeeping anomaly may have simply provided the University enough pretext to target them. Their retaliation claims did not require them to prove the University terminated other employees at the demonstration. The University says the Board inaccurately found Saldivar and Flores did not get instruction on how to log call backs aside from the ambiguous written procedure but the Board’s finding was proper. The University cites (1) testimony from Moore that all employees received training on a new timekeeping device, and (2) forms Saldivar and Flores signed saying they had received written work rules. This evidence does not prove Saldivar and Flores got instruction on how to log call backs outside the written procedure, which is silent about how to handle successive call backs. Saldivar and Flores each testified about their lack of training regarding call back timekeeping. The University has not proven the Board’s finding was inaccurate. The University incorrectly claims Saldivar testified, “he had never experienced union animus in his 11 years at UCLA,” therefore by Saldivar’s admission, “there was no union hostility in his department and thus none can be inferred.” The University supports its proposition with a citation to counsel’s question, “Did you believe that you would be targeted by management?” and Saldivar’s response, “At that moment, no.” On its face, Saldivar’s answer was about a particular time, not about 11 years of time. Saldivar had been testifying about his

31 coworkers’ belief that management would probably target Saldivar because of the flyer. Directly after that, Saldivar said he had believed management would target him at a different time. If anything, this is proof Saldivar experienced union hostility. The University’s assertion Saldivar admitted he never experienced union hostility in this exchange is incorrect. Substantial evidence supports the Board’s rejection of the University’s affirmative defense it would have terminated Saldivar and Flores notwithstanding their union activities. An employer cannot meet its burden of proof if it uncovered its stated reasons for taking an adverse action through an investigation tainted by unlawful motive. (State of California (California Correctional Health Care Services) (2019) PERB Dec. No. 2637-S, pp. 20–21 [43 PERC ¶ 164].) The same factual basis that supports the motive analysis supports the Board’s finding unlawful motive tainted the investigation. Moore did not talk to Saldivar and Flores before escalating the issue, this happened quickly after the employees’ protected activities, and Moore attempted to minimize his knowledge of union activities in the plumbing shop. As Saldivar testified, if Moore had followed his typical process and had talked to him, they could have resolved the issue immediately. The University identified the November 25, 2016 and December 2015 events through the investigation that had been tainted by an unlawful motive. We therefore do not consider those events as proof of the affirmative defense. The University did not prove it would have fired Saldivar and Flores absent their union activity and based on their timekeeping on November 26, 2016. The University offered cursory evidence they had fired other employees for timecard

32 fraud. A University representative briefly mentioned two incidents, one in which it fired two elevator workers for “almost the exact same thing.” The representative did not give any other information about what made that case similar to the case of Flores and Saldivar. The second example included slightly more detail: the University terminated two roofers because one roofer swiped in and out for another roofer when that other roofer was not at work. The first case lacks detail and the second case is distinguishable. Without additional information about the first example, the Board did not need to assume the case involved the same facts as Saldivar’s and Flores’s cases. The frequency of the issue, type of timekeeping issue, tenure at the University, and history of timekeeping issues may well have differed. As for the second case, the University has not suggested Saldivar and Flores did not work during the hours they claimed. Rather, the issue is how they categorized this time. Cursory testimony the University had terminated other employees for time fraud did not require the Board to find the University met its burden of proof for the affirmative defense.

33

DISPOSITION

We deny the University’s petition for extraordinary relief. Saldivar and Flores are awarded their costs on appeal. (Cal. Rules of Court, rule 8.493(a).)

WILEY, J.

We concur:

GRIMES, Acting P. J.

STRATTON, J.

34