Santa Rosa Memorial Hospital

National Labor Relations Board - Administrative Judge Opinions

May 28, 2010

20-RC-018241 (N.L.R.B. May. 28, 2010)

JD(SF)–18–10
Santa Rosa, CA
UNITED STATES OF AMERICA
BEFORE THE NATIONAL LABOR RELATIONS BOARD
DIVISION OF JUDGES SAN FRANCISCO BRANCH OFFICE
SANTA ROSA MEMORIAL HOSPITAL
Employer
and 20-RC-18241
NATIONAL UNION OF HEALTHCARE WORKERS
Petitioner
and
SEIU-UHW, UNITED HEALTHCARE WORKERS-WEST
Intervenor
Gregory W. McClune, Eileen Ridley, and Kristy Marino, Attys.
(Foley and Lardner, LLP) San Francisco, California, for the Employer.
Jonathan H. Siegel, Atty. (Siegel & Le Witter) Oakland, California,
for the Petitioner.
Catherine Ventola, Atty., for the Regional Director, NLRB Region 20.
DECISION AND RECOMMENDED ORDER
ON OBJECTIONS TO ELECTION
Statement of the Case
WILLIAM L. SCHMIDT, Administrative Law Judge. On December 17 and 18, 2009, a
representation election was conducted at the Santa Rosa Memorial Hospital (Employer or
Hospital) under the supervision of the Regional Director for Region 20 pursuant to a Stipulated
Election Agreement that he approved on November 2, 2009.1 The unit includes 120 employee
classifications (detailed in the Order below) that perform technical, service, and maintenance
work at the Hospital and its nearby satellite facilities. The original tally of ballots issued at the
conclusion of the balloting on December 18 reflected 17 challenged ballots, a number sufficient
to affect the outcome of the election. Later, the Employer timely filed ten election objections.
On February 4, 2010, following an administrative investigation, the Regional Director
issued a Report on Challenged Ballots and Objections to Election (RD’s Report). A revised tally
of ballots included in the RD’s Report shows that 283 voters cast ballots for the National Union
of Healthcare Workers (Petitioner or NUHW), 263 cast ballots for neither labor organization, and
13 cast ballots for the for the SEIU-UHW, United Healthcare Workers West (Intervenor or SEIU-

1 If not shown, dates refer to the 2009 calendar year.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
2
UHW-W).2 The RD’s Report sustained 13 of the 17 challenges made to ballots at the election,
leaving the four remaining challenged ballots insufficient to affect the election’s outcome. The
RD’s Report also states that the Employer voluntarily withdrew six of its original objections
during the administrative investigation. The Regional Director issued a concurrent Notice of
Hearing for the purpose of resolving Employer Objections 1, 3, 8, and 9.
I conducted the hearing on objections on February 22, 23, 24, at Santa Rosa, California
and on February 26, at San Francisco, California, pursuant to the Regional Director’s Notice of
Hearing. During the hearing, the parties had an ample opportunity to call and examine
witnesses on material matters, to introduce relevant documentary evidence, to argue procedural
and substantive issues, and to file post-hearing briefs. The Employer and the Petitioner
submitted timely, post-hearing briefs that have been most helpful in my consideration of the
issues in this case.3
After carefully considering the hearing record in light of my credibility determinations,4
and the arguments set forth in post-hearing briefs, I recommend that the Board overrule the
Employer Objections 1, 3, 8, and 9, and that the Board certify the Petitioner as the exclusive
representative of the unit employees under Section 9(a) of the Act based on the following
findings and conclusions.
Findings of Fact
I. Preliminary Findings
As shown by several record references, the SEIU-UHW-W made an unsuccessful effort
to organize this unit of employees five or six years ago. The connection of that organizing effort
to the one involved here is only incidental. However, that campaign and the current NUHW
campaign that began in 2009 were conducted by essentially the same organizers. Between the
two campaigns, an internal dispute developed that culminated with the SEIU president
establishing a trusteeship over the SEIU-UHW-W, a large California local union that represents
thousands of workers in the healthcare industry, and removing several of its officials from office.
Thereafter, the former SEIU-UHW-W officers and many of its professional organizers formed
the National Union of Healthcare Workers and began competing with the SEIU-UHW-W for
bargaining rights at several California healthcare institutions. See e.g., Service Employees v.
National Union of Healthcare Workers, 187 LRRM 3569 (9th Cir. March 15, 2010).

2 Both ballot tallies show approximately 675 eligible voters.
3 The Intervenor did not participate in either the hearing or the pre-hearing conference
conducted. A counsel for the Regional Director participated in all aspects of the hearing as well
as the pre-hearing conference. As the Regional Director’s counsel had only the limited role of
aiding in the development of a full record, she understandably did not file a post-hearing brief.
4 The findings of fact made below incorporate my credibility determinations. Evidence
contradicting these findings has been considered but has not been credited. In general, my
credibility resolutions resulted from the conclusions I reached after considering a witness’
opportunity to be familiar with the subjects covered by their testimony; established or admitted
facts; the impact of bias on the witness’ testimony; the quality of the witness’ recollection;
testimonial consistency; the presence or absence of corroboration; the strength of rebuttal
evidence, if any; the inherent probabilities; reasonable inferences available from the record as a
whole; the weight of the evidence; and witness demeanor while testifying. More detailed
discussions of specific credibility resolutions appear below in those situations that I perceived to
be of particular significance.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
3
The NUHW’s 2009 organizing effort at the Santa Rosa Memorial Hospital led to the filing
of the petition in this case on April 13. Employee Fawn Kruat’s astute observation that the same
organizers conducted both campaigns essentially summarizes what occurred. And, it seems,
many of the area citizens who publicly sided with the SEIU-UHW-W during the earlier campaign
became supporters of the NUHW in the 2009 organizing effort.
The Hospital’s brief emphasizes that the election outcome was extremely close. The
NUHW received three votes more than the number required to avoid a runoff. The closeness of
the election is of particular significance to the opinion testimony of Dr. Jon A. Krosnick (Dr.
Krosnick) that the Hospital adduced in support of Objections 1 and 9. The RD’s Report makes
mention of evidence of an expert witness and Hospital’s counsel addressed this aspect of his
case during the pre-election conference. At that time, I informed the parties that I would permit
the Hospital to call and question its expert witness at the hearing because the limited time
between the Notice of Hearing and the hearing date as well as the lack of a discovery procedure
under the Board’s rules precluded more proper “gatekeeping” but that I would analyze such
opinion evidence using the principles that have evolved from the Daubert case and its progeny.5
As an initial matter, I find Dr. Krosnick qualified as an expert in the field of elections
based on his knowledge, skill, experience, training, and education in that area.6 The Petitioner

5 Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 588, 597 (1993) (Daubert I);
Kumho Tire v. Carmichael, 526 U.S. 137, 141 (1999). These cases require the tribunal to
perform a gatekeeping function to determine the admissibility under Rule 702 of the Federal
Rules of Evidence of expert scientific testimony (Daubert I), and testimony based on technical
and other specialized knowledge (Kumho Tire). In the context of those cases, the gatekeeping
function is designed to determine in advance whether a jury should be exposed to expert
opinion testimony. Although the absence of a jury in an administrative proceeding may reduce
the importance of the gatekeeping function to a degree, the analytical tools that have evolved
out of the gatekeeping required under Daubert I and Kumho Tire are informative and useful to
an administrative tribunal in determining the admissibility of scientific and technical evidence as
well as the weight that should be accorded to if received.
6 Dr. Krosnick received an A.B. degree in Psychology from Harvard University and a M.A.
and Ph.D. degrees in Social Psychology from the University of Michigan, accumulating along
the way extensive training in social psychology and political behavior with a focus on voting and
elections. Over the past 25 years, Dr. Krosnick has conducted research that examined how
social influence occurs, how people influence each other in decision making situations, and how
the formats of questions and ballots influence the votes cast in elections. Dr. Krosnick has
authored 6 books and over 120 peer-reviewed book chapters or journals in the area of social
psychology, many of which relate to the area of elections. He has participated in nearly 300
academic and professional presentations in his field of study, including most recently,
presentations on the activities of the American National Election Studies and a presentation of a
paper on the 2008 American Presidential Election. Dr. Krosnick serves as the co-principal
investigator of the American National Election Studies (ANES) funded by the National Science
Foundation. Since 1948 the ANES has collected data seeking to understand why voters decide
whether to vote or not, and how they decide to vote, and for whom in national elections. The
data from this continuing study (collected under Dr. Krosnick’s direction in 2008) has generated
thousands of publications by scholars from around the world. Dr. Krosnick has testified as an
expert in four prior cases, including those concerning the potential affect on elections by the
actions of persons involved in the campaigns and/or structure of the ballots. Tr. 713-22; Er.
Exh. 16, p. 15-24; Er. Exh. 17; Er. Exh. 19, p 1-2.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
4
does not challenge Dr. Krosnick’s credentials as an expert but, as will be detailed below, it does
challenge the validity and relevance of Dr. Krosnick’s opinions concerning this election.
In Safeway, Inc. 338 NLRB 525-26 (2002), the Board provided the following summary of
the standards applicable in objections cases:
It is well settled that “[r]epresentation elections are not lightly set aside.” NLRB v. Hood
Furniture Mfg. Co., 941 F.2d 325, 328 (5th Cir. 1991) (citing NLRB v. Monroe Auto
Equipment Co., 470 F.2d 1329, 1333 (5th Cir. 1972), cert. denied 412 U.S. 928 (1973)).
Thus, “[t]here is a strong presumption that ballots cast under specific NLRB procedural
safeguards reflect the true desires of the employees.” NLRB v. Hood Furniture Mfg. Co.,
supra, 941 F.2d at 328. Accordingly, “the burden of proof on parties seeking to have a
Board-supervised election set aside is a ‘heavy one.’” Kux Mfg. Co. v. NLRB, 890 F.2d
804, 808 (6th Cir. 1989) (quoting Harlan #4 Coal Co. v. NLRB, 490 F.2d 117, 120 (6th
Cir. 1974), cert. denied 416 U.S. 986 (1974), the objecting party must show, inter alia,
that the conduct in question affected employees in the voting unit. Avante at Boca
Raton, Inc., 323 NLRB 555, 560 (1997) (overruling employer’s objection where no
evidence unit employees knew of alleged coercive incident). See generally Antioch Rock
& Ready Mix, 327 NLRB 1091, 1092 (1999).
With this background, I now turn to consideration of the Hospital’s specific objections before me.
II. Objection 1
This objection asserts that the “NUHW and its agents repeatedly engaged in electioneering at
the polling place on election days and during periods when the polls were open.”
A. The Relevant Evidence
The polling place “electioneering” referenced in this objection pertains solely to the union
buttons and other partisan displays worn by the NUHW election observers while on duty at the
polling sites during the five voting periods required to complete the election. Issues related to
observers wearing partisan displays initially arose at the December 16 pre-election conference
where the assigned election agents reviewed the ground rules with the parties’ representative.
NLRB field examiner Scott Smith served as the lead election agent; NLRB field attorneys
Richard McPalmer, and Paula Katz assisted Smith.
During the conference, the Employer’s attorney, Gregory McClune, proposed a rule
prohibiting the observers from wearing partisan campaign paraphernalia, such as buttons,
insignia, or the like, while on duty during the polling periods. The union representatives present
rejected a total ban but McClune understood that the union agents later agreed to limit the
observers’ partisan displays to a single button about the size of a silver dollar. (Tr. 40-42)
However, Glenn Goldstein, NUHW’s lead organizer for this campaign, categorically denied that
he agreed to any limitations on observers wearing partisan displays. The Hospital’s human
resources vice president Deborah Miller said that the SEIU’s agent, Mike Lauer, also denied
that an agreement had been reached when she raised the subject at one of the voting sessions.
Field attorney McPalmer and field examiner Smith recalled the discussion about the
observers wearing partisan displays during at the pre-election conference but neither had vivid
recollection of the details apart from the fact that McPalmer spoke to this issue when it arose.
He emphasized that the Board strongly discouraged observers from wearing partisan insignia
while serving as an observer but that the Board did not prohibit the wearing of such displays
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
5
outright. However, McPalmer warned that the lack of a strict prohibition did not mean that one
of the parties could not object to an election on that ground, or that the Board would not find
merit to such an objection if the conduct went too far.
Most, if not all, observers ignored McPalmer’s advice. The Employer observers wore
two different buttons, both about two inches in diameter. One read, “Neither,” and the other
read, “Your Vote Counts.” Even though the latter is highly ambiguous as a partisan message, it
had been provided by, and became identified with, the Employer’s position in opposition to
unionization. Several of the NUHW observers wore a single button roughly the size of a silver
dollar that bore the NUHW acronym. Others wore a 1 x 3 inch sticker with printing that stated
“Voice NUHW” on the top line and “Vote December 17-18” beneath. This sticker can be seen
on individuals shown in the photograph at page 5 of Employer Exhibit 7. Two of the NUHW
observers, Stanley (Buddy) Bosenko and Michael Hartnett, wore red NUHW t-shirts to their
assigned voting sessions. The legend on the front of the t-shirt states: “I Choose NUHW” and
the legend on the back states: “Because I Believe In Democracy.”
Nearly all of the observers appeared for their assigned voting sessions wearing at least
one button displaying their election preference. A few had two buttons or, in the case of the
NUHW observers, a button and the rectangular NUHW sticker. Miller recalled that Bosenko
appeared in his red NUHW t-shirt with four or five NUHW buttons attached to it. Miller said
nothing about Bosenko partisan display when she saw it and there is no evidence that anyone
else made an issue of what he wore to the voting session.
However, an issue arose when Hartnett appeared for his designated polling period
wearing the red NUHW t-shirt along with several NUHW buttons.7 Witness estimates about the
number of buttons Hartnett had pinned to his t-shirt varied from five on the low side to two
dozen on the high side.8 The high estimate came from Board agent Smith who remembered
“one gentleman, in particular, who was simply festooned with buttons.” Although Smith could
not recall the observer’s name, he further asserted that this particular observer wore the array of
buttons all through the voting period. (Tr. 695)
Smith is the only witness who recalled that an observer served while wearing a large
array of partisan buttons or other displays. NUHW organizer Peter Tappeiner remember that
the Employer’s attorney began making gestures with his pen as though he were counting the
number of buttons on Hartnett’s t-shirt after he arrived for the meeting held in advance of the
voting session. A discussion ensued about Hartnett’s display. The Employer’s counsel present
for the session complained by saying, “(H)ey, this seems to be going beyond the pale here, you
know, all these buttons, this seems to be electioneering.” Field attorney McPalmer repeated his
warning given earlier that the wearing of campaign materials could lead to objections.
Thereafter, NUHW organizer Tappeiner led Hartnett from Conference Room B where the
meeting was held into the adjacent Conference Room A for a lengthy discussion. At the end of

7 Hartnett did not testify but he appears in several photographs received in evidence
wearing the red t-shirt sans the problematic display of buttons discussed here. See e.g.,
Employer Exhibit 7, page 6.
8 Although confusion reigns among the witnesses about the polling session in which Hartnett
served, in the final analysis, that fact matters little, if at all. Suffice it to say that there were five
voting sessions, three on December 17 and two on December 18. Hartnett either served as
observer at the last session on December 17 (10:30 to 11:30 p.m.), or the final voting session
on December 18.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
6
their talk, Hartnett removed all of the buttons, turned his t-shirt inside out, and then put one or
two of the NUHW buttons back on his t-shirt.9 (Tr. 528-31)
McPalmer remembered that Tappeiner and the observer (whom he could not identify by
name) conferred for some time and that the designated observer removed many of the buttons.
(Tr. 667-68) Deborah Miller, who was present for the meeting generally agreed. She identified
the observer as Hartnett. She recalled that McPalmer’s warning as well as the fact that Hartnett
reversed his t-shirt and wore only one or two buttons for the voting session.10 (Tr. 159-61)
No evidence shows that any observer engaged in any verbal proselytizing with the
voters who appeared to vote nor is there even evidence that any observer engaged in any
unnecessary discussions with the voters at the polling area. Similarly, no evidence establishes
that any voter appeared attracted by or commented on partisan insignia worn by the observers
during the polling periods.
B. The Opinion Evidence of Dr. Krosnick
With respect to the impact of observer’s partisan insignia, Dr. Krosnick offered his
opinion that “the presence of insignias endorsing the NUHW immediately outside the voting
booth is likely to have encouraged voters to cast votes for the NUHW and, thus, increased their
share of the election outcome.” (Tr. 728) Dr. Krosnick based his opinion on two bodies of
evidence: (1) the theories and research found in the scientific literature that existed prior to this
case; and (2) results of an experiment he conducted for purposes of this litigation. (Tr. 725,
728, 730-734)
The Scientific Literature: Dr. Krosnick reported that he reviewed existing literature and
research in order to shed light on whether the presence of insignia in the immediate proximity to
the voting booth might influence voters’ decisions. Dr. Krosnick relied on two general bodies of
theory and research suggesting that partisan insignias may have an impact: (1) conformity and
(2) cognitive priming. (Tr. 730-33)
As the term is used in the relevant scientific literature, conformity describes the
consistent tendency of individuals to “look for guidance about how to behave (based) on the
assumption that others behave intelligently.” For example, Dr. Krosnick said, an insignia such
as ‘I choose NUHW,’ worn by one or more individuals in the vicinity of the voting booth sends a
signal to the voters that others endorse, and support, and would like to see victory for this union.
Sending such a signal, Dr. Krosnick added, “has been shown in many studies to enhance the
likelihood that people will perform behavior consistent with the signal for that reason.”
Cognitive priming is a label that is used in the literature to describe the response by
voters about to enter the voting booth to external stimuli such as partisan insignia, or even a t-

9 Hartnett, as well as the other observers, also wore the official NLRB observer badge that
identified him as the observer for the NUHW. Tr. 506, 531.
10 The Hospital’s brief suggests that Smith’s recollection about an observer “festooned” with
campaign insignia may have been someone other than Hartnett. Employer’s Brief, pp 10-11. I
reject that suggestion as highly improbable where, as here, Miller and Tappeiner identified
Hartnett as the problem observer with the most elaborate display by far. Smith mentioned only
one problematic observer so the notion another observer other than Hartnett caught Smith’s
attention but escaped everyone else’s notice is extremely unlikely. Even more unlikely is the
likelihood that Smith would have never mentioned another individual until the hearing.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
7
shirt in a color closely associated with a particular party. Studies show that cognitive priming
tends to activate in the voter’s mind information previously learned about that party, and thereby
increases the likelihood that the voter will cast a ballot for that choice. Cognitive priming,
according to Dr. Krosnick, provides a “small push that may happen (only) for some voters as
they walk into the voting booth.” (Tr. 733)
Dr. Krosnick’s Experiment: Dr. Krosnick conducted an experiment to test his hypothesis that
the presence of insignia might increase votes cast for the NUHW ballot choice. Dr. Krosnick’s
subsidiary hypothesis “was to investigate the impact of this insignia among two types of voters,
voters who were relatively well informed about the election and voters who were relatively
uninformed about the election.” (Tr. 733-34)
Dr. Krosnick designed and conducted a practical experiment modeled after features
similar to the Santa Rosa Memorial Hospital election (Santa Rosa election). (Tr. 749-750, 781-
783) Dr. Krosnick explained that it would be impossible to replicate the representation election
at the Hospital and that he never intended to do so. (Tr. 749-50, 781-83) However, Dr. Krosnick
claimed that all aspects of the experiment conformed to the scientific method, including survey
data collection, statistics, and experimental design, methodologies and disciplines that have
been recognized in the scientific community for over 70 years. (Tr. 725-27, 758-59)
To perform part of this experiment as well as the ballot design experiment addressed in
the consideration of Objection 9 below, Dr. Krosnick worked with Knowledge Networks, a survey
research company. Knowledge Networks utilizes “random digit dial telephone calls” to samples
of American households for the purpose of securing participation in online surveys. Participants
who join a survey research panel are paid for their time and participation in completing the
computer surveys. (Tr. 736) Dr. Krosnick indicated that he and many of his colleagues in
academia and the commercial world use Knowledge Networks regularly for survey research.
The work done by Knowledge Networks, Dr. Krosnick said, is considered reliable and relevant
within the scientific community in which he works. (Tr. 742)
Unlike the situation with the ballot experiment Dr Krosnick conducted that is discussed
below in connection with Objection 9, all of the hypothetical voters in this experiment were
exposed to the same ballot configuration. Immediately before the hypothetical voters were
asked to cast their votes, half of the voters were exposed to a screen showing the insignia of
one of the hypothetical unions, modeled in color, font, and appearance on the NUHW’s insignia.
This insignia appeared throughout the information the hypothetical voters previously received.
(Tr. 738-739) Dr. Krosnick admitted that that the experiment was limited to the appearance of
NUHW insignia “just moments before the vote was cast.” (Tr. 784) Dr. Krosnick acknowledged
that his insignia experiment did not take into account the presence of other insignia (such as
insignia that said “neither” on it) and admitted that “it’s certainly possible that the presence of
other insignias may have had impact.” (Tr. 784-786)
With respect to the “low information” voters in Dr. Krosnick’s experiment, the presence or
absence of the insignia at the last minute produced no significant difference in their voting
pattern. (Tr. 780) However, with respect to the “high information” voters who saw the insignia
screen immediately before voting, 14 percent voted for the hypothetical union, NUHW, whereas
the number of votes cast for the same union dropped to 8 percent among the high information
voters who did not see the insignia screen just before casting their ballots. (Tr. 746)
Thereafter, Dr. Krosnick performed a test to determine the statistical probability that this
6 percentage-point jump was due to the display of the insignia versus chance alone. (Tr. 746)
Initially Dr. Krosnick said there was a greater than 90 percent chance that the difference was
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
8
due to the insignia display (Tr. 746) but later Dr. Krosnick said that this figure was actually 97
percent. (Tr. 787) This lead Dr. Krosnick to conclude that there was an “extremely high
probability” that the last minute insignia display produced the 6 percent increase in the votes
received by the hypothetical union.11 (Tr. 746)
The fact that the actual election was so close also contributed to Dr. Krosnick’s opinion
that the presence of union insignia in the voting area influenced the outcome of the election. In
his opinion, the close results indicate that some voters, at the time of the vote, were either
conflicted or possessed information which “pushed them to opposite directions with regard to
their vote choice.” The relatively equal number of votes cast for the NUHW versus “neither” in
the actual election suggested to Dr. Krosnick that some of the available information encouraged
a vote for NUHW while other information encouraged a vote for “neither.” Dr. Krosnick further
explained that where “the information is evenly divided and voters see merit on both sides, they
are more often open to being pushed subtlety one direction or another.” (Tr. 747-49)
Finally, in Dr. Krosnick’s opinion, since a greater percentage of all hypothetical voters
cast votes for “neither” in the experimental election than in the actual election (8 to 14 percent
versus 52 percent), the estimated 6 percentage-point difference was most likely an under-
estimate of the impact of insignias on the actual election. (Tr. 749)
C. Further Findings and Conclusions
The Hospital argues that the election should be set aside because the wearing of
partisan insignia by the observers affected the outcome of the election. For this claim, it relies
heavily on the opinion evidence provided by Dr. Krosnick. Its brief also argues that the
presence of at least one observer festooned with pro-union insignia as described by field
examiner Smith merits rerunning the election.
Although the NUHW concedes that its observers wore partisan displays, it also argues
that none exceeded the type the Board has historically tolerated. As for Dr. Krosnick’s opinion
concerning the impact of the partisan displays, the NUHW argues that reliance on this evidence
would be tantamount to setting aside a whole body of law established by the Board and the
courts tolerant of observers who chose to wear insignia similar to those worn here. Moreover,
the NUHW contends that Dr. Krosnick’s experiment supporting his opinion about the effect of
the observers’ insignia on the election outcome is fatally flawed for several reasons.
I have concluded that this objection lacks merit. Contrary to the Employer’s contention,
the evidence does not establish that any individual served as an observer during any voting
period outfitted with such a bevy of partisan displays that would warrant the conclusion that it
amounted to outright electioneering within the confines of the voting area as the Hospital, in
essence, claims.
The Employer reliance on Board agent Smith’s testimony for this claim is misplaced.
Although Smith could not identify the particular observer, Miller, an agent of the Employer and
Tappeiner, an agent of the NUHW, agreed this person was employee Hartnett. These two
agents also agree that Hartnett removed all but one or two of the offensive display buttons and
reversed his T-shirt prior to the start of the voting session. Field attorney McPalmer’s testimony
is consistent as to the removal of almost all of insignia prior to the voting. In view of this general

11 On cross examination, Dr. Krosnick conceded 90 percent is only considered “marginally
statistically significant” whereas 97 percent is considered “statistically significant.” (Tr. 746, 787)
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
9
agreement, I find Board agent Smith’s account less reliable. The fact that Smith’s recollection is
at odds with all the others who testified on this point may well be due to the fact that that
McPalmer, not Smith, took the lead in dealing with the entire topic of observers wearing insignia
whenever it came up. But whatever the explanation for his testimony, I simply cannot credit
Smith’s recollection that one of the observers worked through a voting period festooned, as he
put it, with pro-union insignia.
Although the Board discourages observers from wearing campaign insignia during the
polling session, it does not prohibit that conduct. U-Haul Co. of Nevada, Inc., 341 NLRB 195,
196 (2004). Field attorney McPalmer made this position of the Board emphatically clear to the
parties on at least two occasions. On the second occasion, McPalmer’s emphasis on the
possible consequences obviously caused Tappeiner to persuade Hartnett to shed most of his
campaign paraphernalia before the voting session started. Without more, the Board does not
set aside elections based on the fact that an election observer wore some type of partisan
insignia during the polling period. Van Leer Containers, Inc. v. NLRB, 841 F.2d 779, 788 (7th
Cir. 1988) and the cases cited there. In the absence of any additional conduct by any observer
that accompanied the wearing of a partisan insignia, I find the evidence is insufficient to sustain
this objection.
Nor, in my judgment, does Dr. Krosnick’s opinion testimony warrant a different result.
Federal Rule of Evidence 702 provides for a tribunals consideration of expert opinion testimony
if it will assist the tribunal to “understand the evidence or to determine a fact in issue.” The
opinion of a qualified expert is admissible if (1) it is based upon sufficient facts or data, (2) it is
the product of reliable principles and methods, and (3) the expert has applied the principles and
methods reliably to the facts of the case. Fireman’s Fund Ins. Co.v. Canon USA, 394 F.3d
1054, 1057 (8th Cir. 2005).
As the Van Leer case cited above makes clear, the mere presence of observers in the
polling area wearing campaign insignia is not an issue in this case because the Board does not
prohibit this practice. Instead, to warrant setting aside an election, there must be evidence of
some added conduct accompanying the wearing of the campaign insignia by an observer.
Because the controlling precedent requires added conduct and there is none, Dr. Krosnick’s
opinion testimony is simply not “relevant to the task at hand.” Daubert I, 509 U.S. at 597;
Daubert v. Merrell Dow Pharmaceuticals, Inc., 43 F.3d 1311, 1320 (9th Cir. 1995) (Daubert II).
On that basis alone, I would find the expert opinion evidence submitted in this case inadmissible
under FRE Rule 702.
However, I also find Dr. Krosnick’s opinion as to the effect of the insignia worn by the
NUHW observers fails the reliability test required by Daubert for admissibility under Rule 702.
Although I do not doubt the scientific literature reviewed by Dr. Krosnick contains ample support
for the sociological concepts of conformity and cognitive priming, I am not satisfied that the
insignia experiment he developed to test the impact of these two concepts in a labor
representation election setting produced reliable information about the impact they would have
had on the voters in this election. I have reached this conclusion because his experiment: (1)
included irrelevant data; (2) failed to include sufficient facts necessary for the reliable application
of scientific principles and methods to the facts present in this case; and (3) failed to account for
alternative explanations.
As for the irrelevant data, the experiment included individuals selected from a nationwide
pool so that the vast majority of the participants were from areas well outside the geographic
location of the actual voters. It apparently incorporated individuals who worked in occupations
other than the healthcare field. The inclusion of individuals who differed in significant respects
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
10
demographically from the voters in this election detracts considerably from reliability of the
insignia experiment. City of Tuscaloosa v. Harcros Chemicals, Inc, 158 F.3d 548, 566 (11th Cir.
1998) (including data from other locations problematic regarding transactions that were not part
of the alleged conspiracy because the outside data skews any cumulative measurements, such
as percentages or frequencies, that depend upon the size and characteristics of the database.)
In addition, the insignia experiment did not relate reliably to facts of this case. See
Federal Rules of Evidence, Rule 702; Daubert I, 509 U.S. at 591-93. At the outset, the results
of the hypothetical election did not adequately replicate the results of the actual election to a
sufficient degree as to provide an assurance that the quality and quantity of information
presented to hypothetical voters was comparable to that available to the actual voters. This
principal shortcoming may well have been due to the fact that: (1) the handful of computer
screens the hypothetical voters received about the hypothetical election cannot be said to put
them in a similar position as the workers here who experienced five intense years of
campaigning that included numerous speeches, meetings, rallies, home visits with union
organizers, and intense discussions with co-workers and family members; (2) the randomized
sample of voters in the experiment lacked a similar stake with respect to their immediate terms
and conditions of employment that the workers in the actual election faced; (3) the results
showing a significant effect on high information voters by the appearance of the NUHW insignia
immediately prior to voting but no significant effect on low information voters would appear to be
at odds with scientific principles of conformity and cognitive priming tested by Dr. Krosnick;12
and (4) exposing hypothetical voters to a screen showing an NUHW insignia immediately before
casting their vote, fails to take into account the manner and degree to which the NUHW
insignias were displayed to the voters in the actual election.13
Furthermore, Dr. Krosnick’s experiment did not account for the concurrent presence of
the Hospital observers’ competing insignia. Even Dr. Krosnick acknowledged that this fact “may
have had impact.” (Tr. 784-86) The failure to account for alternative explanations detracts from
the reliability of the experiment on which the Hospital relies to establish the impact of the NUHW
insignia worn by that union’s observers. See e.g. Claar v. Burlington Northern R.R., 29 F.3d
499, 502-04 (9th Cir. 1994) (trial court properly excluded affidavits that did not attempt to rule
out other possible causes of plaintiffs’ injuries); McCorvey v. Baxter Healthcare Corp., 298 F.2d
1253, 1256-57 (11th Cir. 2002) (expert’s testimony was properly excluded as unreliable in a
strict products liability action against a manufacturer because the expert failed to test alternative
designs or test for a failure that could have come from other sources).
Finally, I note that Dr. Krosnick’s experiment was conducted for purposes of this litigation
and has not been peer reviewed or published. While that fact alone is not a sufficient basis to
reject his opinion, courts tend to be especially skeptical of scientific evidence that has not been
thoroughly peer reviewed. See Clausen v. M/V New Carissa, 339 F.3d 1049, 1056 (9th Cir.
2003); Brock v. Merrell Dow Pharmaceuticals, Inc, 874 F.2d 307, 313 (5th Cir. 1989).
Regardless, for reasons detailed above, I find that Dr. Krosnick’s opinion is not relevant to this
issue. But, even if it had relevance to the issue here, the experiment Dr Krosnick used to
quantify the effect of the partisan insignia worn by the union observers in this close election was

12 For example, as described below in the discussion related to Objection 9, Dr. Krosnick
found font differences among the choices on the ballot impacted low information voters but not
high information voters. The relationship between the voters’ knowledge level and the different
reactions to various visual stimuli has not been adequately explained.
13 Dr. Krosnick admitted that his experiment did not account for how, when, and in what
quantity the NUHW insignias had been displayed to the voters. (Tr. 784-786)
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
11
not sufficiently related to the facts of this case to merit consideration by the tribunal.
Accordingly, I recommend that the Board overrule Objection 1.
III. Objection 3
Objection 3: alleges that NUHW and its agents “engaged in surveillance of employees
attempting to exercise their Section 7 rights, videotaping employees attending employer-
sponsored meetings.”
A. The Relevant Evidence
The Hospital’s evidence concerning this objection focuses on two discrete incidents that
occurred on December 2 and December 7. It claims that on those two occasions NUHW agents
or representatives “engaged in surveillance and attempted to disrupt Employer sponsored
meetings.” (Er. Br., p. 9) Petitioner denies the agency status of the persons responsible for the
conduct involved and otherwise claims that the evidence fails to show the kind of interference
that would warrant setting aside the election results.
Between December 2 and December 11, the Hospital conducted 17 “Informed Choice”
meetings designed “to present the employees with factual examinations of the employees’
choices.” (Er. Br., p. 9) Employees received a schedule of the Informed Choice meetings by
email and a copy of the schedule was posted in several Hospital departments.
Each meeting lasted about 40 minutes or so. The format included a Power Point
presentation by Hospital staffers designed to provide employees with information about the
voting process, the participating unions, union dues, existing employee benefits, and the
Hospital’s position on unionization. The Employer paid employees for their time spent at an
Informed Choice meeting but attendance was not required.
The two meetings under scrutiny here were held in Conference Room A at the main
Hospital facility. In total, nine of the Informed Choice meetings were held in Conference Room
A, a rectangular enclosure located adjacent to a public cafeteria along a hallway leading from
the Hospital’s entrance lobby.
1. The December 2 Incident
On December 2, three non-employees, Martin Bennett, JoAnn Consiglieri, and Father
Raymond Decker, made separate, unsuccessful efforts to gain admittance to Conference Room
A, ostensibly to monitor the Informed Choice presentation scheduled for 12:30 p.m.
Rene May, the executive assistant to human resources vice president Miller, provided
logistical support for the instructors at the first Informed Choice meeting that day. After
arranging the room and setting up the equipment required for the presentation, she remained as
a monitor. May stationed herself near the entrance to the room to check that persons arriving
for the meeting had proper employee identification. As Bennett, Consiglieri, and Father Decker
each arrived and sought to enter, May politely told each that the meeting was for employees
only and that they could not attend.
All three lodged a mild protest. May said Bennett seemed annoyed at not being
admitted and told her that he planned to file a complaint the administration. (Tr. 193) Consiglieri
also appeared annoyed to May. She told May that the hospital must be hiding something if
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
12
townspeople were not being allowed to attend the meeting. (Tr. 195-96) Similarly, Father
Decker asked May, “(W)ell what are you hiding?” (Tr. 428)
May’s encounter with each was very brief.14 None of the three actually entered the
Conference Room A. None of the three made any effort to evade May, or to enter the meeting
after she prohibited them from entering. Although they may have been visible to employees
already admitted to the meeting, no evidence shows that May’s encounter with any one of the
three attracted any particular attention of those already at the meeting location. Likewise, no
evidence shows that this effort in any way disrupted the meeting or that they unduly blocked any
employee from entering Conference Room A to attend the meeting.
Michael Aparicio, a reporter from an internet publication called the Empire Report,
videotaped interviews of all three immediately outside the entrance to Conference Room A after
they had been denied admittance. He sought to obtain information about what they had
attempted to do, why they were doing it, and what had happened. Each made a brief statement
to Aparicio that he posted on You Tube. Father Decker claimed not to know Aparicio. He also
denied knowing about any prearranged plan to videotape the effort to gain admittance to the
Informed Choice meeting.
While videotaping his interviews, Aparicio captured views of others in the vicinity, some
employees and some not. From the videos in evidence as Employer Exhibit 8(b) 1 and 2,
Deborah Miller identified the Hospital employees. Those she identified are seen entering
Conference Room A, entering the restrooms across the corridor from Conference Room A, or
walking down the corridor.
Fawn Kraut, a business coordinator in the surgical services department who opposed
unionization in this election and in the earlier campaign five years ago, attended the first
Informed Choice meeting on December 2. She left her work place on the second floor of the
hospital close to the scheduled meeting time, proceeded to the lobby area on the first floor and
turned down the corridor leading from the lobby to Conference Room A. As she entered the
corridor she saw a group of five or six individuals, one of whom had a shoulder-held camera
within a couple of feet of the conference room door. At first, Kraut did not recognize any of the
individuals but she assumed they were physicians who regularly meet in that conference room.
This assumption led Kraut to also assume that the meeting probably had been relocated.
She walked past the group in the corridor to check for the meeting in one of the three other
conference rooms on down that corridor. After failing to locate the Informed Choice meeting in
these alternate places, Kraut walked back toward Conference Room A. As she neared that
location, she finally recognized some members of the group in the corridor. That group, Kraut
said, included Monsignor Brinkel as well as Father Decker and “a host of others who were part
of this support group for NUHW.” She suspected they were “trying to stage something to try to
get into the meeting” as had been done earlier when Steven Harper, another of the non-
employee NUHW supporters “kind of snuck in (to an earlier union-related hospital meeting) and
stood up and . . . asked questions . . . (before he was) asked to leave.”15 Kraut walked by the

14 Following the events in the hospital corridor as described below, this group went to
Hospital executive offices in the adjacent hospital administration building. There they sought a
meeting with the Hospital’s CEO and then dispersed after learning the CEO was out that day.
15 No further light was shed on the Harper incident. However, it probably accounts for
Miller’s caution in posting her assistant at the door to Conference Room A to insure that only
employees were admitted.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
13
group and on into the meeting. She estimated that she probably would have been a couple of
minutes late for the meeting in any circumstance but had been made even later because she
assumed at first that the meeting had likely been moved.
No one identified any organizer or other agents as a part of the group in the corridor
when the attempts were made to by the three non-employees to enter the Informed Choice
meeting on December 2. However, Philip Yunker, the northern California labor relations director
for the St. Joseph Health System (the Hospital’s parent company), claims that he saw Bennett,
Consiglieri, and Father Decker in the Hospital cafeteria in the company of union organizer
Tappeiner shortly before lunchtime that day. Tappeiner denied being present as claimed by
Yunker.16 He claimed that he had spent the day in a San Francisco courtroom for proceedings
involving one of the lawsuits the SEIU brought against NUHW.
Martin Bennett lives in the Santa Rosa area, works as an instructor at the Santa Rosa
Junior College, serves as a delegate to the North Bay Labor Council, and is active in a local
organization called the Fair Wage Coalition. He is among several academics nationwide who
have signed letters endorsing the NUHW in its many organizational campaigns and its disputes
with the SEIU.
JoAnn Consiglieri is a former member of the Order of Saint Joseph of Orange, the
Catholic order that owns the Saint Joseph Health Care System. She lives in the Santa Rosa
area and now works as a therapist. Consiglieri actively supported the Hospital workers efforts
to organize from 2005 onward.
Of the three turned away from the December 2 Informed Choice meeting only Father
Decker testified.17 He is a retired priest from the Archdiocese of San Francisco who now lives
in Sebastopol, California, a small, nearby city southwest of Santa Rosa. Father Decker, a
Roman Catholic priest for about 50 years, has been involved in social justice issues for many
years. He once served as the Director of Social Policy for the Catholic Diocese of Oakland.
Beginning with the SEIU-UHW-W organizing effort five years ago, Father Decker sought
to obtain the adoption of a “fair election agreement” by the Hospital that prohibits employer
intimidation of workers seeking union representation. According to Father Decker, a majority of
Catholic institutions in America have signed such an agreement. He has never been employed
by a labor organization and has never understood that he had authority to act on behalf of a
labor organization. However, Father Decker admittedly became involved in the unionization

16 Yunker identified Bennett, Consiglieri, and Father Decker from photos (Er. Exhs. 9 and
10) that he found on the Empire Report internet site. Yunker claimed that these two photos
depict a part of the group he observed at the cafeteria on December 2. However, whether the
photos were actually taken on December 2 is not clearly established. Indeed, whether these
two photos were even taken on the same day is not altogether certain, at least in my mind,
especially where, as here, the group pictured on Father Decker’s side of the table in Employer’s
Exhibit 9 is decidedly different than the group pictured on his side of the table in Employer’s
Exhibit 10. Regardless, Tappeiner identified the two persons depicted in Employer Exhibit 10 at
the far end of the table to the right of Father Decker as NUHW organizers Colleen Fewer and
Peter Clayton.
17 The Petitioner had Bennett available to testify at the outset of the afternoon session on
February 23. However, Bennett’s schedule and a delay in resuming at the appointed time
prevented the Petitioner from actually calling Bennett at that time. Petitioner made no known
attempt to secure Bennett’s testimony at a later time.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
14
campaigns at the Hospital because he lives in the area and because of his strong beliefs about
the application of Catholic social principles at Catholic institutions. He claimed that he has had
considerable correspondence with the Superior General of the Sisters of St. Joseph of Orange
about the application of Catholic social principles to the organizing campaigns at the Hospital
but seemed frustrated that the Hospital’s executives employed advisors who interpreted these
principles at odds with his understanding.
At some point before the election, both Bennett and Father Decker videotaped
endorsements of the NUHW. See Employer Exhibit 8(b)-4 and 5. On December 10, a group of
local residents, including Bennett, Consiglieri, and Father Decker, announced the formation of a
group calling itself the Fair Election Oversight Commission (FEOC). The FEOC, an ad hoc,
non-governmental community organization, lacked authority of any kind in connection with the
election process and otherwise served no known purpose other than as a front for local activists
to insinuate themselves into the process for resolving this question concerning representation.18
During the pre-election conference held on the evening December 16, the union parties
agreed to a proposal by the Hospital’s counsel that the no union agents would be permitted on
the Hospital’s premises during the periods the polls were open. Counsel also proposed that the
ban apply to the FEOC. Glen Goldstein stated at the time that the NUHW did not represent the
FEOC but the Hospital’s counsel disagreed and requested that Goldstein inform FEOC
members of the agreement the next time he saw them. There is no evidence that Goldstein
responded to this request concerning the FEOC. (Tr. 39-40)
At the conclusion of the pre-election conference meeting, the SEIU-UHW-W agent
questioned Deborah Miller about the enforcement of the ban as to the FEOC members if there
would be no management representatives in the public cafeteria area. Miller responded that
she would have a human resources employee check for FEOC members from time to time.
Miller said that Goldstein, who was standing next to her at the time but had not been a part of
the conversation, remarked, “(D)on’t worry, we won’t be sending them in.” (Tr. 114-15)
2. Gabriella Martinez and the December 7 Informed Choice Meeting
Until September 2009, Gabriella Martinez worked for Mission Hospital, an acute care
institution in Orange County California also owned and operated by the St. Joseph Health
System. During her employment at Mission Hospital, Martinez had been active in an on-going
organizational drive by the NUHW at that institution and served on an in-house employee
organizing committee there.
Beginning sometime in November, Martinez made several trips to the Santa Rosa area
to work on the Hospital campaign as a volunteer. She often visited the public cafeteria at the
Hospital to talk with employees about supporting the NUHW. Martinez as well as other
volunteers who supported the NUHW’s campaign at the Hospital attended some of the union’s
planning meetings along with the staff organizers assigned to this project. Tappeiner, one of the

18 In its brief, the Hospital sought to lump the activities of Aparicio, Bennett, Consiglieri, and
Father Decker together under the rubric of the FEOC from December 2 onward. Er. Br. 26, fn
12. The formation of the FEOC was not publicly announced until a week or more later, and no
evidence shows that any of the three non-employees who sought admission to the December 2
Informed Choice meeting or their interviewer referred to the FEOC in connection with the events
of that day. Moreover, the Employer’s December 8 letter protesting the events of that day to the
NUHW makes no reference to the FEOC. Er. Exh. 1.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
15
principal NUHW organizers during this organizing effort, denied that Martinez ever became an
agent of the NUHW like himself.19
On December 7 Kathleen DeLaney, who works for the Hospital as a human resource
services representative, provided the logistical support for the Informed Choice meeting in
Conference Room A that commenced at 3:45 p.m. Monica Scolieri, another human resources
employee, assisted DeLaney in arranging the room and setting up the computer for the Power
Point presentation. Sue Pierce and Michelle Huntley served as instructors for the session.
A window wall on the south side of Conference Room A separates it from an outdoor
area called the Healing Garden, a location set aside for reflection by patients’ families and
visitors. DeLaney and Scolieri arranged the meeting chairs perpendicular to the window wall.
When the meeting began, DeLaney and Scolieri sat in chairs along the wall opposite the
window wall with a view out into the Healing Garden.
A few minutes after the start of the meeting, Scolieri called DeLaney’s attention to two
women in the Healing Garden. DeLaney had not noticed anyone in that area when the meeting
began. Scolieri told DeLaney that it appeared as though they “were taking pictures of the
activity going on in the room.” One of the women, later identified to DeLaney as Martinez,
appeared to have a cell phone in her hand. 20 Occasionally, she would “get up from the bench,
walk around in certain areas and point (the device in her hand) towards the window.” Then, she
would go sit down on the bench and look at the device.
Although Delaney speculated at first that Martinez appeared to be taking pictures, she
later conceded that her motions were also consistent with reading e-mail on a cell phone.
DeLaney and Scolieri observed the two women in the Healing Garden for about ten minutes
before Scolieri closed the drapes so the persons in the Healing Garden would not distract the
employees in the meeting.21
DeLaney did not know if any employee at the meeting ever saw the two women in the
Healing Garden. No evidence shows that instructors Pierce or Huntley ever noticed the two
women in the Healing Garden nor is there evidence that their presence interfered with or
disrupted the progress of the Informed Choice presentation under way at the time. If any
photographs were taken that depict the meeting in progress that day, DeLaney never saw them
nor could she say for certain that any photographs were actually taken. Likewise, no testimony
was proffered from anyone claiming to have seen photographs of an Informed Choice meeting
in progress depicting a view from the Healing Garden on that date or any other date.

19 Tappeiner claimed that he worked without pay during this campaign. Pressed on cross-
examination to explain the difference between himself and Martinez, he quickly noted that he
was an agent and she was not. He went on to explain that, unlike Martinez, he had an
employment history with the predecessor organization, could “make decisions about the
direction of the campaign,” and had been delegated responsibilities by the NUHW leadership for
certain aspects of the campaign such as producing campaign materials. Tr. 568.
20 At the time, DeLaney did not recognize either of the women but about a month later
DeLaney identified Martinez as the woman with the cell phone or camera in her hand on
December 7 in a picture sent to Scolieri. (Tr. 311-14; Er. Exh. 7, p. 5, woman in green on the
left, Gabriella Martinez)
21 Vice president Miller said that Scolieri filed a report concerning the matter but, if so, it was
not offered in evidence.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
16
B. Further Findings and Conclusions
The Hospital contends that (1) Aparicio, Bennett, Consiglieri, Father Decker, and
Martinez are all agents of the NUHW; (2) the videotaping by Aparicio on December 2 and
Martinez’ conduct on December 7 amounted to surveillance of employees engaged in activity
protected by Section 7, i.e., attending an Informed Choice meeting, that warrants overturning
this election; and (3) NUHW agents concertedly planned to disrupt the December 2 meeting
with Bennett, Consiglieri, and Father Decker.
Petitioner denies that the individuals involved in the events of December 2 and 7 were
ever its agents. It also argues that the Hospital’s evidence fails to establish that any
surveillance occurred on those dates. Finally, Petitioner contends that employees were not
engaged in protected, Section 7 activities by attending the Hospital’s Informed Choice meetings.
In determining whether an individual is an agent of another, the Board applies the
common law principles of agency. Allegany Aggregates, 311 NLRB 1165, 1165 (1993);
Dentech Corp., 294 NLRB 924, 925–926 (1989). The burden of proving an agency
relationship is on the party asserting its existence. See Cornell Forge Co., 339 NLRB 733
(2003), and the cases cited there. An individual can be a party’s agent if the individual has
either actual or apparent authority to act on behalf of another party. Id.
I find that the Hospital failed to meet the burden of establishing that Michael Aparicio,
Martin Bennett, JoAnn Consiglieri, Father Raymond Decker, and Gabriella Martinez were
agents of the NUHW at any relevant time. No evidence shows that they have ever been
employed by the NUHW, or the SEIU-UHW-W. There is no evidence to show that the NUHW
said or did anything to convey actual authority on any of these individuals. Nor is there any
conduct by the NUHW that could have given employees reason to believe that these individuals
acted on the union’s behalf so as to support a finding of apparent authority. Kraut’s
characterization of them as a part of a longstanding union “support group” indicates that
employees recognized these individuals as nothing more than local citizens who backed
unionization. Nor is there evidence of any kind establishing that they have any authority to
speak or act on behalf of the NUHW or that they ever claimed to have such authority.
As for Martinez, the Employer argues she “acted in the same role as other NUHW
representatives during the campaign” even though it might appear that she amounted to little
more than an unpaid volunteer. (Er. Br., p. 48) Insofar as is known, she received no pay for her
efforts on behalf of the NUHW during the organizing campaign at the Hospital. However, the
Employer points out that even the NUHW organizers admitted to be agents were not paid. But
clearly, Martinez is not in that class as she had never worked as an organizer for the Intervenor
prior to the trusteeship as did all of the other admitted NUHW organizers serving this campaign.
Moreover, the mere fact that Martinez actively promoted the NUHW to the Hospital’s
employees, or served on an employee organizing committee in support of the NUHW at her
former place of employment fails to establish that she was an NUHW agent during this
organizing drive. United Builders Supply Co., 287 NLRB 1364, 1364 (1988) (prounion
employees do not become union agents merely because of their vocal and active union
support.); Tuf-Flex Glass v. NLRB, 715 F.2d 291, 296 (7th Cir. 1983) (same); Advance Products
Corp., 304 NLRB 436, 436 (1991) (employee members of an in-plant organizing committee are
not, simply by virtue of such membership, agents of the union); Uniroyal Technology Corp. v.
NLRB, 98 F.3d 993, 999–1000 (7th Cir. 1996) (same). Both Goldstein and Tappeiner credibly
denied that Martinez had any authority to speak for or bind the NUHW. Likewise, no evidence
supports a conclusion that Martinez was vested with apparent authority to act on behalf of the
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
17
NUHW as there is no showing that anyone with authority from the NUHW said or did anything
that would lead a reasonable person to believe that she had authority to act for the NUHW.
Other minor incidents on which the hospital relies also fall far short of showing the
agency relationship required here. For example, even assuming the parties reached an
informal verbal agreement that only employees and agents would be admitted to the small room
where the vote count was held on December 18, lead organizer Goldstein’s effort to secure the
admittance of Martinez and Consiglieri does not amount to an admission that they were NUHW
agents in any legal sense. Similarly, evidence showing that Martinez may have been in a group
that included the leader of the NUHW on one occasion in the Hospital’s cafeteria is also
insufficient to vest her with authority to act on behalf of that union with respect to anything.
Absent a showing that these individuals were agents of the NUHW or any other party in
this proceeding, their conduct is measured by the principles applicable to outside parties. The
Board will set aside elections based on outside party conduct but only where it is so aggravated
as to create a general atmosphere of fear and reprisal so as to render a free election
impossible. Robert Orr-Sysco Food Services, 338 NLRB 614, 615 (2002), citing Westwood
Horizons Hotel, 270 NLRB 802 (1984) (Westwood). Usually, less weight is accorded to third
party conduct because it tends to have less effect upon voters than similar conduct by a party
and because employers and unions usually cannot prevent misdeeds by persons over whom
they have no control. Deffenbaugh Industries, Inc., v. NLRB, 122 F.3d 582, 586 (8th Cir. 1997).
The conduct shown to have occurred on December 2 and 7 that are the subject of this
objection falls far short of that required under the standard applicable to conduct by an outside
party. The effort on the part of Bennett, Consiglieri, and Father Decker to gain entrance to the
December 2 Informed Choice meeting caused no disruption and interfered with no employee
gaining access to the meeting insofar as is known. Their conduct, though determined to a point,
remained dignified and respectful throughout the period under scrutiny. Likewise, Aparicio
made no deliberate effort to film employees seeking to attend the meeting. To the extent that
employees were filmed, it was incidental to the interviews he conducted with Bennett
Consiglieri, and Father Decker in the Hospital corridor that is used by employees and the public
alike. No evidence shows that any employee suffered any form of retribution as a result or that
any attempt was made to intimidate employees filmed during the interviews.
Nor does the conduct attributed Martinez on December 7 meet the Westwood standard
described above. Even assuming that she photographed the meeting as charged, no evidence
shows that anyone other than DeLaney and Scolieri was made aware of that fact then or at any
time prior to the election. The conduct did not disrupt the meeting in any fashion. No showing
has been made that retribution of any kind was visited on any unit employee who attended the
meeting that day and no evidence shows that any photographs from the meeting whether from
this potential source or any other source were used in any manner let alone a manner designed
to intimidate or frighten unit employees from attending the Informed Choice meetings.
For the foregoing reasons, I find insufficient evidence to support the allegation in
Objection 3. Accordingly, I will recommend that this objection be overruled.
IV. Objection 8
Objection 8 alleges that “Board agents failed to monitor and prevent improper conduct by
employees in the voting area.”
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
18
A. The Relevant Evidence
The Hospital adduced evidence about three incidents that occurred during the initial
voting session on December 17 and claims that the Board Agents improperly handled the
situations that arose. The Employer relies primarily on accounts of three incidents provided by
its observer Cathi Charlberg. Board agents McPalmer and Smith testified in connection with
this objection as did Donald Fugate, the NUHW’s observer at that session. The third Board
agent, field attorney Katz, did not testify.
As explained earlier, the polling area consisted of two adjacent conference rooms. Field
attorney McPalmer oversaw the arrival of voters in the outer room. Field examiner Smith and
field attorney Katz worked in the other room where the observers, three voting booths, and a
ballot box were located.
According to Charlberg, about 30 to 45 minutes into the voting, three female employees
came to the polling area. The first two checked in, received a ballot, and entered two adjacent
voting booths.22 Charlberg, who does not speak or understand Spanish, said that shortly
thereafter the two women began talking in Spanish to one another from their respective voting
booths. Board Agent Katz, stationed a short distance from the voting booths near the ballot
box, promptly told the two women to stop the conversation, that talking was not allowed in the
voting booth. When they continued their talk, Katz again admonished them to stop. At that, the
one of the two women, who was never identified by name, came out of the voting booth with her
ballot in hand and began to walk away. Katz stopped her, informed the woman that she could
not “leave with her ballot,” and asked if she could help. The woman told Katz that she was not
ready to vote yet.
Another Board agent, probably field examiner Smith, approached about that time and
asked if he could be of assistance. The woman repeated that she was not yet ready to vote.
Charlberg claims that one of the Board agents then told the woman “that if she left now she
couldn’t vote.” Tr. 245. When the woman she started to leave again, Katz told her she could
not take the ballot. With that, the voter held the ballot out to Katz. Purportedly, Katz then
warned the woman that she could not vote if she left. Tr. 246.
Charlberg recalled the second voter then came out of the booth and spoke to the other
voter in Spanish. One of the Board agents warned again that there was to be no talking in the
voting area. Then the Board agent fluent in Spanish (McPalmer) spoke to the first voter in
Spanish for a brief period. With that the voter turned and went to the observers’ table where
she paused briefly in front of the NUHW observer Fugate, gave a puzzled shrug, said that she
was ready to vote, and proceeded back into the voting booth. Charlberg said the second of the
two voters also returned to the adjacent voting booth. Tr. 247.
To Charlberg it appeared as though the voter who earlier wanted to leave stepped
sideways toward booth of the other voter and reached her arm out full length with her index
finger pointing in the direction of or into the booth occupied by the other voter. Charlberg did not
see a ballot in the hand of the voter that reached out toward the other voting booth and the two

22 The three voting booths were positioned in a row adjacent to one another slightly forward
and to the left of the observers’ table. As described by Charlberg, there would have been some
space between each voting booth. The voting booth entrances faced a wall on the side
opposite from the observers table so that the observers would only have been able to see
voters from the waist down after they entered a booth.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
19
again began speaking in Spanish. (Tr. 267-69) Board agent Katz admonished the voters that
only one voter at a time was allowed in a voting booth and that they were not allowed to speak
in the voting area. However, the voters continued their conversation until Katz warned them
again. (Tr. 248)
Meanwhile, throughout this prolonged incident, Charlberg said the third voter who
arrived in their company remained at the observer table until the first voter finally emerged the
second time from the voting booth to deposit her ballot in the ballot box. Shortly thereafter the
second of the voters emerged and put her ballot in the ballot box and the two left the voting
area. Only then did the observers check off the name of the third voter so she could vote.
During the second hour of voting, Charlberg said that another Spanish speaking woman
approached the table, got a ballot, and went into the voting booth. Shortly after, she came out
and said in English that she was confused. With that the Spanish speaking Board agent
(McPalmer) “asked in English and Spanish if she needed clarity, if she needed more time or
whatever the issue was.” The woman then stated, according to Charlberg, that she wanted to
leave but the “female agent (Katz) told her she couldn’t leave with the ballot, without voting”
after which the voter proceeded to vote. (Tr. 251-53)
Near the end of this voting period, Charlberg said another Spanish speaking employee
approached the observers and asked for a ballot in Spanish. The Spanish speaking Board
Agent came forward and spoke to her in Spanish, and then took her aside for a further
conversation in Spanish. After their conversation, the employee then voted.23
NUHW observer Fugate’s recollection differs significantly from Charlberg’s. Fugate
recalled that the two Hispanic employees were processed through at the observers table,
received ballots from the Board agent, and then entered adjacent voting booths to mark their
ballots. He first noticed a problem existed when one of the two stepped back out into the voting
area from the booth with a ballot in her hand and had a conversation in Spanish with one of the
NLRB agents. After at brief exchange in Spanish, the second woman came out of her voting
booth “and started speaking in English trying to explain . . . that the first lady wasn’t sure how
she wanted to vote, that . . . she wanted time to think about it.” (Tr. 363) According to Fugate,
the Board agent then stated: “[W]e’ve already marked her off the list, she’s received a ballot,
she has to vote now or not at all, at any rate, she certainly cannot leave the area with her ballot.”
There followed, Fugate said, a brief discussion between the two women in Spanish after
which they both returned to their respective voting booths. Back in their voting booths, the two
women again spoke in Spanish to each for a short period until the Board agent told them to stop
talking. Fugate never saw either one reach into or enter the other’s voting booth. He recalled
that the woman who had explained to the Board agent the difficulty her acquaintance was

23 The NUHW objected at the hearing and in its post-hearing brief to any consideration of
these latter two incidents described by Charlberg on the ground that they were not addressed in
the RD’s Report. I overruled that objection at the hearing and reaffirm that ruling here. In its
brief, the NUHW cites Precision Products Group, 319 NLRB 640 (1995); Bell Halter, Inc., 276
NLRB 1208 (1985); and Iowa Lamb Corp., 275 NLRB 185 (1985) in support of its position that
this evidence should not be considered. As I concluded at the hearing, and continue to believe
now, these two additional incidents described by Charlberg merit consideration because they
arguably relate closely to the claim in this objection that the Board agents “failed to monitor and
prevent improper conduct by employees in the voting area.” I find the cases cited by the NUHW
factually distinguishable from the situation here.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
20
having a short time earlier, exited the voting booth, put her ballot in the ballot box, and left the
area. Shortly thereafter the voter who earlier sought added time exited her voting booth,
deposited her ballot in the box, and left.
Fugate denied there were other occasions during the first voting session when a Board
agent spoke to a voter in Spanish. He also denied that any one of the Board agents took a
voter aside to another part of the room and spoke in Spanish to the voter.
With respect to first incident Charlberg described, field examiner Smith recalled that two
women presented themselves, and went through the regular process to obtain a ballot and vote.
After they entered their respective voting booths, Smith overheard them speaking to one in
Spanish. When Smith took a few steps in the direction of the booths to warn them against
talking, both women emerged from the voting booths. One of the women cast her ballot
immediately. The second walked up to Smith, held her ballot out to him, and told him in broken
English that she wanted to vote later. Smith said that he never took the woman’s ballot.
Instead, he told the woman she should vote now but was uncertain if she understood him so he
ask field attorney Katz to summon field attorney McPalmer to talk with the woman in Spanish.
McPalmer said that when he entered the main voting room, he saw field examiner Smith
with a ballot in his hand, standing in front of the voting booths with two women. Smith explained
to McPalmer that one of the women had returned from the voting booth asking to return her
ballot and he wanted McPalmer to speak with her. A second woman stood nearby. McPalmer
had the impression that the second woman had already cast her ballot but could not recall how
he gained that impression.
McPalmer asked the woman having difficulty about her problem. She told him that she
wanted to go to lunch and speak with someone about how she should vote. McPalmer told the
woman that the ballot contained all the information about the election in English and Spanish so
she should be able to vote. He added the, “it’s necessary for you to vote at this time.”
McPalmer said that he paused and asked if it was possible for her to vote. Then McPalmer
again told the woman, with added emphasis, that it was necessary for her vote at that time.24
(Tr. 682) With that, the woman retrieved her ballot from Smith, and entered a voting booth.
McPalmer and the second woman waited a few feet away from the voting booths until the voter
emerged, deposited her ballot in the ballot box, and left with the other woman. (Tr. 661-63)
About half an hour later, field attorney Katz summoned McPalmer back to the main
voting area and told him that “we have the same problem, this woman is trying to give her ballot
back and she appears to be a Spanish speaker.” (Tr. 664) McPalmer again explained to the
voter (also unidentified) that all of the instructions had been given in English and Spanish and
then asked, “(I)s it possible for you to vote at this time?” The woman apparently did not
respond. Instead, she retrieved her ballot from Katz, reentered the voting booth, and came out
shortly thereafter and put her ballot in the ballot box.
Neither of the observers nor any of the parties’ representatives at the election sought an
explanation from any of the Board agents overseeing the election as to what the voters had

24 When asked to explain his choice of words, McPalmer explained that it was the Spanish
that came to him at the time. Tr. 683. McPalmer is not a native speaker of the Spanish
language nor does he profess to have a broad facility with that language. His knowledge of the
language comes largely from high school and college courses as well as his personal effort to
learn the language.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
21
been told in the Spanish language conversations Smith conducted. Both Smith and McPalmer
claim these were the only two incidents that took place during the first voting session.
B. Further Findings and Conclusions
The Hospital contends that Board agents did not adequately monitor and prevent
improper conduct by employees in the voting area. The Hospital claims two voters talked in the
voting area on multiple occasions and were observed in the in the same voting booth at the
same time. This conduct and the fact that Board agents were observed handling the voter
ballots, the Hospital claims, compromised the secrecy of at least two ballots. The Hospital also
argues that due to the closeness of the election result, the ballots of the two voters who sought
to leave should not have been comingled with the other ballots so that the issues pertaining to
their unusual conduct could have been resolved before their ballots were counted. Finally, the
Hospital argues that at least three confused voters forfeited their right to vote because the
Board agents compelled them to vote at a particular time.
The NUHW, relying on the accounts provided by Board agents McPalmer and Smith
rather than either its own observer or the Hospital’s, disputes the Hospital’s charge that the two
voters were ever in the same voting booth together. In addition, NUHW argues that the Board
agents did not abuse their discretion by telling the reluctant voters who had already been given
ballots that it was necessary to vote before leaving the voting area. But even assuming that the
Board agents may have committed a procedural irregularity in their handling of the reluctant
voters, the NUHW argues it was not sufficient to warrant setting aside the election because the
Hospital has failed to raise a reasonable doubt as to the fairness and validity of the election. As
for the other two incidents described by Charlberg, the NUHW asserts even her version shows
that they were unremarkable.
Although the accounts provided about the events that occurred during the first voting
session vary in their detail to a surprising degree, I find McPalmer’s account the most reliable.
He was called upon, in effect, to resolve the unusual situations that developed because he was
the only Board agent present with some understanding of the Spanish language and the
situations involved voters more conversant in that language. In view of his central role, I find it
more likely that the details of the incidents would have made a stronger impression on him than
any of the others in the vicinity. Moreover, McPalmer’s demeanor reflected confidence about
those facts that he could recall as well as an honest acknowledgment of facts he could not
recall. The balance between the two I found consistent with the passage of time and the role
that he played at that session.
Hence, on points critical to the arguments the parties have made on this issue, I find with
respect to the first incident that only one voter returned to the voting booth after the two
emerged, and that the more hesitant voter gave her ballot to field examiner Smith when she first
emerged and retrieved it from him after deciding to return to the voting booth. I do not credit
Charlberg’s claim that this hesitant voter approached the observer’s table before reentering the
voting booth and made a puzzled gesture in front of NUHW observer. Finally, I find that only
two incidents rather than three occurred during this voting session, and that both involved voters
who made a last second effort to vote at a later time. Charlberg, in my judgment, most likely
saw the second incident as two separate events because the voter first spoke with field attorney
Katz who then left to retrieve McPalmer to speak with the voter in Spanish.
I am satisfied that that the secrecy of the voters’ ballots involved in these incidents
remained inviolate. Contrary, to the Hospital’s assertion, the evidence fails to establish that any
voter entered the booth of another voter in a manner that would have jeopardized ballot
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
22
secrecy. Even crediting CharIberg’s account of the extended arm, this evidence fails to
establish anything untoward occurred that might serve as a signal between the adjacent voters
about their voting intentions. In fact, the record lacks any evidence showing that the second
voter ever became aware of the extended arm or that she reacted in any physical fashion this
conduct in any observable way. Simply put, I find that the reaching-out gesture that Charlberg
described is far too ambiguous to have any significant meaning at all.
I find the Board agents acted within their discretion when they requested that two of the
first session voters cast the ballots that had been given before leaving the polling area. Both
voters had been checked in on the eligibility list, given a ballot, and entered the voting booth
once before becoming overcome with apparent indecisiveness. Although the Hospital argues
that no Board rule would have precluded the Board agents from acceding to the voters’ requests
to leave and return at a later time, following such a procedure would greatly increase the
likelihood of confusion, errors, disputes, and added delays due to indecisive election results.
This dispute is about an accommodation the two voters sought to vote later rather than
an emergency requiring either of them to depart immediately before casting the ballot already
provided. This dispute does not concern a voter who decided belatedly that she did not want to
vote at all. On the other hand, it does not concern a situation where the election procedures
made it impossible to accommodate the requests for added time to consider their franchise.
Ordinarily, when a voter appears to vote, the observers locate the voter’s name on the
official voter eligibility list and each will check that name off by placing a mark next to the name
to signify that the voter has been cleared to vote. Most of the time, but not always, one or both
of the observers can tell the Board agent that the voter is who she says she is, and the Board
agent will then give the voter a ballot and direct her to an available voting booth. Arguably, the
only thing necessary in a representation election when a voter requests to leave and return later
after having received a ballot would be to return the unmarked ballot and cross out the marks
used on the eligibility list by the observers to signify that the person had voted.
Simple as that might sound, I am confident that detached observers of representation
elections would agree that such a procedure would be fraught with dangers. In the highly
charged atmosphere often found in these elections, such a highly unusual election procedure
would likely produce protracted disputes about whether such voters had or had not voted. For
example, it would be naive to believe such a process could be routinely accommodated without
quarrels arising from time to time between observers as to whether an employee had already
voted that would leave the Board agent entirely confounded. The probability that this would
occur most likely would increase where, as here, multiple election sessions had been planned
and the parties used different observers at each session. Such conflicts and confusion would
most likely lead to an increase in the number of challenged ballots and more indecisive
elections than already occur. Although the evidence fails to show whether voters were
permitted to vote on work time here, by accommodating voters in the manner sought by these
two employees could lead to abuse and disputes in those situations.
Hence, I recommend that the Board overrule Objection 8 and abide by the judgments
made by the election agents at the scene. The evidence fails to show that they acted arbitrarily
or capriciously by insisting that employees who already had received their ballots vote at that
time rather than later. Mead Southern Wood Products, 337 NLRB 497, 498 (2002) (Board
agent has broad discretion in deciding the details of an election and Board will abide by the
agent’s judgment unless it is arbitrary or capricious). Similarly, I reject the Hospital’s assertion
that the ballots of the two voters involved should have been segregated. No one claims they
were ineligible voters and no other reason existed to segregate their ballots.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
23
V. Objection 9
Objection 9 alleges that the “Board created a ballot that displayed the employees' choices in
different fonts and font styles thereby confusing voters and/or creating the impression that the
NLRB disfavored the employer and considered a vote for ‘Neither’ was not on equal terms and
was of lesser significance to a vote for either union.”
A. The Relevant Evidence
To poll the unit employees, the Regional Director’s agents utilized the Board’s standard
English/Spanish, three-choice ballot format. The Stipulated Election Agreement (SEA) specified
the location of the choices on the ballot from left to right as follows: 1) National Union of Health
Care Workers (NUHW); 2) Neither; and 3) SEIU-HWA, United Healthcare Workers-West.
Employer’s Exhibit 2, a copy of the ballot declared void at the vote count, shows the
ballot created by the Regional Office for use in this election. It reflects that the three choices
appeared horizontally on the ballot. As required by the terms of the SEA, the NUHW choice
appeared as the ballot choice on the left, the “Neither/Ninguna” choice (Neither choice)
appeared in the middle, and the SEIU-UHW-W choice appeared on the right. The two union
choices ballot appeared on the ballot in a serif font with bolding and the Neither choice
appeared in a sans serif font without bolding.
The appearance of the ballot became known, or should have become known, to all
parties a week or more prior to the election when the Regional Director forwarded copies of the
standard Notice of Election to the Employer for posting around the facilities involved. The
Notice of Election contains an exact duplication of the ballot to be used in the election albeit in a
reduced size to fit the standard Notice of Election form.
The standard Notice of Election form, which measures 14 inches high and 25-1/4 inches
wide, contains a two-line heading spread across the width of the document. The first line reads
“UNITED STATES OF AMERICA (five point star) NATIONAL LABOR RELATIONS BOARD”
and the second line reads: “NOTICE OF ELECTION” with the Seal of the United States
appearing on the left and the Seal of the National Labor Relations Board on the right. A three-
line footer in capital letters also runs the width of the page warning against defacing the notice
of election, disclaiming Board responsibility for any extraneous markings that might appear on
the notice or the sample ballot contained on the notice, and concluding with the statement that
that the NLRB “does not endorse any choice in the election.” This exact disclaimer language
also appears on the right panel of the Notice of Election form in the paragraph below the
heading “The National Labor Relations Board protects your right to a free election.”
In addition, the ballots used in this election contain the standard disclaimer mandated by
the Board in Ryder Memorial Hospital, 351 NLRB 214 (2007). That disclaimer reads: “The
National Labor Relations Board does not endorse any choice in this election. Any markings that
you may see on any sample ballot have not been put there by the National Labor Relations
Board.” (Er. Exh. 2)
Between the header and the footer, the Notice of Election contains three panels of text
that addresses election details, employee rights, prohibited election conduct, and the role of the
National Labor Relations Board in the election. The two side panels contain pre-printed text that
appears on all of the Board’s representation election notices. The center panel contains text
about the particular election that is inserted mechanically by the regional office before the
notices are mailed to the employer for posting. The center-panel information shows the
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
24
categories of employees eligible to vote, those categories not eligible to vote, the date, time and
place of the election, and a sample ballot.
The classification-by-classification unit description the parties chose in the SEA
precluded printing the center panel information in the ordinary manner. Instead, the regional
office personnel responsible for preparing the election material prepared an attachment with the
necessary election-specific information and affixed it to the Notice of Election form. The sample
of the ballot used in the election appears at the bottom of the attachment shrunken to fit the
attachment. This smaller version contains the same typeface for the ballot choices that
appeared on the ballots used in the election. As the attachment extended below the notice of
election form, the lower portion was not affixed to the form so that the bottom portion of the
attachment could be lifted in order to read the warning language along the bottom of the form.
The Hospital’s counsel conceded that he had received the Notice of Election containing
a shrunken reproduction of the ballot in advance of the election. However, he first called
attention to the difference in fonts and the lack of bolding for the Neither choice when he saw
the void ballot during the vote count following the closing of the polls. (Tr. 52) Insofar as is
known, the Hospital posted numerous copies of the Notice of Election in compliance with the
Board’s rules.25 Hospital’s counsel conceded that he did not object to the ballot format after
receiving the Notice of Election and no evidence establishes that any other party objected or
called attention to the ballot format in advance of the actual balloting.
NUHW election observer Fugate claimed that one of the Board agents (presumably field
attorney McPalmer based on Fugate’s limited identification) displayed an unmarked ballot and
solicited objections to its use. No one objected to the ballot. (Tr. 361) Those present,
according to Fugate, included Employer attorney McClune. McPalmer, admittedly present for
this conference, had no recollection that he or any other Board agent displayed a ballot to the
representatives during this gathering. (Tr. 657-58) The Hospital and the NUHW later stipulated
that McClune had not been present in Santa Rosa for the pre-balloting conference that morning.
B. The Opinion Evidence of Dr. Krosnick
In addition to the evidence concerning the ballot format, the Hospital also offered the
opinion of Dr. Krosnick concerning the impact of the ballot design on the voters’ selection.
Dr. Krosnick testified that in his opinion the ballot format in this election with the two union
choices appearing in a serif font with bolding and the Neither choice appearing in a sans serif
font without bolding “increased the number of votes cast for the NUHW and reduced the number
of votes case for ‘(N)either.’” (Tr. 728, 753, 755, 758; Er. Exh. 19, p. 11-12) It was Dr. Krosnick’s
further opinion that if the Neither option had been in bold, the NUHW would have lost at least
four percentage points of votes so that it would not have won the election outright. (Tr. 761) Dr.
Krosnick admitted he could not offer an opinion about whether the ballot was or was not
confusing and admitted that his opinion admitted had nothing to do with whether the ballot

25 Notices of Election must be posted in conspicuous places at least 3 full working days (a
working day means an entire 24-hour period excluding Saturdays, Sundays, and holidays) prior
to 12:01 a.m. of the day of the election. See Section 103.20 of the NLRB Rules and
Regulations. Compliance with this posting requirement would mean that the notices needed to
be posted by midnight December 13. As the election time neared, the regional office mailed the
Employer different Notice of Election forms as there apparently had been discussion or
speculation that the Intervenor might withdraw at the last minute. That never happened. Hence,
the Hospital posted the original Notice of Election forms sent out by the regional office.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
25
conveyed a preference by the NLRB for a particular election choice. (Tr. 791, 794) Dr. Krosnick
based his opinions on two bodies of evidence: (1) the theories and research found in the
scientific literature published prior to this case; and (2) the results of an experiment he
conducted for purposes of this litigation. (Tr. 725, 755-58)
The Scientific Literature. According to Dr. Krosnick, the scientific research and literature
supports the conclusion that “small aspects of the design of a ballot can change voters’
behavior, when casting their votes.” (Tr. 756; 755-58; Er. Exh. 19, p. 4) Dr. Krosnick relied on
two general areas of research concerning ballot design generally labeled order and prominence.
(Tr. 756, 790) With regard to order, Dr. Krosnick testified that “[t]he findings of (published, peer-
reviewed) studies are overwhelmingly consistent with a simple conclusion: candidates listed first
on the ballot have an advantage.” (Er. Exh. 19, p. 4-6; Tr. 756) Dr. Krosnick said the literature
shows that a candidate whose name appears first on a ballot will gain approximately three
percent more votes than would have occurred had he/she been listed later on the ballot and
conversely that the second or later choice will receive three percent less votes, a “swing” of six
percent. (Tr. 756-57) He said that “[t]here is very little research evidence inconsistent with this
claim, and veritable mountain of evidence consistent with it.” (Er. Exh. 19, p. 4-6)
As for the impact of prominence on voter behavior, Dr. Krosnick stated that studies have
shown that trivial aspects of a ballot, such as wording, may distort an election outcome. (Er.
Exh. 19, p. 8-9; Tr. 757-58, 790) In this connection, Dr. Krosnick said that the principle of
perceptual salience recognized in the literature on cognitive and social psychology comes into
play. (Tr. 757, 790; Er. Exh. 19, p. 8-9) The notion of perceptual salience, Dr. Krosnick said,
explains why a human brain responds as it does when exposed to a particular visual display. A
display, Dr. Krosnick explained, causes the brain to focus “on the visually salient stimuli that
(are) larger, more colorful, more active, (and) outstanding against the background.” (Tr. 757)
Dr. Krosnick asserted that the bolding of the two union names made them, by definition, more
salient on the ballot so that individuals likely focused their perceptual attention, on those more
visible, prominent aspects of the ballot and paid less attention to the Neither option, not by
choice but simply by the way the brain works. (Tr. 757-58) Even though the Neither choice on
the actual ballot used in this election differed from the two union choices, Dr. Krosnick asserted
that did not make it more salient. The key factors, he explained, was that the Neither choice
“was different and lighter, less visible, more receding in size and prominence, and that’s what
will make it less salient.” (Tr. 805-06)
Dr. Krosnick’s Experiment. Dr. Krosnick devised an experiment to test the hypothesis he
developed from the scientific literature that the bolding of the first selection on the ballot
(NUHW) and the failure to bold the second selection (Neither) increased votes for the NUHW
and decreased votes cast for Neither. (Tr. 725, 759) He modeled his experiment after features
he deemed similar to the Santa Rosa Memorial Hospital election (Santa Rosa election). (Tr.
749-750, 781-783) However, Dr. Krosnick acknowledged that he did not intend to replicate the
Santa Rosa election, as that would be impossible. (Tr. 749-750, 781-783) But Dr. Krosnick
claimed that all aspects of his experiment conformed to the scientific method, including survey
data collection, statistics, and experimental design, all three being methodologies and
disciplines the scientific community has recognized for over 70 years. (Tr. 725-27, 758-59)
Dr. Krosnick again used the services of Knowledge Networks to perform part of his
experiment. (Tr. 736) Knowledge Networks created a random nationwide sample that reflected
some of the demographic characteristics of the individuals eligible to vote in the Santa Rosa
election. (Tr. 737-738, 802) Dr. Krosnick surveyed a total of approximately 550 respondents
(hypothetical voters), who were given information that they would be voting in a hypothetical
election for hospital employees. (Tr. 737, 798) Dr. Krosnick divided the hypothetical voters in
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
26
his experiment into high and low information categories based on separate levels of information
provided before they were asked to cast ballots. (Tr. 779) The varying levels of information
provided to the hypothetical voters were modeled after the information provided at the Hospital’s
Informed Choice meetings, and the messages contained in various campaign materials
produced by both unions. (Tr. 737-738, 759)
Low information voters saw 3 to 4 screens of information for each hypothetical election
choice, with 3 to 4 bullet points on each screen. (Tr. 800-801, 810-811) High information voters
saw additional screens of information but it is unclear how many screens they were shown
altogether. (Tr. 801) Dr. Krosnick cautioned that in terms of knowledge about the election he
could not say the low information hypothetical voters replicated the actual voters in any way.
(Tr. 801) After they were shown the information provided, the hypothetical voters were asked
to cast their ballot. (Tr. 738) However, half of the hypothetical voters saw a ballot as appeared
in the Santa Rosa election with the Neither choice in a sans serif font without bolding. The other
half of hypothetical voters were provided a ballot with all three choices appearing in the same
bold serif font used for the two unions in the actual ballot format. (Tr. 759-760, 803)
About 70 percent of the low-information hypothetical voters who saw the Neither choice
without a bold font voted for that choice whereas 19 percent voted for the hypothetical NUHW
choice. (Tr. 749, 760-761, 819-820) About 74 percent of the low-information hypothetical
voters who saw the Neither choice in the same bold font as that used for the unions voted for
that choice, and about 15 percent voted for the hypothetical NUHW. (Tr. 749, 760-761, 819-
820) The results for the high information voters were essentially the same regardless of the
ballot design. (Tr. 779, 803)
Dr. Krosnick used a statistical analysis to determine whether this four percentage-point
drop for low information voters resulted from the bolding of the Neither option in the experiment.
(Tr. 760, 798) Based on his analysis, Dr. Krosnick said there was a greater than 80 percent
chance that the difference was due to the bolding as opposed to chance alone. (Tr. 760, 798)
Dr. Krosnick said that 80 percent is “informative,” or an “extremely high probability” and,
therefore, one which would lead him to conclude that it was due to bolding. (Tr. 760, 799)
Dr. Krosnick was never asked to define the term “informative.” However, on cross-
examination he conceded that the 80 percent value historically has not been considered
“statistically significant.” (Tr. 799) Dr. Krosnick asserted that the term “statistically significant”
came to be defined as a probability value of 95 or above primarily because of an opinion
expressed by statistician R.A. Fisher many years ago. Dr. Krosnick said some scientists now
believe that Fisher adopted the 95 percent value arbitrarily as a convenience for a book he
published in order to avoid endless tables with of lots of numbers. In the last 15 years, Dr.
Krosnick explained, a large amount of literature has developed arguing that the term
“statistically significant” pegged at Fisher’s 95 percent value should largely be ignored. The
better practice today, Dr. Krosnick said, is to describe the actual probability and to base a
judgment on that.26 (Tr. 787-788)

26 However, courts use the concept of “statistical significance” in applying Daubert. See
Brock, 874 F.2d at 312 (plaintiff’s epidemiological evidence of causation excluded because it
lacked “statistical significance”); Miller v. Pfizer, Inc., 356 F.3d 1326, 1333-35 (10th Cir. 2004)
(lack of statistically significant results considered in court’s Daubert analysis to determine
whether expert’s opinions were founded on scientific principles).
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
27
Because the experiment found that a greater percentage of all hypothetical voters cast
votes for the Neither choice than in the Santa Rosa election, Dr. Krosnick also concluded that
the estimated difference of 4 percentage-points due to the bolded name is probably an under-
estimate of the impact of the bold/non-bold fonts in the actual election. (Tr. 761) Dr. Krosnick
noted that, even if he had underestimated the impact of the bold font in the Santa Rosa election,
because of the closeness of the election, he was confident that the effect of the bold/non-bold
fonts was sufficiently large to have changed the outcome of the election. (Tr. 761)
Dr. Krosnick explained that the experiment “points to . . . the idea that individuals who
had relatively little information about the campaign had their attention focused both –on the two
bolder options, less on the ‘neither option’ . . . [a]nd that the NUHW being presented first [had]
another advantage between the bold options.” (Tr. 795) Dr. Krosnick went on to state the
following:
Now, I’m not here to testify that the bolding of those two options should have led to equal
numbers of votes being cast for the two bold options and that any quality of those votes
cast somehow challenges the idea that they were more salient than the Neither option, it
would not challenge that. But what this suggests is that for these voters, who had
relatively low information, that attracting their attention to the NUHW may have caused,
in fact probably causes, some of those individuals to vote that way as opposed to the
‘neither’ option.” (Tr. 795)
C. Further Findings and Conclusions
The Hospital, relying largely on the opinion evidence provided by Dr. Krosnick, argues
that the lack of a uniform font for the choices on the ballots used in the election amounted to a
defect that increased the number of votes cast for the NUHW and reduced the number of votes
cast for “Neither”. The Hospital contends that various state election laws requiring the use of
ballots with the choices presented in a uniform style and/or font represents persuasive authority
supporting a conclusion that the ballot used here was flawed and that the election should be
declared void. The Hospital also argues that it had no duty to object in advance concerning the
defective ballot used in this election because the Board rather than the parties is responsible for
the accuracy of election notices and, by inference, the ballots used.27
The NUHW argues that this objection should not be considered substantively because
the Hospital waived any objection to the ballot design by failing to object to it when it received
the Notices of Election well in advance of the election date and posted copies of them
throughout its facilities. The NUHW contends that Dr. Krosnick’s expert opinion should not be
considered because it is premised on irrelevant considerations and is unreliable.
I find merit to the NUHW’s waiver contention. In the case it cites, Liquid Transporters,
Inc., 336 NLRB 420 (2001), the Board precluded an employer from objecting for the first time in

27 In support, the Hospital cites R Casehandling Manual Section 113314 (sic). Er. Br., p. 75.
Section 11314 of the R Casehandling Manual provides in relevant part:
The completion of (the Notice of Election) requires insertions made by use of a template,
which is a function of the support staff, but the basis for such action is the election order
sheet . . . prepared by the assigned Board agent. The responsibility for the accuracy of
the finished notice of election is the Board agent’s.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
28
post-election objections to the union’s use of a Section 2(11) supervisor as an election observer,
a practice long prohibited by the Board. I find that case analogous to the situation here. The
Hospital knew, or should have known, about the ballot design on which this objection is based
when it first received the sample ballot attached to the Notices of Election well in advance of the
election. Not only did it fail to raise any objection at that time, it also failed to address the
problem, or even request an inspection of the full-sized ballot, during the pre-election
conference the Board agents conducted the evening before the first election session when time
still remained to correct the ballot design to the Hospital’s satisfaction.
The Hospital’s contention that it had no duty to call attention to the ballot design (based
on language from the R Casehandling Manual) that it now chooses to call “defective” lacks
merit. The Board, with court approval, has long required parties to provide timely notice of
potential election problems rather than remaining silent until post-election objections are filed.
See e.g. NLRB v. Corrugated Container Corp., 431 F.2d 1196, 1197 (1st Cir. 1970) (employer
and losing union’s silence in advance of election and after the election notices had been posted
for a week as to employees’ problems in understanding the choices offered amounts to either
laches or, more preferably, compelling evidence that the procedure used was sufficient). The
cited R Casehandling Manual provision does not pertain to the parties’ duties and
responsibilities at all; it merely allocates responsibilities between the Board agent and the office
support staff for the preparation of the Notice of Election. Accordingly, I recommend that the
Board overrule Objection 9 on the ground that the parties’ waived objections to the ballot design
by failing to call attention to it in advance of the election when it could have been corrected.
But that reason aside, I recommend that the Board overrule this objection because the
Hospital failed to sustain its heavy burden of proving that the ballot design affected employees
in the voting unit. Safeway, Inc., supra. at p. 525. The claim in the second prong of the
Hospital’s objection that the ballot created “the impression that the NLRB disfavored the
employer and considered a vote for ‘Neither’ was not on equal terms and was of lesser
significance to a vote for either union” should be rejected as a matter of law in view of the
disclaimer language contained on the ballot itself as well as on the Notices of Election. In
response to issues raised in the altered sample ballot cases – a situation far more suggestive
and egregious than the font differences at issue here – the Board adopted the requirement that
the ballot itself contain disclaimer language so that henceforth there would be no confusion as
to “the Board’s neutrality in the election process,” and so that “employees will not reasonably be
misled into believing that the Board supports a particular party . . . or promotes that party’s
cause.” Ryder Memorial, 351 NLRB 214, 216. To conclude that the election notice and ballot
disclaimer language is inadequate to prevent employees from mistakenly concluding that the
Board favored one of the parties in an election based on the ballot design used here but is
adequate to prevent employees from concluding “reasonably” that the Board favored one party
or the other in situations involving an altered sample ballot would be completely irrational.
I also find that the evidence is insufficient to establish the first prong of Objection 9
asserting that “the Board created a ballot that displayed the employees' choices in different
fonts and font styles thereby confusing voters.” No voter in this election claimed at the hearing
to have been confused by this ballot design. Indeed, there is no evidence that any voter even
noticed the problematic ballot design. Instead, the Hospital relies on Dr. Krosnick’s opinion that
the ballot design probably had an impact on the choice made by enough voters to affect the
outcome of this election based on the reactions he observed of a few so-called low information
voters in a hypothetical election.
I have concluded Dr. Krosnick’s opinion concerning the impact of the ballot design
should not be accorded significant weight because the experiment he conducted to support that
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
29
opinion did not relate reliably to the facts of this case as required by FRE Rule 702 and the
Daubert case. For reasons approximating those already noted in connection with the insignia
experiment, the hypothetical election did not adequately replicate the actual election. The low
knowledge, high knowledge divide based on the number of information screens shown on a
computer is an insufficient approximation of the influences present in the typical representation
election where, for example, a variety of personal exchanges and personal biases play such a
significant role. Additionally, the low information voter construct used in the experiment cannot
be correlated to any identifiable group of voters in the actual election. No explanation has been
provided as to whether this group would include four voters or four hundred voters in the actual
election. Nebraska Plastics, Inc. v. Holland Colors Americas, Inc., 408 F.3d 410, 416-17 (8th
Cir. 2005) (trial court properly excluded expert’s calculation of anticipated future damages
because calculation failed to take into account “a plethora” of relevant facts); Conopco, Inc. v.
Cosmair, Inc., 49 F.Supp.2d 242, 254 (S.D.N.Y. 1999) (consumer confusion survey flawed
because it did not present products in conditions that even vaguely replicated marketplace).
Give the variety of shortcomings underlying Dr. Krosnick’s ballot design experiment, I am
unconvinced that a sufficient correlation exists between the empirical data he derived from that
experiment and the actual election. Therefore, I conclude that this evidence is insufficient to
warrant setting aside the election results here.
On these findings of fact and on the entire record, I issue the following recommended28
ORDER
Objections 1, 3, 8 and 9 filed by the Employer, Santa Rosa Memorial Hospital, are
hereby overruled in their entirety, and the National Union of Healthcare Workers is hereby
certified as the exclusive collective bargaining representative under Section 9(a) of the Act in
the following appropriate unit:
Including: All full-time and regular part-time employees employed by the Employer at its California
facilities located at 1165 Montgomery Dr, Santa Rosa; 1287 Fulton Rd, Santa Rosa; 500 Doyle Park
Dr, Santa Rosa; 525 Doyle Park Dr, Santa Rosa; 1450 Medical Center Dr, Rohnert Park in the
following job classifications:
Clerk-Imaging
Clerk-Receiving Distribution
Clinical Lab Asst I-CPT
Clinical Lab Asst II-CPT
Clinical Lab Asst III-CPT
Clinical Lab Asst II-Outreach
Clinical Lab Asst III-Outreach
CNAl Care Partner I
CNAl Care Partner I, II, III
CNAl Care Partner II
Financial Counselor, Admitting Dept.
Inventory Database Controller
Lead Resp. Care Practitioner
Lead Secretary
Lead Secretary Laboratory
Lead-Cook
Lead-Courier
Lead-Env. Services Rep.
Lead-Nutrition Services Aide
LVN I
LVN II
Tech II-Urgent Care
Tech I-Pharmacy
Tech I-Radiology
Tech I-Sterile Processing
Tech I-Urgent Care
Tech Patient Care
Tech Patient Handler
Tech Telemetry
Tech-Anesthesia
Tech-Cardio Radiologic
Tech-Cardio/Pulmonary

28 Within fourteen (14) days from the issuance of this decision and recommended order, any
party may file exceptions to it with the Board in Washington, D.C. If a party files exceptions,
Section 102.69 requires that an original and eight (8) copies be submitted to the Board together
with a supporting brief, if desired. Any party filing exceptions must immediately serve a copy
thereof on all other parties, and the Regional Director. Exceptions, if any, must be received by
the Board in Washington, D.C. on or before June 11, 2010.
JD(SF)–18–10
5
10
15
20
25
30
35
40
45
50
30
CNAl Care Partner III
Cook/Caterer
Coordinator-Bio Medical
Coordinator Cancer Registry
Coordinator-CME
Coordinator-Database
Coordinator-Diet
Coordinator-Engineering
Coordinator-Events & Dev. Asst
Coordinator-Health Info. Mgt
Coordinator Business I,
Orthopedics,
Obstetrics/Gynecology,
Pharmacy
Coordinator Business II
Coordinator-LLC Comm.
Coordinator-Medical Library
Coordinator -Medical Staff Svcs
Coordinator-Pyxis
Coordinator-Quality Assurance
Coordinator-Resource
Coordinator-Surgical Services
Coordinator-Transcription
Coordinator-Utilization Mgmt
Courier
Diet Coordinator
Env Services Rep
Env Services Specialist
ER Patient Access Rep
LVN III
LVN IV
Medical Interpreter
Nutrition Services Aide I
Nutrition Services Aide II
Outpt Registrar/Clinical Assistant
Pathology Lab Assistant I
Pathology Lab Assistant II
PBX Operator I
PBX Operator II
Physical Therapy Assistant I
Physical Therapy Assistant II
Registrar Trauma
Rehab Coordinator I
Rehab Coordinator II
Resp Care Practitioner I
Resp Care Practitioner II
Scheduler
Secretary
Secretary-Operating Room
Secretary-Staffing Services
Spec-Perinatal
Spec-Surgical Database
Tech I, II and III Surgical
Tech III-Radiology
Tech II-Pharmacy
Tech II-Radiology
Tech II-Sterile Processing
Tech-Cardiovascular
Tech-Central Supply
Tech-Cytology
Tech-Echo
Tech-EKG
Tech-EKG/EEG
Tech-Emergency Dept
Tech-Endoscopy
Tech-Equipment
T ech-Laparoscopic
Tech-Medical Laboratory
Tech-MRI
Tech-Nuclear Medicine
Tech-OB
Tech-Ortho Trauma
Tech-Pulmonary
Tech-Rehab
Tech-Surg Service Patient Care
Tech-Surgical Services Equip
Tech-Ultrasound
Transcriber
Transcriber II
Transporter I-Patient
Transporter Supply
Transporter-Patient
Unit Sectry II/Care Partnr III
Ward Clk/Unit Secretary I
Ward Clk/Unit Secretary II
Ward Clk/Unit Secretary III
Excluding: All Physicians, Registered Nurses, Professionals, Skilled Maintenance Employees,
Business Office Clerical Employees, Guards and Supervisors as defined in the Act.
Dated at Washington, D.C. May 28, 2010

WILLIAM L. SCHMIDT
Administrative Law Judge
JD(SF)-18-10
31
UNITED STATES OF AMERICA
BEFORE THE NATIONAL LABOR RELATIONS BOARD
DIVISION OF JUDGES SAN FRANCISCO BRANCH OFFICE
SANTA ROSA MEMORIAL HOSPITAL
Employer
and 20-RC-18241
NATIONAL UNION OF HEALTHCARE WORKERS
Petitioner
and
SEIU-UHW, UNITED HEALTHCARE WORKERS-WEST
Intervenor
TABLE OF CONTENTS
Page
Decision and Recommended Order On Objections to Election………………….. 1
Statement of the Case………………………………………………………………… 1
Findings of Fact………………………………………………………………………... 2
I. Preliminary Findings………………………………………………......…………... 2
II. Objection 1…………………………………………………………………………. 4
A. The Relevant Evidence…………………………………………………………… 4
B. The Opinion Evidence of Dr. Krosnick………………………………………….. 6
C. Further Findings and Conclusions ……………………………………………... 8
III. Objection 3………………………………………………………………………… 11
A. The Relevant Evidence…………………………………………………………… 11
1. The December 2 Incident…………………………………………………………. 11
2. Gabriella Martinez and the December 7 Informed Choice Meeting………….. 14
B. Further Findings and Conclusions………………………………………………. 16
IV. Objection 8………………………………………………………………………… 17
A. The Relevant Evidence…………………………………………………………… 18
B. Further Findings and Conclusions………………………………………………. 21
V. Objection 9…………………………………………………………………………. 23
A. The Relevant Evidence…………………………………………………………… 23
B. The Opinion Evidence of Dr. Krosnick………………………………………….. 24
C. Further Findings and Conclusions……………………………………………… 27
ORDER…………………………………………………………………………………. 29