One Wisconsin Now v. Kremer, Jesse et al

Filing 80

OPINION AND ORDER granting 27 Motion for Summary Judgment; denying 56 Motion for Summary Judgment; granting in part and denying in part 38 Motion to Strike Expert Report; granting 69 Motion for Leave to File Sur-Reply Brief; denying as moot 74 Motion to Strike Jury Demand; granting 77 Motion to Supplement Motion for Summary Judgment. All remaining deadlines in this case are STRUCK, except for the briefing schedule on relief set forth above and defendant Kremer's motion to dismiss previously set by the court. The parties are also to hold open March 11, 2019, for possible argument and presentation of evidence as to the appropriate remedy, if necessary. Signed by District Judge William M. Conley on 1/18/2019. (kwf)

IN THE UNITED STATES DISTRICT COURT
FOR THE WESTERN DISTRICT OF WISCONSIN
ONE WISCONSIN NOW,
Plaintiff,

OPINION AND ORDER

v.
17-cv-0820-wmc
JESSE KREMER, JOHN NYGREN,
ROBIN VOS, in their official capacities,
Defendants.
Plaintiff One Wisconsin Now alleges that defendant Jesse Kramer, John Nygren and
Robin Vos -- all elected members of the Wisconsin State Assembly -- violated the First
Amendment by blocking it from their respective Twitter pages. Before the court are the
parties’ cross-motions for summary judgment. (Dkt. ##27, 56.) For the reasons that
follow, the court will grant plaintiff’s motion and deny defendants’ motion for summary
judgment, leaving only the question of the appropriate scope of relief under the First
Amendment to be decided.

PRELIMINARY ISSUE
Before turning to the parties’ cross-motions for summary judgment, the court first
addresses defendants’ challenge to the testimony pf plaintiff’s expert Kathleen Bartzen
Culver.

(Defs.’ Mot. to Strike (dkt. #38).)

Dr. Culver has a Ph.D. in mass

communications and teaches at the University of Wisconsin’s School of Journalism and
Mass Communication. Defendants do not seek to strike portions of her report that fall
within her area of expertise -- namely, how Twitter works -- but rather, to strike a number
of paragraphs in her report, which they contend constitute inadmissible legal opinions.

Specifically, defendants challenge Culver’s opinions that: (1) political speech is afforded
special deference in the United States; (2) Twitter presents a public forum; (3) defendants’
use of Twitter is not government speech; (4) defendants’ blocking of plaintiff constitutes
state action, implicating First Amendment concerns; and (5) there are lesser restrictive
alternatives. (Id. at 3-4 (citing Culver Rept. (dkt. #26); id. at 4 (listing specific paragraphs
defendants seek to strike).) In response, plaintiff acknowledges that Culver’s testimony
“occasionally and necessarily touch on the law,” but argues that the “opinions are not so
purely comprised of legal conclusions to be inadmissible.” (Pl’s Opp’n (dkt. #52) 1.)
Generally, the court agrees that portions of the testimony defendants challenge
either provide context for her subsequent discussion or use the law as a framework for
discussing specific features of Twitter. For example, the first sentence of paragraph 11
simply summarizes her understanding of the questions presented by the current case and,
as such, provides no opinions, legal or otherwise.

Paragraph 18 also describes her

understanding as to how Twitter activities can constitute political speech, and while relying
on this understanding of the legal framework to couch her description of Twitter activities,
the thrust of her opinions concern how citizens can use Twitter to engage in political
discussions. Finally, paragraph 20 provides her overview of the different types of public
forums, which then guides her analysis of Twitter in paragraphs 23, 24 and 25.
All of this challenged testimony is akin to other experts who, for example, set out
the Georgia-Pacific factors or trademark infringement factors before analyzing each one.
Indeed, the court finds the substance of Culver’s descriptions as to the different parts of
Twitter, which content is controlled by the Twitter account owner versus by others, and

2

how an observer would understand the difference, potentially helpful and within her area
of expertise. (Culver Rept. (dkt. #26) ¶¶ 27-28.)1 Similarly, paragraphs 30, 32-34 describe
the impact of blocking, as compared to other Twitter actions like muting or setting policies
controlling Twitter feed comments. All of this also falls within her area of expertise and is
helpful to the court.2
On the other hand, there are entire paragraphs of the report that simply describe
caselaw in a way that is neither helpful to the court, nor within Culver’s area of expertise.
As such, the court will grant the motion -- and has not considered -- the following
paragraphs: ¶ 10 (describing the Supreme Court’s holding in Packingham v. North Carolina,
137 S. Ct. 1730 (2017)); ¶¶ 16-17 (describing the importance of political speech and
reasons for that protection espoused by Alexander Meiklejohn); and ¶¶ 21-22 (describing
political speech in other contexts and the Supreme Court’s recognition of social media as
a forum for political speech in Packingham).

Buried in these paragraphs, and others, Culver does, at times, offer her opinions on what are
ultimately legal question (e.g., “These public officials’ use of social media channels and the feeds
specifically do not amount to ‘government speech.’”). (Culver Rept. (dkt. #26) ¶ 27.) The court
agrees that this testimony is not helpful, but other parts of the same paragraph are helpful and
within her area of expertise. See Grussgott v. Milwaukee Jewish Day Sch., Inc., 882 F.3d 655, 662 (7th
Cir. 2018), cert. denied, No. 18-125, 2018 WL 3626102 (U.S. Nov. 5, 2018) (“Courts do not
consult legal experts; they are legal experts.”) (citing United States v. Knoll, 785 F.3d 1151, 1156
(7th Cir. 2015); Fed. R. Evid. 702(a)). Since the court is certainly capable of discerning the
difference, there would be little point, and a great deal of wasted time and energy, to attempt to
excise the offending language from paragraphs containing otherwise appropriate relevant
information. Accordingly, the court will simply give no weight to Culver’s legal opinions, rather
than bother to do so.
1

Here, too, Culver extends her testimony to ultimate legal questions, e.g., whether an action
constitutes “state action.” (Culver Rept. (dkt. #26) ¶ 30.) The court opts to simply ignore this
impermissible extension rather than the appropriate areas of testimony in that same paragraph.
2

3

Accordingly, the court will grant in part and deny in part defendants’ motion. For
the purposes of deciding the parties’ cross motions for summary judgment, I have not
considered paragraphs 10, 16, 17, 21 and 22, nor portions of other paragraphs in which
Culver opines on ultimate legal questions or otherwise misstates current law.

UNDISPUTED FACTS3
A. The Parties
Plaintiff One Wisconsin Now (“OWN”) is a 501(c)(4) non-profit corporation based
in Madison, Wisconsin. OWN self-identifies as a liberal organization, engaged in “issue
advocacy and works to promote a vision of a Wisconsin with equal economic opportunity
for all.” (Pls.’ PFOFs (dkt. #29) ¶ 2.)
At the time the parties filed for summary judgment, defendants Jesse Kremer, John
Nygren and Robin Vos were all Wisconsin State Assembly Representatives. Kremer resides
in Kewaskum, Wisconsin, and represents the 59th Assembly District; Nygren resides in
Marinette, Wisconsin, and represents the 89th Assembly District; and Vos resides in
Burlington, Wisconsin, and represents the 63rd Assembly District. All three defendants
maintain offices in Madison, Wisconsin.

B. Twitter
1. Overview
Twitter is an online social media platform where users are able to “publish short

Unless otherwise noted, the court finds the following facts to be material and undisputed for
summary judgment purposes.
3

4

messages, to republish or respond to others’ messages, and to interact with other Twitter
users in relation to those messages.” (Pls.’ PFOFs (dkt. #29) ¶ 7.) The social media site
has over 300 million users worldwide. There are approximately 70 million Twitter users
in the United States.4
A Twitter user is an individual or entity who has created a Twitter account. Each
Twitter user has an account name, which always consists of an “@” symbol and a unique
identifier (e.g., @repvos). The account name is called a Twitter “handle.” Each user also
has her own timeline that displays her tweets, the date she joined Twitter, and “a button
that invites other to ‘Tweet to’ the user.” (Id. at ¶ 10.) The timeline may also include
a short biographical description; a profile picture, such as a
headshot; a “header” image, which appears as a banner at the
top of the webpage; the user’s location; a button labeled
“Message,” which allows two users to correspond privately; and
a small sample of photographs and videos poster to the user’s
timeline, which link to a full gallery.
(Id. ¶ 11.) Provided by plaintiffs, the image below is an example of a Twitter user’s
timeline.

4

These numbers were as of the dates of the parties’ summary judgment submissions.

5

(Id. ¶ 12.)
An individual tweet can be up to 280 characters in length. In addition to the 280
characters, a tweet will contain the user’s account name (which links to the user’s timeline),
the user’s profile picture, the date and time the user generated the tweet, and the number
of times other Twitter users have replied to, retweeted or liked/favorited the tweet. Tweets
can also contain multimedia content such as photographs, videos and links.
2. Interactive Capabilities
All of a user’s own tweets are displayed on the user’s timeline. A user’s most recent
tweets are automatically published on the top of her timeline. A user can also display
content from other Twitter users on her timeline by retweeting or quoting another user. If
a user retweets another user, the other user’s original tweet will appear on the retweeter’s
timeline along with the user’s own tweets. The image below illustrates this concept:

6

(Id. at ¶ 9.) Users are also able to quote other users in their own tweets.
Twitter users can directly engage with other users in several ways. The most basic
form of interaction is “following,” which allows Twitter users to subscribe to other users’
messages. Users that follow another user’s Twitter account (called “followers”) are notified
when the user they follow publishes tweets on her timeline. Twitter users also can engage
in more direct user interaction by replying to another user’s tweet. Like a tweet, a reply
can contain up to 280 characters and can include photographs, videos, and links. For
instance, if User A tweets “I support candidate X,” User B can reply directly to User A’s
tweet and say, “I disagree with you, I support candidate Y.” When User B replies to User
A’s tweet, the reply will appear in two places: (1) User B’s timeline under a tab called
“Tweets & replies” and (2) on User A’s timeline in a “comment thread” under the original
tweet. Once a reply appears in a comment thread, any other Twitter user with access to
User A’s timeline can reply to either User A’s tweet or User B’s reply to the tweet. “Other
users’ replies to the same tweet will appear in the same comment thread, nested below the
replies to which they respond.” (Id. at ¶ 17.) Another way Twitter users can engage with

7

each other is by “liking,” also referred to as “favoriting,” another user’s tweet. (See, e.g.,
Pl.’s Reply to Pl.’s PFOFs (dkt. #49) ¶ 19.) Finally, a user can mention another user in a
tweet by including the other user’s Twitter handle in a tweet.
The original content of these interactive mechanisms (tweets, retweets, replies, likes
and mentions) are controlled by the user who generates them. No other users “can alter
the content of any retweet or reply,” and other users “cannot prescreen tweets, replies,
likes, or mentions that reference their tweets or accounts.” (Id. at ¶ 20.)5
3. Accessibility
Generally, Twitter users’ timelines are visible not only to other Twitter users, but
also everyone with internet access, including non-Twitter users. While non-Twitter users
can see a user’s account, “they cannot interact with users on the Twitter platform and may
not see tweets in real time.” (Id. at ¶ 14.) A Twitter user can also elect to “block” another
user’s access to her timeline. When a Twitter user is blocked, she is no longer able to
see, or reply to the blocking user’s tweets, retweet the blocking
user’s tweets, view the blocking user’s list of followers of
followed accounts, or use the Twitter platform to search for the

In their reply brief, defendants represented that they may delete others’ tweets in addition to
blocking users. (Defs.’ Reply (dkt. #66) 15-16.) In response, plaintiff filed a motion to file a surreply addressing this argument and the proposed brief. (Dkt. ##69, 69-1.) In responding to that
motion, defendants stated that they do not oppose the sur-reply and acknowledged that the “cited
evidence does not support” the statement that “a user may delete a follower’s reply tweet.” (Defs.’
Resp. (dkt. #70).) As such, the court will grant the motion to file a sur-reply and has considered
the proposed brief and defendants’ response. More recently, defendants filed their own motion to
supplement the summary judgment record further to add plaintiff’s acknowledgement in response
to interrogatories that “[w]hen a Twitter user deletes one of his own tweets, to which a follower has
replied, from his feed, that user’s tweet and the follower’s reply are removed from view on that
user’s feed.” (Defs.’ Mot. to Suppl. (dkt. #77) 2.) The court will grant this motion as well, and it
has considered its import as reflected in the opinion below. See infra Opinion p.29 n.14s.
5

8

blocking user’s tweets.
(Id. at ¶ 21.)
While users are not notified when they are blocked, a user can see whether she is
blocked by visiting the blocking user’s Twitter timeline. If blocked, the user will then “see
a message indicating that the other user has blocked him or her from following the account
and viewing the tweets associated with the account.” (Id.) The following is an example of
such a message:

(Id.)
Twitter also allows users to “mute” other users. A muted user is still able to “follow
the muting user, retweet the muting user’s tweets, and participate in comment threads
created by the muting user.” (Id. at ¶ 23.) However, the muting user is unable to see any
of the muted user’s Twitter activity that would otherwise appear on that user’s timeline.

9

C. The Parties’ Twitter Accounts
1. Plaintiff OWN’s Account
OWN created its Twitter account in September 2008. The account is operated by
OWN’s staff under the username @onewisconsinnow. OWN’s Twitter account has nearly
11,500 followers and follows over 2,200 other accounts, including the Twitter accounts of
both “liberal and conservative politicians.” (Id. at ¶ 39.)
OWN uses Twitter for several purposes. First, OWN’s Twitter account allows it to
“engage in speech, monitor Wisconsin’s elected officials, and deliver news and analysis to
the public.” (Id. at ¶ 37.) OWN publishes several Tweets a day “almost exclusively on
matters of state and federal politics and policy from a liberal perspective.” (Id. at ¶ 38.)
Second, OWN uses Twitter to “receive news of, reply to, and ask questions” about the
activities of liberal and conservative politicians. (Id. at ¶ 39.) Specific to this lawsuit,
before being blocked from the defendants’ Twitter accounts, OWN would monitor,
comment on, and reply to defendants’ tweets, retweet defendants’ tweets, and mention
defendants’ accounts in its own tweets.
2. Defendant Kremer’s Account
Kremer “operates and/or oversees the operation” of the Twitter account
@RepJesseeKremer (“Kremer Account”). (Id. at ¶ 4.) The Kremer Account was created in
January 2015. The Kremer Account has over 1,000 followers. The Kremer Account is
registered to “Rep. Jesse Kremer, Official account for Wisconsin State Representative Jesse
Kremer (R-Kewaskum).” (Id. at ¶ 28.) The account’s page links to Kremer’s “official
legislative website . . . established by the Wisconsin Legislative Technology Services

10

Bureau.” (Id.)
Kremer possesses and uses the login information for the Kremer Account. Kremer
also controls the content of tweets and decides who to block. Members of Kremer’s staff,
in particular a research assistant, post content on the Kremer Account. Kremer’s staff do
so “from their offices on state time.” (Id. at ¶ 66.)
Kremer uses Twitter to “notify the public about his public listening sessions,” where
members of the public can listen to Kremer “talk about things that are going on in
Madison” and share their ideas with Kremer. (Id. at ¶ 81.) Kremer also uses Twitter “to
notify the public about topics such as legislation, upcoming legislative hearings, and
government reports.” (Id. at ¶ 82.) More generally, Kremer told a staff member that the
Kremer Account “is a forum for constituents.”6 (Id. at ¶ 135.) In contrast, he has stated
that the Kremer Account “is not for Dane County liberals to carry on conversations with
me.”7 (Id. at ¶ 128.)
3. Defendant Nygren’s Account
Nygren “operates and/or oversees the operation of” the Twitter account @rep89
(“Nygren Account”). (Id. at ¶ 5.) The Nygren Account was created in May 2009. The
Nygren Account is registered to “John Nygren, ‘Wisconsin State Representative, Joint
Finance Co-Chair, Proudly Serving Marinette, Oconto, and Brown Counties.’” (Id. at ¶

6

Kremer made this comment after OWN filed its lawsuit. (Id. at ¶ 135.)

By “Dane County liberals,” Kremer explained he is referencing “‘the mind set for a region,’ because
‘Dane County is obviously a very liberal, socialist area, progressive area.’” (Id. at ¶ 129.)
7

11

30.)

As with Kremer’s, the “@rep89” refers to Nygren’s representation of the 89th

Wisconsin State Assembly District. Nygren’s account’s page links to Nygren’s “official
legislative website . . . established by the Wisconsin Legislative Technology Services
Bureau.” (Id.) The Nygren Account has over 2,900 followers.
The Nygren Account is primarily operated by Rep. Nygren, although members of
Nygren’s legislative office staff also play a role in its operation.8 Moreover, the staff
operates the Nygren Account “on state time and with state computers.” (Id. at ¶ 73.)
Nygren created this Twitter Account to “communicate with his legislative
constituents.” (Id. at ¶ 85.) The Nygren Account tweets about policies, but also about
topics of less policy-orientated nature “in part because as a state representative he wants
his constituents to know him more fully.” (Id. at ¶ 88.) Nygren also uses Twitter to notify
the public about town hall meetings.
4. Defendant Vos’s Account
Finally, Vos “operates and/or oversees the operation of” the Twitter account
@repvos (“Vos Account”). (Id. at ¶ 6.) The Vos Account was created in March 2009 and
now has over 4,000 followers. The Vos Account is “publicly registered to Robin Vos, ‘State
Representative representing parts of Racine County.’” (Id. at ¶ 31.)
Vos created this account as “‘a minority member of the Joint Committee on Finance
looking for additional ways to get my views out to the public’ and want[ing] ‘the

Defendants object to this fact because plaintiff relies in part on testimony from an individual who
no longer works on Nygren’s staff. (Id. at ¶ 71.) Absent contrary evidence, however, this witness’s
personal knowledge still forms an adequate basis for the finding of fact.
8

12

opportunity to talk directly to people.’” (Id. at ¶ 91.) The parties disagree over the Vos
Account’s current use and who operates it, but there is no dispute that Vos “created the
Vos Account with the intention that anyone could follow it, whether or not they were a
constituent of [Vos’s] Assembly District,” and that he “continues to use the Vos Account
for both tweets about policy and personal tweets.” (Id. at ¶¶ 95, 190.) Unlike the Kremer
and Nygen Accounts, however, the Vos Account does not contain a link to Vos’s legislator
homepage, a state-sponsored website.
Vos also has another account with the handle “@SpeakerVos,” which he created
after assuming the role of Speaker of the Wisconsin State Assembly.

That account

“functions in Vos’s ‘role as Speaker . . . speaking on behalf of the Assembly and Republicans
in general.’” (Id. at ¶ 93.) Vos’s legislator homepage contains a link to the @SpeakerVos
Twitter Account, which did not block One Wisconsin Now on Twitter.

D. Defendants Block OWN’s Access to Their Twitter Accounts
OWN first followed the Vos Account in 2011, the Nygren Account in September
2013, and the Kremer Account in April 2015. (Id. at ¶¶ 44, 50, 56.) In 2017, all three
defendants blocked the OWN Account from their respective Twitter Accounts as discussed
above.9 While defendants have offered various reasons for blocking the OWN Account,
there is no dispute that each defendant personally made the decision to do so. (Id. at ¶¶
126, 162, 177.) Kremer claims he blocked the OWN Account “to stop spamming, stop

In June 2017, OWN noticed that its Twitter account was blocked by the Kremer Account. (Id. at
¶ 124.) Nygren blocked the OWN Account on or about August 9, 2017. (Id. at ¶ 126.) Vos blocked
the OWN Account sometime before March 16, 2017. (Id. at ¶ 127.)
9

13

the posting of tweets unrelated to the topic of the original tweets he posted, and to stop
tweets of an inappropriate and unprofessional nature.” (Id. at ¶¶ 180.) Nygren “claims to
have blocked OWN . . . for ‘[c]rude comments on Wisconsin politics’” that included
“numerous vulgarities,” such as “personal attacks, not necessarily against [Nygren] but
other figures.” (Id. at ¶¶ 152, 154.) Finally, while Vos does not remember why he blocked
the OWN Account (id. at ¶ 162) his staff noted that the only appropriate grounds to block
another Twitter user was “the use of vulgarity and profanity.” (Id. at ¶ 176.)
All three defendants have blocked other Twitter accounts besides the OWN
Account from their respective Twitter Accounts, but none have adopted official policies on
blocking other Twitter users. (Id. at ¶¶ 109, 158, 161.) Not including the OWN Account,
Kremer has blocked five Twitter accounts; Nygren has blocked 37 Twitter accounts; and
Vos has blocked three Twitter accounts. Id.

OPINION
The court may grant summary judgment if the moving party “shows that there is
no genuine dispute as to any material fact and the movant is entitled to judgment as a
matter of law.” Fed. R. Civ. P. 56(a). In deciding whether to grant summary judgment,
the court views all facts and draws all inferences in the light most favorable to the
nonmoving party. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986). A factual
dispute can preclude summary judgment, but only if the “facts might affect the outcome
of the suit under the governing law.” Id. at 248. The party moving for summary judgment
carries the initial burden of showing that there is no genuine issue of material fact that that
it is entitled to relief. Celotex Corp. v. Catrett, 477 U.S. 317, 323 (1986).

14

Plaintiff One Wisconsin Now claims defendants engaged in unconstitutional
content or viewpoint discrimination when they blocked its Twitter account, thereby
removing plaintiff’s access to the interactive portion of defendants’ respective Twitter
accounts. Plaintiff’s claim presents a novel question of law for this court: is the interactive
portion of a government official’s Twitter account a designated public forum?
While there is no settled law on whether a government actor’s social media account
is a designated public forum, two federal district courts have now held that government
officials’ social media accounts can constitute public forums. See Knight First Amendment
Inst. at Columbia Univ. v. Trump, 302 F. Supp. 3d 541 (S.D.N.Y. 2018); Davison v. Loudon
County Bd. of Supervisors, 267 F. Supp. 3d 702 (E.D. Va. 2017). More recently, the Fourth
Circuit became the first federal court of appeals to review this question in upholding the
Eastern District of Virginia’s Davison decision, similarly finding that the interactive portion
of a local government official’s Facebook page constituted a public forum. Davison v.
Randall, No. 17-2002, 2019 WL 114012 (4th Cir. Jan. 7, 2019). Finally, the Supreme
Court has also acknowledged that social media is a “vast democratic forum” analogous to
traditional public forums, such as parks. See Packingham v. North Carolina, 137 S. Ct. 1730,
1735–36 (2017).
Relying on the three-step analysis set forth in these decisions, the court similarly
finds here that: (1) defendants acted under color of state law in creating and maintaining
their respective Twitter accounts in their capacity as members of the Wisconsin State
Assembly; (2) the interactive portion of defendants’ Twitter accounts are designated public
forums; and (3) defendants engaged in content-based discrimination when they blocked

15

the plaintiff’s Twitter account. As a result, plaintiff is entitled to a grant of summary
judgment on its liability claims.

I. State Actor Requirement
As an initial matter, to state a claim under § 1983, a plaintiff must establish that
the defendant acted under color of state law. Cruz v. Safford, 579 F.3d 840, 843 (7th Cir.
2009). In general, a public employee “acts under color of state law while acting in his
official capacity or while exercising his responsibilities pursuant to state law.” West v.
Atkins, 487 U.S. 42, 50 (1988). A court may also find action by a private actor to be under
color of state law if there is a “sufficient nexus between the state and the private actor” to
show that “the deprivation committed by the private actor is ‘fairly attributable to the
state.’” L.P. v. Marian Catholic High Sch., 852 F.3d 690, 696 (7th Cir. 2017) (quoting Lugar
v. Edmondson Oil Co., 457 U.S. 922, 937 (1982)). The sufficient nexus is determined based
on the totality of the circumstances. Skinner v. Ry. Labor Executives’ Assoc., 489 U.S. 602,
614–15 (1989).
Defendants contend that the act of blocking plaintiff on Twitter was not state action
because there is no Wisconsin law that specifically grants them the power to block Twitter
users. (Defs.’ Opp’n (dkt. #47) 4.) Defendants’ interpretation of the color of state law
test is unworkable, narrow, and, simply put, silly. The law does not and has never required
that a state action be specifically authorized by statute before being subject to examination.
If it did, the force of the protection promised by the First Amendment and the Bill of
Rights would be gutted. Instead, for individuals acting in a public capacity, the color of
state law test examines whether the defendant (1) explicitly or implicitly invoked state

16

authority or (2) “could not have acted in violation of the plaintiff's constitutional rights
but for his state authority.” Luce v. Town of Campbell, 113 F. Supp. 3d 1002, 1016 (W.D.
Wis. 2015), rev’d in part on other grounds, 872 F.3d 512 (7th Cir. 2017). Furthermore, to
find whether a private actor is acting under color of state law, the test is whether there is a
“sufficient nexus between the state and the private actor” to show that “the deprivation
committed by the private actor is ‘fairly attributable to the state.’” Marian Catholic High
Sch., 852 F.3d at 696.
Plaintiff argues that defendants acted under color of state law when they created
and operated their Twitter accounts. To support its argument, plaintiff cites both the
Davison and Knight district court decisions. Although Knight held that the operation of the
defendant government actor’s social media account, including the power to block other
users, constituted state action, it did not engage in a separate color of state law analysis.
Id. at 567–68. Thus, for purposes of the color of state law analysis, the court will focus on
the analysis in Davison.
Applying totality of the circumstances test approved by the Supreme Court in
Skinner, the Davison court found that a government official acted under color of state law
when she created and operated a Facebook page. Id. at 711–14. While the Davison court
noted that several aspects of the defendant’s social media account seemed “entirely
private,”10 these factors were substantially outweighed by factors suggesting the creation

Specifically, the court acknowledged that: “Defendant's enumerated duties do not include the
maintenance of a social media website. The website in question will not revert to the County when
Defendant leaves office. Moreover, Defendant does not use county-issued electronic devices to post
to the ‘Chair Phyllis J. Randall’ Facebook page, and much of Defendant’s social media activity takes
place outside of both her office and normal working hours.” Davison, 267 F. Supp. 3d at 712.
10

17

and operation of the social media page constituted state action. Id. at 712. Specifically,
the Davison court found the following dispositive: (1) the social media page’s obvious
public, not private, purpose (defendant’s election to public office and subsequent use as “a
tool of governance”); (2) defendant’s use of government resources, including government
employees, to maintain the page; (3) the connection between defendant’s official
newsletters and the page; and (4) defendant’s efforts to “swathe” the page “in the trappings
of her office.” Id. at 712–14.
As to the final factor in particular, Davison identified at least eight ways in which
the social media page was bound up with the trappings of her office:
(1) the title of the page includes Defendant’s title; (2) the page
is categorized as that of a government official; (3) the page lists
as contact information Defendant’s official County email
address and the telephone number of Defendant’s County
office; (4) the page includes the web address of Defendant’s
official County website; (5) many—perhaps most—of the posts
are expressly addressed to ‘Loudoun,’ Defendant’s
constituents; (6) Defendant has submitted posts on behalf of
the Loudoun County Board of Supervisors as a whole; (7)
Defendant has asked her constituents to use the page as a
channel for ‘back and forth constituent conversations’; and (8)
the content posted has a strong tendency toward matters
related to Defendant’s office.
Id. at 714.
In affirming the Eastern District of Virginia’s decision in Davison, the Fourth Circuit
also identified a number of factors to be considered when determining whether purportedly
private action bears a sufficiently close nexus with the State to constitute action under
color of state law. No. 17-2002, 2019 WL 114012, at *7. These factors include: (1)
whether defendant’s actions are linked to events arising out of his or her official status; (2)

18

whether defendant’s “status” as a public official enabled defendant to act in a manner that
a private citizen could not have done; (3) whether the defendant’s action occurred in the
course of performing an actual or apparent duty of his or her office; and (4) whether
defendant used the power and prestige of his or her state office to damage the plaintiff. Id.
After analyzing the totality of the circumstances, the Fourth Circuit agreed with the district
court that on balance the defendant had acted under color of state law. Id.
As in Knight and Davison, there are a few facts in this record that might support
finding the defendants’ creation and operation of their Twitter accounts was private, rather
than public, action: (1) creating the Twitter accounts was not one of the defendants’
enumerated duties; (2) the Twitter page will not become state property when the
defendants leave office; and (3) some of the defendants’ social media activity likely takes
place outside of normal working hours. However, as in those cases, there are far more
factors that support finding defendants created and operated their respective Twitter
accounts under the color of state law under the totality of the circumstances test, therefore,
plaintiff prevails in establishing state action. To illustrate this, the court will set forth the
record evidence as to Kremer’s and Nygren’s Twitter accounts together and then turn to
Vos’s account given its arguably unique facts.

A. Kremer and Nygren’s Twitter Accounts
There is little question that Kremer and Nygren’s accounts rely on “the power and
prestige of . . . state office” and were “created and administered . . . to perform actual or
apparent duties of [Kremer and Nygren’s] office[s].” Davison, 2019 WL 114012 at *7. As
plaintiff notes, Kremer and Nygren’s Twitter accounts are intertwined with their public

19

responsibilities: Kremer and Nygren created their accounts while in office; both defendants
use their accounts to share information on legislative matters, policies, political
appearances, and events; and both use government resources to maintain their Twitter
accounts.
Also weighing in favor of finding state action is the fact that Kremer’s and Nygren’s
accounts are “swathed in the trappings of their office.” Davidson, 267 F. Supp. 3d at 714.
Defendant Kremer’s account, is registered to “Rep. Jesse Kremer,” described as the “Official
account for Wisconsin State Representative Jesse Kremer (R-Kewaskum),” and identified
by the handle “@RepJesseeKremer.” Similarly, Kremer’s account features a picture of the
Wisconsin state capitol accompanied by “images of the word ‘State Representative’ and of
the American flag.” Unlike his separately maintained personal account (“@JesseforWI”),
Kremer’s Twitter account also includes a link to his official legislative website. As for
Nygren’s account, identified by the handle “rep89,” it describes him as a “Wisconsin State
Representative, Joint Finance Co-Chair, Proudly Serving Marinette, Oconto, and Brown
Counties.” Nygren’s account also features a picture of several people, including Nygren,
standing in front of a sign for “Wisconsin State Representative John Nygren.” Finally, like
Kremer’s account, Nygren’s account also links to his official legislative website.

B. Vos’s Twitter Account
In fairness, the facts of record on summary judgment as to defendant Vos’s Twitter
account do not as cslearly indicate that it was created and operated under color of state
law. Two, additional issues complicate the analysis for the Vos account: (1) whether it
utilizes government resources in its operation and (2) the existence of Vos’s

20

“@SpeakerVos” account.
Beginning with the existence of the “@SpeakerVos” account, Vos created it after he
was elected Speaker of the Wisconsin State Assembly and the description of the
“@SpeakerVos” account suggests it exists for Vos to speak on behalf of Republican
representatives in his capacity as Assembly Speaker, not solely as a state representative,
particularly since Vos still uses the “@repvos” account to communicate with the public as
a state representative. Furthermore, while the parties disagree on the extent of Vos’s use
of state resources to maintain his original representative account, it is still sufficiently
linked to public circumstances and swathed in the trappings of his office to constitute state
action under a totality of the circumstances test.
First, Vos created and continues to use his “@repvos” for public purposes.
Specifically, it is undisputed that Vos created the account “because he was ‘a minority
member of the Joint Committee on Finance looking for additional ways to get my views
out to the public’ and wanted ‘the opportunity to talk directly to people.’” (Pl.’s Opening
Br. (dkt. #28) 10.) And he “continues to use the Vos Account for both tweets about policy
and personal tweets” to constituents. (Id. at 11.)
Second, the “@repvos” account is heavily “swathed” in the trappings of Vos’s office.
The account is registered to “Robin Vos, ‘State Representative representing parts of Racine
County.’” (Id. at 10.) The account features an image of uniformed individuals and an
American flag. These two facts alone show that the account still represents Vos in his
official, not personal, capacity.
Finally, with or without the existence of Vos’s speaker account, the essential purpose

21

and function of Vos’s account remains essentially the same as Kremer’s and Nygren’s
Twitter Accounts: to perform actual and apparent duties as state assemblyperson using
the power and prestige of that office to communicate legislative matters and other issues
with the public. Therefore, the court find that Vos acted under color of state law when he
created the account and continues to act under color of state law when operating the
account.

II. Designated Public Forums
The court must next decide whether the interactive components of defendants’
Twitter accounts are designated public forums.11 Under the First Amendment, members
of the public retain varying degrees of free speech rights on government property. Pleasant
Grove City, Utah v. Summum, 555 U.S. 460, 469 (2009). The traditional illustration of
speech in these public forums is assembly or communication on public streets or in public
parks. Id. As noted, however, the Supreme Court has extended public speech protection
to less traditional, designated public forums, defined as “locations or channels of
communication that the government opens up for use by the public for expressive activity.”
Surity v. Hyde, 665 F.3d 860, 869 (7th Cir. 2011).
A government cannot create a designated public forum “by inaction or by permitting
limited discourse.” Walker v. Texas Div., Sons of Confederate Veterans, Inc., 135 S. Ct. 2239,
2250 (2015). Rather, the government must intentionally open “a nontraditional forum

Plaintiff narrows its challenge to only the interactive components of Twitter, which makes sense
in light of the district court’s extensive discussion in Knight First Amendment Inst., 302 F. Supp. 3d
at 573.
11

22

for public discourse.” Id. (citing Cornelius v. NAACP Legal Defense & Educ. Fund, Inc., 473
U.S. 778, 802 (1985)). To determine whether a government has intentionally created a
designated forum, courts examine (1) the “policy and practice of the government” and (2)
“the nature of the property and its compatibility with expressive activity.” Id. Given the
nature of Twitter and its compatibility with expressive activity, now highlighted by the
Supreme Court’s decision in Packingham v. North Carolina, 137 S. Ct. 1730 (2017), the
interactive portions of defendants’ respective Twitter accounts plainly constitute
designated public forums. Id. at 1736; see also Knight First Amendment Inst., 302 F. Supp.
3d at 575 (finding that “[t]he interactivity of Twitter is one of its defining characteristics”
and similarly concluding that the interactive portion constitutes a designated public
forum); Davison, 2019 WL 114012, *8-9 (concluding that the interactive component of a
government official’s Facebook page constituted a public forum because it was
“intentionally opened . . . for public discourse” and was “compatible with expressive
activity”).
Indeed, in Packingham, the United States Supreme Court suggested the internet
generally, and particularly social media, is a new space for public discourse analogous to
traditional public forums:
A fundamental principle of the First Amendment is that all
persons have access to places where they can speak and listen,
and then, after reflection, speak and listen more . . . . A basic
rule, for example, is that street or park is a quintessential forum
for the exercise of First Amendment rights . . . . While in the
past there may have been some difficulty in identifying the
most important places (in a spatial sense) for the exchange of
views, today the answer is clear. It is cyberspace—the ‘vast
democratic forums of the Internet’ in general . . . and social
media in particular.

23

137 S. Ct. at 1735–36. In particular, the Supreme Court observed that when one creates
a social media account, one subscribes to the fundamental purpose of social media: to
create “a digital space for the exchange of ideas and information.” Davidson, 267 F. Supp.
3d at 716. Here, there can be no reasonable dispute that defendants intentionally created
the interactive social media accounts at issue in order to communicate with members of
the public about news and information related to their roles as public officials, and are
continuing to operate them as such.
Nonetheless, defendants argue at summary judgment that by creating a Twitter
account, they did not intentionally create designated public forums because there was no
“policy” in place to do so. Rather, defendants suggest, the Twitter accounts were intended
only as a way for defendants “to get their messages out to their constituents.” (Defs.’
Opp’n (dkt. #47) 16.) Even if this is the only reason the defendants created their Twitter
account, they have not taken any steps to limit access to their accounts to their
constituents, nor have they limited access by the general public, save perhaps for a few
individuals. Regardless, Twitter remains a social media platform, and interaction remains
a key component of this platform. If defendants truly had no intention to create a space
for public interaction and discourse, they would not have created public Twitter accounts
in the first place. Instead, they could simply broadcast their views, schedules and other
information, for example, through a non-interactive blog. Having opted to create a Twitter
account, however, and benefit from its broad, public reach, defendants cannot now divorce
themselves from its First Amendment implications and responsibilities as state actors.
Defendants offer two other arguments against finding a designated public forum:

24

(1) because Twitter is a private corporation, the accounts are not public property, and thus
cannot be a public forum; and (2) defendants’ tweets constitute government speech.
Neither argument holds much merit, for the same reasons recently articulated the Fourth
Circuit. First, the private property claim fails because, as plaintiffs rightly note, public
forums can be found on public or private property. See, e.g., Cornelius, Cornelius v. NAACP
Legal Defense & Educ. Fund, Inc., 473 U.S. 778, 801 (1985) (providing a “speaker must seek
access to public property or to private property dedicated to public use to evoke First
Amendment concerns.”); Davison, 2019 WL 114012, at *9 (“[T]he Supreme Court never
has circumscribed forum analysis solely to government-owned property.”).
Second, as for the government speech claim, even plaintiff acknowledges that the
defendants’ own tweets may constitute government speech, but this does not mean that all
speech associated with their Twitter accounts is government speech. Defendants attempt
to side-step this issue by simply stating that the “more proper conclusion is to treat a state
representative’s entire Twitter page as a whole—and consider it government speech.”
(Defs.’ Opp’n (dkt. #47) 20.) However, defendants fail to provide any reason for requiring
such limited analysis. Moreover, as detailed in plaintiff’s expert’s report, it is easy to
distinguish the parts of Twitter that reflect the defendants’ respective views (e.g., their own
tweets and reactions to other tweets, from the actions of other citizens on their feeds).
(Culver Rept. (dkt. #26) ¶¶ 27-28.) Accordingly, the court agrees with the district court
in Knight and the Fourth Circuit in Davison finding that the interactive portion is severable
from the rest of the Twitter account and not subject to the government speech exception.
See Davison, 2019 WL 114012, at *11 (rejecting the government speech argument because

25

the “interactive component of the Chair’s Facebook Page—the portion of the middle
column in which the public can post comments, reply to posts, and ‘like’ comments and
posts—is materially different” than the portion of the page containing plaintiff’s own
posts); Knight First Amendment Inst., 302 F. Supp. 3d at 572 (holding the content of
Trump’s tweets is not amenable to forum analysis because it is government speech, but
finding “[t]he same cannot be said . . . of the interactive space for replies and retweets
created by each tweet sent by the @realDonaldTrump account”).12
For all of these reasons, the court concludes that the interactive portions of the
defendants’ respective Twitter accounts constitute designated public forums.

III. Content-Based Discrimination
Finally, when the government creates a designated public forum, there are limited
circumstances in which it can restrict speech. Perry Educ. Assn., 460 U.S. at 45–46 (noting
that designated public forums are subject to the same standards as traditional public
forums). Specifically, the government can impose “reasonable time, place and manner
regulations” on speech within the forum. Id. at 46. What the government cannot do is
exclude speech based on its content, unless it can meet that the extraordinarily high burden

12

As described above, defendants filed a motion to supplement the record this week, seeking to add
an undisputed fact that a Twitter user can delete his or her own tweet, which in turn would delete
the replies by other Twitter users. (Dkt. #77.) The court infers that this proposed supplemental
fact is material to defendants’ government speech argument. The court is unpersuaded, however,
that the availability of this delete function could give rise to a reasonable inference that a Twitter
user endorses the position of replies to his or her tweets. See Price v. City of New York, No. 15 CIV.
5871 (KPF), 2018 WL 3117507, at *14 (S.D.N.Y. June 25, 2018) (“The City Defendants cannot
credibly suggest that the public would confuse Plaintiff’s posts criticizing the City as being the
City’s own speech.”).

26

of proving exclusion is “necessary to serve a compelling state interest and that it is narrowly
drawn to achieve that interests.” Id. at 45. Said more succinctly, the government’s reason
for content-based restrictions must satisfy strict scrutiny. Id.
Here, plaintiff argues that by blocking its Twitter account, defendants
impermissibly restricted its speech because of its viewpoint and content. Because there is
a question as to what specific action motivated each defendant to block OWN from their
Twitter accounts, the court opts to consider plaintiff’s content-based challenge. In Surita
v. Hyde, 665 F.3d 860 (7th Cir. 2011), the Seventh Circuit found an impermissible
content-based restriction where a mayor blocked an individual from speaking about an
issue at a city council meeting because of something the speaker said two days before the
meeting.

Id. at 872. In particular, the Seventh Circuit found the mayor’s action was

content-based, rather than content-neutral, because he “excluded a speaker within the class
to which the designated public forum was available.” Id. at 870. In addition, the court
found the mayor’s actions were not a reasonable time, place, or manner restriction because
(1) the plaintiff “approached the microphone at the appropriate time,” and (2) there was
no indication the plaintiff’s speech would be inappropriate. Id. at 871. Finally, the court
concluded the mayor’s reason for employing the content-based restrict (the plaintiff’s prior
speech) did not satisfy strict scrutiny and was thus an impermissible content-based
restriction. Id. at 872.
Like the plaintiff in Surita, defendants in this case blocked plaintiff because of its
prior speech or identity. Indeed, all three defendants indicated, either directly or indirectly,
that they do not approve of plaintiff’s liberal perspective. For instance, Kremer testified

27

that his Twitter feed is not “for Dane County liberals to carry on conversations with me,”
and while Nygren claims he blocked plaintiff for its “crude comments on Wisconsin
politics,” he was unable to identify any, actual “crude” comments. (Pl.’s Opening Br. (dkt.
#28) 32; Pl.’s Reply to Pl.’s PFOFs (dkt. #49) ¶ 155 (“Nygren has no record or recollection
of any particular offensive comments made by OWN.”).) For his part, Vos offers no
speech-related reason for blocking OWN; instead, he simply avers that he cannot recall the
reason. Given the context, however, the only reasonable inference is that Vos blocked
OWN because of its prior activity on his Twitter page. Furthermore, all three defendants
have only blocked a select number of Twitter accounts, supporting plaintiff’s assertion that
defendants selectively blocked plaintiff for a specific reason and not as a matter of
happenstance.
Even if the reason for blocking OWN were not obviously based on content,
defendants’ bear the burden to come forward with a compelling state interest for blocking
plaintiff, Perry Educ. Assn., 460 U.S. at 45–46, and that they have wholly failed to do. In
fact, as plaintiff notes, defendants have articulated no reason for blocking plaintiff at all,
except by vague innuendo or by claiming that they must have had a valid reason but can
no longer remember why they did so. Indeed, in their opposition brief, defendants simply
declined to respond to plaintiff’s content-based challenge, arguing that the government
speech doctrine applies to the Twitter accounts as a whole. (Defs.’ Opp’n (dkt. #47) 33.)13
For these reasons, the court concludes that defendants’ content-based restrictions are

Even if defendants had articulated a reason for blocking OWN from their respective Twitter
accounts, defendants also failed to meet their burden by showing why blocking -- as opposed to
muting, for example -- is the least restrictive means to achieve their undefined goal.
13

28

impermissible and in violation of plaintiff’s First Amendment rights.

IV. Relief
Having found that defendants violated plaintiff’s First Amendment rights in
blocking it from defendants’ respective Twitter accounts, the only remaining question is
the appropriate relief. In its complaint, plaintiff seeks “a permanent injunction requiring
Defendants to unblock One Wisconsin Now from Defendants’ Accounts, and prohibiting
Defendants from blocking One Wisconsin Now or others from the Accounts.” (Compl.
(dkt. #1) p.18.) This request, for the most part, appears reasonable, though the court
would lack authority to prohibit defendants from blocking others from their accounts for
legitimate reasons that do not implicate the First Amendment.
Accordingly, the court will direct the parties to submit short briefs on the remaining
issue of relief. Plaintiff’s brief is due January 30, 2019, with defendants’ response due
February 8, 2019. In addition, the briefing deadlines remain with respect to defendant
Kremer’s recent motion to dismiss because he is no longer a member of the Wisconsin
State Assembly.14 Finally, the parties should hold open the first day of trial, March 11,
2019, should the court decide that any argument or testimony is necessary to decide on an

In plaintiff’s opposition to defendants’ motion for extension to respond to plaintiff’s motion for
summary judgment, OWN represented that Kremer had announced his intention to not seek reelection to his position as a state representative in November 2018 (Pl.’s Opp’n (dkt. #35) 1); and
it appears from local news reports that another individual was elected to represent the 59th
Assembly District. See Sara Razner, Fond du Lac, Dodge County Results, FDL Reporter, Nov. 6, 2019,
https://www.fdlreporter.com/story/news/2018/11/06/wisconsin-elections-rep-jeremy-thiesfeldtwins-52nd-assembly-seat/1909148002/ (noting Timothy Ramthun won the election). As such, any
requested relief against Kremer would now appear to be moot. The court notes that defendants
have recently filed a motion to dismiss plaintiff’s claim against Kremer as moot. (Dkt. #71.) The
court will take up that motion in a later opinion and order.
14

29

appropriate remedy.

ORDER
For the reasons state above, IT IS ORDERED that:
1) Plaintiff One Wisconsin Now’s motion for summary judgment (dkt. #27) is
GRANTED.
2) Defendants Jesse Kremer, John Nygren and Robin Vos’s motion for summary
judgment (dkt. #56) is DENIED.
3) Defendants’ motion to strike plaintiff’s expert (dkt. #38) is GRANTED IN
PART AND DENIED IN PART as set forth above.
4) Plaintiff’s motion to file a sur-reply brief (dkt. #69) is GRANTED.
5) Defendants’ motion to strike plaintiff’s jury demand (dkt. #74) is DENIED AS
MOOT.
6) Defendants’ motion to supplement their motion for summary judgment (dkt.
#77) is GRANTED.
7) All remaining deadlines in this case are STRUCK, except for the briefing
schedule on relief set forth above and defendant Kremer’s motion to dismiss
previously set by the court. The parties are also to hold open March 11, 2019,
for possible argument and presentation of evidence as to the appropriate remedy,
if necessary.
Entered this 18th day of January, 2019.
BY THE COURT:
/s/
__________________________________
WILLIAM M. CONLEY
District Judge

30