A Survey of Official and Unofficial

Law Enforcement Twitter

Accounts in Canada, the United
Kingdom, and the United States

A Report of

Laura Madison
Christa M. Miller
Chris Worden

© 2010 Canadian Association of Police on Social Media

CAPSM provides leadership for law enforcement on the proven, reliable and most effective best
practices for Internet based communication tools relating to both internal and external purposes,
and to advocate for the use such technologies for the communities they serve and the members
they represent.

TABLE OF CONTENTS

Abstract…………………………………………………………………………………………………….3
Introduction………………………………………………………………………………………………..3
Background………………………………………………………………………………………………..4
Internet Usage……………………………………………………………………………………4
Twitter Usage…………………………………………………………………………………….5
International Twitter Usage……………………………………………………………………..5
Methodology………………………………………………………………………………………………6
Results……………………………………………………………………………………………………..7
Authoritative Visuals: Use of Twitter Features………………………………………………..8
Use of Third-Party Tools……………………………………………………………………….11
Tweet Automation………………………………………………………………………………13
Legal Disclaimers and Policy………………………………………………………………….13
Twitter Activity and Subject Matter……………………………………………………………14
Conclusions and Recommendations………………………………………………………………….19
References……………………………………………………………………………………………….21
Appendix…………………………………………………………………………………………………22
ABSTRACT

To use Twitter to its fullest potential for public communications, emergency management, and
other functions, law enforcement agencies must first understand the medium -- not only how
citizens use it, but also how their peers use it both officially and unofficially. This study, a survey
of 1,089 police and police-related Twitter accounts, used 25 different criteria to show how
agencies and officers are using Twitter, where they can improve, and implications for their future
use.

Keywords: law enforcement, Twitter, crisis informatics, social media policy, third-party Twitter
tools, public information, community relations

INTRODUCTION

The way law enforcement agencies use social networking for public relations and
communications is not well understood. It is easy to look at a handful of Twitter and Facebook
accounts, and think they represent all law enforcement agencies on those particular social
networks, whether they are active or inactive or positive or negative in tone.

However, this does a disservice to the 1,000+ law enforcement agencies and officers on Twitter,
as well as the communities they serve. The fact is that agencies approach the way they use
these tools in a variety of ways: personal accounts both official and unofficial; official department
accounts; via mobile devices, third-party tools and the basic web interface; and in conjunction
with other social networks. They may assign one person or a team to maintain their accounts,
opt for a positive or neutral tone, or try an account briefly before ending its use.

The goal of this study is to refine our current understanding of the way law enforcement
agencies in Canada, the United States and the United Kingdom use the social networking tool
Twitter. Popular because of its messages’ brevity, Twitter allows for one-to-many
communication, at the same time facilitating an unprecedented means for two-way
communication between police and public.

Currently, most if not all analysis of law enforcement Twitter use is qualitative and anecdotal.
This study quantifies that use and analyzes levels of engagement, such that law enforcement
commanders, public information officers, members of the media, and members of the public
(among others) will be able to assess their local agencies’ and officers’ use compared to others’.
The study will also help police to develop best practices regarding Twitter use, whether personal
or official.

Researchers wanted to find out whether police users:

● Fully utilized Twitter’s many features, such as professionally made backgrounds and
appropriate avatars, to create a better sense of engagement with their followers.
● Accessed Twitter via the standard web interface, or used third-party and mobile-device
 tools. This would indicate whether they had a good command of social media
capabilities as well as whether they accessed their accounts from the field.
● Automated any aspect of their Twitter engagement, including whether they tied their
accounts to other social tools such as Facebook or twitterfeed. This could indicate not
just a good command of social media capabilities, but also ensure postings even when
the officer was unavailable.
● Included legal disclaimers or other policy about their social media use.
● Tweeted information about their on-duty activities including arrests; information about
wanted or missing persons, local traffic advisories, or community volunteer initiatives; or
links to local news media.
● Tweeted personal opinions about crime or criminal justice; hobbies or other personal
interests; potentially sensitive information about subjects or ongoing cases; or prejudicial
statements.

The study also assessed how talkative, or engaged, police users were.

BACKGROUND

Although no law enforcement-specific social media research currently exists, plenty of studies
have been done which can be applied to police social media use.

Internet Usage

Notably, law enforcement agencies must foremost be aware of how the Internet is actually used,
not just in their own communities but also nation- and worldwide. Digital media measurement
firm comScore Inc. notes that nearly three-quarters of global internet users monthly use social
networking sites. However, the World Internet Project, conducted by the Center for the Digital
Future at the USC Annenberg School for Communication & Journalism, found that just half of 10
reporting countries reported more than a majority of their respondents use the internet. In the
United States, 78 percent of respondents were internet users.

About equal numbers of men and women in the United States use the internet, according to the
WIP, as do 42 percent of American respondents aged 65 or older. Of those who do not, most
are not interested or do not see the internet as useful. Fifty-seven percent of American
respondents believe that half or less of the information they receive online is valuable.

These findings are important for law enforcement agencies for a variety of reasons. They cannot
expect to rely solely on social media for their community relations, but can consider social
media to be an enhancement to their communications efforts. If about 22 percent of the
population and more than half of elderly adults do not use the internet, and if more than half of
users do not trust online information reliability, then police must consider the effect on their
communications about crime trends, disasters, issues of public trust, and even more mundane
events like traffic incidents. With that said, it is better to be available to supply correct
information to citizens who do use the internet and social media, then to fail to be in those online
spaces.

Twitter Usage

With regard specifically to Twitter, comScore Inc. reports that in June 2010, nearly 93 million
Internet users visited Twitter.com from their homes or workplaces. Exclusive of third-party
Twitter applications like TweetDeck, this represented an increase of 109 percent from the
previous year. Regionally, Europe saw 22.5 million Twitter visitors in the same time period, an
increase of 106 percent, while North America’s 25 million visitors represented a 22 percent
increase for that region.

In comScore’s six mobile markets (the United States, the United Kingdom, France, German,
Spain and Italy), Twitter use is growing among smartphone users. For the purposes of this
study, 4.2 million (8.3 percent) of US smartphone users accessed Twitter.com in one month
from their mobile devices, while 5.8 percent of smartphone users in the United Kingdom did
likewise.

On the other hand, a recent survey showed that half of Twitter users tweet less than once per
day, and nearly a third have never tweeted. Meanwhile, only ten percent of users tweet five or
more times per day, despite Twitter activity increasing.

Again, police must consider the implications of these statistics for their online communications --
this time for technology-savvy users in their communities, who are more likely to report about
their involvement in large- or small-scale disasters. The nascent field of “crisis informatics” as
described by Palen et al. (2010) takes as a matter of course that members of the public will
respond to and become involved in crises, including actively communicating information to one
another.

Part of this is the practice of “retweeting,” or copying tweets with credit given to the originator(s).
Retweeting serves a valuable purpose in that it can spread the reach of information much
further than a single user’s network; thus, government agencies must consider it valuable, and
take pains to tweet content that their followers will want to retweet, in particular information that
has URLs and hashtags included (Suh 2010).

International Twitter Usage

Apart from emergencies, it should not go unnoticed that Indonesia, Brazil and Venezuela had
the highest numbers of people visit Twitter.com (out of the 41 countries which comScore
reports). With Indonesia at 20.8 percent market penetration, Brazil at 20.5 percent, and
Venezuela at 19 percent, and countries like the Philippines, Mexico and Singapore not far
behind at 14.8 percent, 13.4 percent, and 13.3 percent respectively, police users in North
America and the UK must remember that these countries rank high for cyber criminal activity.

It follows that, even if the actual investigation of crimes remains the purview of federal and state-
level task forces, reporting may well be initiated with local law enforcement when citizens
become aware that the innocuous-looking link they clicked from Twitter downloaded malware to
their computer or smartphone. Even if local police never find the malware’s origin, higher market
penetration in countries already known to host cyber criminals increases the likelihood of North
American and British crime victims, and thus of reporting.

Indeed, as a recent Barracuda Labs report states, activity is increasing on Twitter. More people
are joining, and the nearly 29 percent of “True Twitter Users” along with casual users are
becoming more active -- resulting in more malicious activity. The Twitter Crime Rate, or
accounts that ended up being suspended by Twitter, was 1.67 percent through the first half of
2010.

METHODOLOGY

Researchers estimated the total number of police Twitter users across Canada, the United
Kingdom and the United States at 1,089, and analyzed this number of accounts. Although it is
possible that some police users joined Twitter as the study was ongoing and therefore were
missed, this number is likely to be small and not affecting the overall sample size.

Researchers began to “collect” or follow law enforcement Twitter accounts in April 2010, a
process that continued through July 2010. Accounts were not separated by whether they
belonged to an officer or to an agency, although they were separated by country.

Account assessment began in July 2010 and ended in August 2010. Researchers coded each
account according to 25 different criteria. These assessments were entered into a Microsoft
Excel spreadsheet for each country and then tabulated. Analysis of the research questions was
completed using spreadsheet tabulation.

Researchers did not code for responsiveness -- whether and how a Twitter user replies to
followers. As the study was intended only to be a survey of use, correlations were not
completed.

The study also does not try to compare Twitter accounts with actual numbers of law
enforcement agencies and officers in each country. That is because some individual officer
accounts are official, and others unofficial; they do not all represent an accurate cross-section of
police in a particular country or even regions within.

A marginal number of unofficial agency accounts, including police unions, fraternities and
civilian-created “scraper” feeds of computer aided dispatch calls, were excluded from the study
because researchers did not believe they accurately represented answers to the research
questions. Also excluded were police accounts from outside the Canada, the United Kingdom,
and the United States.
RESULTS

Of the 1,089 police accounts currently in existence, a majority of nearly three-quarters (800
accounts, or 73 percent) belong to police officers or agencies in the United States. One-fifth
(227, or about 21 percent) belong to police in the United Kingdom, with the remaining 62
accounts (about 6 percent) belonging to Canadian police.

Total Accounts
Canada UK US
6%
21%

73%

At the close of the study in August 2010, accounts in all three countries had tweeted a total of
376,821 times. Percentages were roughly proportional to account numbers in each country:
about 75 percent came from the United States, about 16 percent from the United Kingdom, and
about 9 percent from Canada.

Total Number of Tweets
Canada UK US
9%

16%

75%

Accounts in all three countries have a total of 489,470 followers, with 83 percent of them in the
United States, 10 percent in the United Kingdom, and 7 percent in Canada.
 Total Number of Followers
Canada UK US
7%
10%

83%

However, the proportions shift somewhat when it comes to the number of accounts the police
are following. Police follow a total of 98,128 accounts; just 69 percent of these are in the United
States, while 21 percent are in the United Kingdom and 10 percent in Canada.

Total Number of Following
Canada UK US
10%

21%

69%

However, it is often said that follower numbers are not an appropriate measure of engagement
because they do not provide the entire picture. Indeed, although they have the fewest accounts,
Canadian police have an average of 590 followers (compared to 80 for British police and 506 for
American). And they follow an average of 156 accounts (compared to 93 in the United Kingdom
and just 84 in the United States).

Authoritative Visuals: Use of Twitter Features

Researchers wanted to find out whether police users fully utilize Twitter’s many features, such
as professionally made backgrounds and appropriate avatars, to create a better sense of
engagement with their followers.
 Na
Na
m m

50%
100%

0%

0%
50%
100%
e & e &
  Ra
 R
an
nk k
Bi
Bi
o  o  C
Co om
m pl
pl et
et
e In e
In sig
sig Na
n n m
e 
&
ia  ia 
 R
0%
20%
40%
60%
80%
100%

or or
 B  B an
ad ad Bi k
o 
ge ge Co
In m
Ba Ba sig p le
ck ck ni te
a 

Canada Yes
gr gr or
ou ou Ba  B
a
Grand Totals Yes

nd nd ck dg

UK Yes
 C  C gr
ou e
ra ra nd
. .. . ..  C
ra
Ph ft
ed
Ph Ph ot
Lin o 
ot ot k  of
o o  O

UK No
to
 of  B ffi
lo ce
 of
 O r

Canada No
 O g 
o
Grand Totals No

f r
fi c ffi Av
Ve
r  W
er ce at ific eb
r ar at sit
Lin Lin  is io e
Totals ‐ All Nations

k k  B n 

UK NA
 t o ad by
 T
 t o ge
 B  B  o w
lo lo r o itt
er
g o g o ffi
r r ce

Authoritative Visuals ‐ UK
... ... r p
Grand Totals NA

ho

Canada NA
Ve Ve to
rif rif
ic i
Authoritative Visuals ‐ Canada

at ca
io tio
n  n
by  b
 ... y .
..
Av Av
at at
ar ar
 i s B  is
ad  B
ge ad
 . . ge
 . .
 Authoritative Visuals ‐ US

100%
80%
60%
40%
20%
0%
Name & Rank Bio Complete Insignia or Badge Background Photo of Officer Link to Blog or Verification by Avatar is Badge
Crafted Website Twitter or officer photo

US Yes US No US NA

Use of name and rank is important to transparency, or an officer’s or agency’s willingness to

take responsibility for its content. Overall, only 32 percent of accounts state name and rank. Of
all three countries, the United Kingdom was strongest at using this detail; nearly half of its
accounts made the statement. In the United States, that percentage dropped to 28 percent, and
in Canada, was lowest of all at 16 percent.

A completed biography -- the 160-character notation in the upper right-hand corner of a Twitter
page -- can be a qualitative indicator of transparency. Twitter bios convey a degree of
personality, to the extent that they may be entirely professional -- or take a tone of silliness,
sarcasm, satire, and so forth. Left blank, they can leave a viewer guessing as to the account’s
purpose or even its legitimacy.

Nearly three-quarters of all accounts had completed bios. The percentage was not far off in the
United States and Canada -- with 72 percent and 69 percent, respectively -- but in the United
Kingdom, 81 percent of accounts showed completed bios.

A police officer or agency that uses a badge, patch, or photo of a cruiser or department as an
avatar and/or part of a background may convey a sense of authority, and perhaps more
importantly, distance. Conversely, an officer or agency that uses a photo of him- or herself in the
avatar or background may be seeking to convey a sense of humanity. Officers who use pictures
of themselves in uniform likely seek a balance between humanity and authority.

Insignia or badges were present in a little more than half of all accounts. This percentage was
about the same in the United States, but rose to nearly 70 percent in Canada. Meanwhile, only
40 percent of British accounts used insignia or badges. However, about 47 percent of all British
accounts used officer photos, compared with only 24 percent of Canadian accounts and 13
percent of American accounts.

Whether the account’s avatar contained an officer photo or badge was also measured. These
depictions were present in about 72 percent of all accounts. The percentage was similar in
Canada (71 percent), the United Kingdom (68 percent), and the United States (73 percent).

Twitter users have a choice between uploading their own background image, or using the web
interface default. For a police department, an official-looking background that is crafted with
insignia, officer image, contact information and/or legal information can help to convey a sense
of professionalism and authority that is not present in a generic default background.

Nearly half (43 percent) of all accounts featured a crafted background. This percentage was
slightly higher in the United States with 48 percent, but it was lower in Canada with 34 percent
and lower still in the United Kingdom with 26 percent.

Twitter users also have the option of linking to a website from their account profiles. For law
enforcement users, that might be a department website or a blog. The presence of either
indicates users who seek to communicate with others beyond 140-character tweets; linking to a
website can provide additional context to department activities, while a blog -- if regularly
updated -- conveys the desire to provide more dynamic content to the public and can also help
with verifying that its actually a police account if links on the blog or website are also present.

Nearly two-thirds of all accounts linked to a website or blog. This percentage was similar in the
United States and in the United Kingdom, but was closer to three-quarters in Canada.

Finally, a verified account indicates that the Twitter user is indeed who s/he says s/he is. This is
most common among celebrity users, but can be a useful brand management tool for
government agencies too -- particularly law enforcement agencies, which may be susceptible to
pranksters or even users with more malicious intentions.

However, researchers found that only eight of those law enforcement accounts indexed in this
study were verified by Twitter staff -- one in the United Kingdom, and the other seven in the
United States.

Use of Third-Party Tools

Researchers wanted to know whether police users accessed Twitter via the standard web
interface, or used third-party and mobile-device tools. This would indicate whether they had a
good command of social media capabilities as well as whether they accessed their accounts
from the field.

Totals ‐ All Nations

100%
80%
60%
40%
20%
0%
Web Only 3rd Party API in use RSS in use Facebook to Nixle in use Mobile in use
Platforms Twitter

Grand Totals yes Grand Totals no Grand Totals na

 Tools Utilized ‐ Canada
100%
80%
60%
40%
20%
0%
Web Only 3rd Party API in Use RSS in Use Facebook to Nixle in use Mobile in use
Platforms Twitter

Canada yes Canada no Canada na

Tools Utilized ‐ UK

100%
80%
60%
40%
20%
0%
Web Only 3rd Party Platforms API in use RSS in use Facebook to Twitter Nixle in use Mobile in use

UK yes UK no UK na

Tools Utilized ‐ US
100%
80%
60%
40%
20%
0%
Web Only 3rd Party API in use RSS in use Facebook to Nixle in use Mobile in use
Platforms Twitter

US yes US no US na

A little more than half of the accounts in all countries utilize the web interface only. About one-
quarter use third-party tools like TweetDeck, HootSuite, and the like; American and British
adoption of these tools, which are 23 percent and 26 percent respectively, represent a lower
rate of adoption than in Canada, 40 of whose accounts use these tools.

All third-party tools use Twitter’s Application Programming Interface (API), but not all tools
change the API setting to reflect use of their particular brand name. Researchers recorded “API”
separately from other third-party tools to account for this.
Fifteen percent of all accounts use API tools. The rate of American law enforcement agencies
using API tools is much higher -- nearly 18 percent -- than in either Canada (6.5 percent) or the
United Kingdom (3.5 percent) due in large part to American use of the service Nixle. Nixle,
which uses the National Law Enforcement Telecommunications System (NLETS) platform, was
touted as a way to broadcast community information without the need to engage, as
conventional Twitter use demands. Of the American law enforcement agencies that use Twitter,
12 percent use Nixle to push updates to Twitter.

Finally, researchers coded for whether mobile devices (and applications) were in use. (It should
be noted that Nixle is accessible via mobile device, but was not counted this way because there
is no way to tell when it is used in the field vs. from a stationary location.) Twenty-seven percent
of all law enforcement Twitter users use mobile devices. The rate of adoption is highest in the
United Kingdom at 39 percent, followed by Canada with 35 percent mobile use. Only about one-
quarter or 23 percent of American police tweeters, however, use mobile devices to tweet.

Tweet Automation

Researchers wanted to find out whether police users automated any aspect of their Twitter
engagement, including whether they tied their accounts to other social tools such as Facebook
or twitterfeed. This could indicate not just a good command of social media capabilities, but also
ensure postings even when the officer was unavailable.

Researchers separated the use of RSS, or tools such as Twitterfeed, from other third-party
tools. Use of RSS tools indicates the desire to share news -- most commonly blog entries, but
also any news items published via RSS, or Really Simple Syndication. Close to one-fifth of all
active law enforcement Twitter users use RSS; this is broken down into eight percent of
Canadian accounts, about six percent of American accounts, but two-thirds of all British
accounts.

Whether Facebook updates are pushed to Twitter is important because it reflects a different
level of engagement, a desire to seek a broader audience by pulling one group of people
(Twitter followers) towards another medium in which to engage. Nevertheless, Facebook-to-
Twitter adoption rates are low: just about ten percent overall, with Canada and the United
Kingdom sharing similar rates (eight and six percent, respectively) and 11 percent of agencies
in the United States using Facebook-to-Twitter.

Legal Disclaimers and Policy

Researchers wanted to know whether police users included legal disclaimers or other policy
about their social media use. Both make clear what followers should and should not expect from
interacting with police on Twitter.

However, none of the 62 agencies or officers accounts in Canada included either disclaimers or
policy. Nine American agencies had legal disclaimers, but no policy links. In the United
Kingdom, meanwhile, no agencies posted legal disclaimers. Only one agency -- the London
Metropolitan Police -- included a link to its Twitter policy. (Interestingly, however, at the time of
the survey the Met had not tweeted, despite its detailed policy.)

Twitter Activity and Subject Matter

Researchers assessed how talkative, or engaged, police users were. They defined “talkative” as
accounts that tweeted one or more times per day, most every day of the week, “active” as
accounts that tweeted several times per week, and “sparse” as accounts with only few tweets
per month or less. Accounts with no tweets were counted along with locked or protected
(accessible only to approved followers) accounts.

Totals ‐ all Nations
Talkative Active Sparse None Locked

6% 3% 13%

40%
38%

Overall, most accounts fell into the “Active” or “Sparse” categories with 38 and 41 percent of
accounts, respectively. Only 12 percent could be considered “Talkative,” but just six percent had
no tweets, and only three percent were locked.

Police Twitter users in the United States showed very similar proportions. However, Canadian
agencies had a higher proportion of “Talkative” accounts, while in the United Kingdom, the
differences were profound -- 21 percent of accounts were “Talkative,” 42 percent were “Active,”
and only 31 percent were “Sparse.”
 Canada
Talkative Active Sparse None Locked

5% 3%
18%

42%
32%

UK
UK  UK Talkative UK Active
2% UK Sparse UK None UK Locked

5% 0%
21%

31%

41%
 USA
US  US Talkative US Active US Sparse US None US Locked
3%
6% 0% 10%

37%
44%

Researchers wanted to find out whether police users tweeted information about their on-duty
activities including arrests; information about wanted or missing persons, local traffic advisories,
or community volunteer initiatives; or links to local news media.

Qualitative & Quantitative ‐ Totals
100%
80%
60%
40%
20%
0%
s
ty
sts

sts
ew

...
..

 ...
fe
is.
re

. ..

re
..

en
l n
Sa

ive
 a.
m
 M
r

te

m
r a

c/

ca

ns
Co

in

sit
or

m
ffi
y o

lo

s, 
o
 

en

co
d 

e &
a

i
ts 

in

ie
ut

 Tr
te

y s

al 
op
ee

bb
c
t d

an

ici
all
ol
ou

ho
al
t W
ou

ud
t P

i
et
 ab

on

nt
 
ut
ab

/r

ej
ou

te
ou

rs
s

o
s

pr
ts 

po
t

pe
ab

ab
ee
ab

ee

ts 
ee

ts 
ts 

ts 

ts 
ts 

Tw

Tw

ee
Tw

ee
ee

ee

ee
ee

Tw
Tw
Tw

Tw

Tw
Tw

yes no na
 Tw
ee
ts 
ab
ou Tw
Tw t d ee
ee ut ts 
ts  y  o ab

0%
20%
40%
60%
80%
100%
ab r a Tw ou
ou . .. ee
0%
50%
100%

Tw t d
ee t W ts  u.
ab .
ts 
ab
an
te Tw o ut
ou d. ee  W
Tw t T .. ts  ...
ee r af ab
ts  fic Tw ou
ab /S ee t T
ou ... ts  r..
Tw t P ab .
ee o Tw o ut
lic ee  P

yes
yes
ts/
re e . ts/ o.
tw .. re ..
Tw ee tw
ee Tw

no
no

ts  ee ee
ts  lo ts.
pe ca ts  ..
Tw r so l... Tw p er
s

na
na

ee na ee on
ts  l o ts  a.
ab pin ab ..
ou .. . Tw ou
Tw t h ee t h
ee ob ts  ...
ts  bi po
po es Tw te
te , ..
Tw nt . ee nt
ial
ee ial ts  ..
l
Qualitative & Quantitative ‐ UK

ts  y s pr
pr en ej
ej ud
Qualitative & Quantitative: Canada

ud .. . ici
ic i ..
al 
c..
.
 Qualitative & Quantitative ‐ USA
100%

50%

0%

.
.

e.
...

..
...

.. .

..
y  .

..

...

a.
p.
lo

bi
an

ic.
fi

ly 
ut

ic i
l o
af

ob
ts 
ol
t W
t d

ial

ud
r

ee

na

t h
t P
t T

nt
ou

ej
ou

o
tw
ou

ou

ou

te
rs

pr
ab

ab

re

po
pe
ab

ab

ab

ts 
ts 

ts/
ts 

ts 

ts 

ts 

ts 

ts 

ee
ee

ee

ee
ee

ee

ee

ee

ee

Tw
Tw

Tw

Tw
Tw

Tw

Tw

Tw

Tw
yes no na

About half of all users tweeted about on-duty activities, including arrests. In Canada, this
percentage was closer to 63 percent; in the United Kingdom it was as high as 70 percent, while
in the United States it was only about 42 percent.

Nearly 38 percent of users tweeted about wanted or missing persons. While in the United
States, the percentage was similar at 34 percent, in the United Kingdom the rate was higher at
46 percent, and in Canada higher still at 52 percent.

About 54 percent of users tweeted about traffic or other public safety advisories. American
users tweeted these at a slightly lower rate (50 percent), Canadian users at a slightly higher rate
(58 percent), and British users at nearly two-thirds (64 percent).

A little more than one-third (35 percent) of all accounts tweeted about police and community
volunteer activities. Sixty-nine percent of Canadian users tweeted on these topics; likewise
British users. In the United States, however, this rate fell to just under one-quarter.

Users were not likely to tweet links to local news media stories. Overall, only about one-fifth of
users did so. This was about the same in the United States, and higher in Canada (27 percent),
but lowest of all in the United Kingdom (seven percent).

Researchers also wanted to know whether police users tweeted personal opinions about crime
or criminal justice; hobbies or other personal interests; potentially sensitive information about
subjects or ongoing cases; or prejudicial statements.

Only nine percent tweeted personal opinions about crime or criminal justice. This percentage
remained consistent across Canadian, British and American users.
Nearly one-fifth of all users tweeted about personal interests or hobbies. This was about the
same in the United States, slightly lower in Canada at 17 percent, but higher in the United
Kingdom at 22 percent.

Whether police users tweet potentially sensitive legal information -- such as clues that could
identify suspects or victims of crimes -- was of interest, in part because of media stories about
police officers posting pictures of crime victims and scenes online. Overall, 99 percent of users
avoid such tweets. In the United Kingdom and Canada, only two accounts apiece tweeted
sensitive information; in the United States, only 11 accounts did.

Finally, accounts were assessed for whether they tweet prejudicial or marginalizing statements,
such as grouping “people on that side of town.” or statements that could be interpreted as racist
or sexist. Overall, only three percent of users made such statements. Only one Canadian
account had made them; no accounts in the United Kingdom had. However, in the United
States, 29 accounts made such statements at least once.

CONCLUSIONS AND RECOMMENDATIONS

Police in none of the three countries have a “better handle” on Twitter use than in others.
Instead, each group of police has its own strengths and focuses, some of which may be cultural
or political while others may reflect better command of social media and how it can serve the
community.

For example, British police have the fewest followers on average, but have the highest
proportion of “talkative” as well as “active” Twitter accounts. They also have the highest
proportion of RSS use, indicating a degree of automation which may not be of as much use to
followers as more personalized tweets.

On the other hand, Canadian police, even with the fewest accounts of the three, have the
highest average number of both followers and followees, and a higher proportion of “talkative”
accounts. Meanwhile the United States, with the most accounts and most tweets, have the
fewest average followees as well as more “sparse” accounts.

In general, then, police should follow more people and companies in their own region -- not just
for the sake of gaining followers or with the intent of reading every tweet, but to increase their
reach: the likelihood of their tweets being retweeted during times of crisis or other need. In
addition, law enforcement agencies should commit to tweeting at least 2-4 times per week in
order to stay on their followers’ radar (including responding to tweets or retweeting interesting
local information). They should experiment with tweeting at different times of day to see which of
their followers (for instance, private citizen vs. company) engages via retweet or conversation;
they should also experiment with content, including links and even hashtagging.

On average, British police tweet the most about their on-duty activities and traffic and public
safety advisories, while American police are strongest of the three countries at tweeting missing
and wanted persons information, but are least likely to tweet community and police volunteer
activities. All are worthwhile types of content.

So are links to local news stories, yet agencies across all three countries tweet them least of all.
In fact, no law enforcement agency should discount the value of local media, just because social
media make it easier to tell their own stories. While in the United States, the police/media
relationship can be adversarial, traditional media are important conveyances of information in
times of crisis. Engaging them can take a variety of forms: following them, tweeting their police-
related story links (even if they are redundant to an agency’s own press release or blog entry).
An agency that shows it is paying attention to its community earns social capital and credibility.

However, it is important that public information officers not be the sole representatives of law
enforcement on Twitter. Individual officers with their own accounts have important voices; policy
can and should drive their tone, but in general, because the vast majority of officers using
Twitter do not make legally sensitive or marginalizing remarks, they should be trusted to make
the right decisions and perhaps even support the agency’s efforts in disseminating information.
However, they should also be trained to tweet, whether on duty or off, with a particular focus on
not tweeting “just anything” but rather having a strategy for Twitter activity.

This kind of use can be defined and governed via internal guidelines, which should depend only
partly on “best practices.” It should also be specific to the agency’s culture and individual officer
preferences. In fact, the lack of legal disclaimers (e.g. “The opinions expressed here do not
reflect those of my employer” or “This account is not monitored 24/7; please do not use it to
report crimes”) and social media policy (which may include disclaimers as well as purpose and
other information) is of concern. Forming and then publicizing such statements helps clarify the
account’s purpose along with that of each person who maintains it. The statements also help
citizens understand what to expect.

Law enforcement users in all countries can improve their use of visual cues to convey purpose,
professionalism, transparency and humanity. These include officer photos and/or images of
badges and other insignia, along with complete biographies and official-looking backgrounds. In
general, British police do a better job of using their own photos and of filling out their bios, while
American police are more likely to have official-looking backgrounds and their names and/or
ranks clearly displayed. Canadian police have the most links to their websites or blogs. (Web
presences should also have prominent links to Twitter and other social accounts.)

Whether a law enforcement account needs to be verified should be a function of its size and
reputation. It is not unheard of for pranksters or even helpful citizens to “brandjack” an agency
account. Some have tweeted only computer aided dispatch information, but others (most
notably in Austin, Texas, USA) have made fake accounts look official, and caused brand
damage. Thus while verification is not necessary for every single agency, it is worth careful
consideration.

Finally, police Twitter users should try to experiment more with third-party and mobile Twitter
platforms. American users in particular need to school themselves on tools that exist and how to
use them; for example, British police use of RSS-to-Twitter far surpasses use in either Canada
or the United States (although again, care must be taken not to over-rely on any one tool). In
multilingual communities, meanwhile, TweetDeck’s translation tool may come in handy.

American police especially need to work harder at adopting mobile Twitter use. Being able to
tweet from the scene of a bad incident or disaster can be critical to successful crisis monitoring
and management.

That more police are using Twitter to connect with their communities is encouraging. However,
they should take care not to use it only because it is popular or because the neighboring agency
signed on. Instead, they must recognize it in the context of public communication at large: who
uses it, how they use it, and where it can fit into various types of police operations.

REFERENCES

Barracuda Labs. “2010 Midyear Security Report.” 2010.

http://www.barracudalabs.com/downloads/BarracudaLabs2010MidyearSecurityReport.pdf.
Accessed August 20, 2010.

Center for the Digital Future. USC Annenberg School for Communication & Journalism. 2010.
http://www.digitalcenter.org/WIP2010/wip2010_long_press_release_v2.pdf. Accessed August
20, 2010.

ComScore, Inc.
http://comscore.com/Press_Events/Press_Releases/2010/8/Indonesia_Brazil_and_Venezuela_Lead_Glo
bal_Surge_in_Twitter_Usage. Accessed August 20, 2010.

Palen, et al. “A Vision for Technology-Mediated Support for Public Participation and Assistance
in Mass Emergencies and Disasters.”
http://www.cs.colorado.edu/~palen/computingvisionspaper.pdf. Accessed August 20, 2010.

Suh, B.; Hong, L.; Pirolli, P. L.; Chi, E. H. Want to be retweeted? Large scale analytics on
factors impacting retweet in Twitter network. Second IEEE International Conference on Social
Computing (SocialCom); 2010 August 20-22; Minneapolis, MN.
http://www.parc.com/publication/2489/want-to-be-retweeted-large-scale-analytics-on-factors-
impacting-retweet-in-twitter-network.html. Accessed August 21, 2010.
APPENDIX

CAPSM Research Framework for Data Analysis

Totals: US, UK, Canada

1. Total Tweets
2. Total Followers
3. Total Following

Authoritative Visuals: Visual representations and links located on police Twitter

accounts.

1. Name & Rank

2. Bio complete
3. Insignia or badge
4. Background crafted
5. Photo of officer
6. Avatar is badge or officer photo
7. Link to blog or website
8. Verification by Twitter

Tools Utilized by Police Officers and Agencies When Using Twitter.

1. Web Only
2. Third Party Platforms
3. API in use
4. RSS in use
5. Facebook to Twitter
6. Nixle in use
7. Mobile in use

Quantity and Quality of Communications: Police Officers’ Activity and Topics.

1. Talkative, active etc

2. Tweets about duty or arrests
3. Tweets about wanted or missing persons
4. Tweets about local traffic or safety
5. Tweets about police and community volunteer activities
6. Tweets or Retweets local news Tweets from Main Stream Media
7. Tweets personal opinions about Crime or Criminal Justice System
8. Tweets about hobbies, food, travel, sports or exercise.
9. Tweets potentially sensitive legal information or that may ID subjects
10. Tweets prejudicial comments or statements that could marginalize an identifiable group