22
DE GRUYTER MOUTON Communications 2014; 39(2): 129–149 Christopher Buschow, Beate Schneider and Simon Ueberheide Tweeting television: Exploring communication activities on Twitter while watching TV Abstract: “Social TV”, described as the use of social media such as Twitter or Facebook stimulated by TV programs, is highly topical in the television indus- try. Communication research has fallen behind in addressing this issue. In this paper we explore the simultaneous communication activities of Twitter users while watching TV. Additionally, we tested whether different TV programs stim- ulate different communication activities. The main findings of our quantitative content analysis of approximately 30,000 messages show that communication within the Twitter community as well as evaluations of shows and actors are the main subjects of the explored tweets. We also found that different TV pro- grams evoke different communication activities. While talent shows produce expressions of fandom and critiques of the candidates in the show, live events evoke a critical debate about the show itself and what’s happening on screen. Political talk shows can stimulate a public discourse. Keywords: television, communication activities, Twitter, Social TV, content analysis DOI 10.1515/commun-2014-0009 1 Introduction Due to the emergence of new alternatives to linear TV reception such as web video, video on demand, and Internet TV, the future of traditional linear televi- Christopher Buschow, Research Associate at the Hanover University of Music, Drama and Media, Germany, E-mail: [email protected] Beate Schneider, Professor for Communication and Media Studies at the Hanover University of Music, Drama and Media, Germany, E-mail: [email protected] Simon Ueberheide, Research Associate at the Hanover University of Music, Drama and Media, Germany, E-mail: [email protected]

Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Communications 2014; 39(2): 129–149

Christopher Buschow, Beate Schneider andSimon UeberheideTweeting television:Exploring communication activities onTwitter while watching TVAbstract: “Social TV”, described as the use of social media such as Twitter orFacebook stimulated by TV programs, is highly topical in the television indus-try. Communication research has fallen behind in addressing this issue. In thispaper we explore the simultaneous communication activities of Twitter userswhile watching TV. Additionally, we tested whether different TV programs stim-ulate different communication activities. The main findings of our quantitativecontent analysis of approximately 30,000 messages show that communicationwithin the Twitter community as well as evaluations of shows and actors arethe main subjects of the explored tweets. We also found that different TV pro-grams evoke different communication activities. While talent shows produceexpressions of fandom and critiques of the candidates in the show, live eventsevoke a critical debate about the show itself and what’s happening on screen.Political talk shows can stimulate a public discourse.

Keywords: television, communication activities, Twitter, Social TV, contentanalysis

DOI 10.1515/commun-2014-0009

1 Introduction

Due to the emergence of new alternatives to linear TV reception such as webvideo, video on demand, and Internet TV, the future of traditional linear televi-

Christopher Buschow, Research Associate at the Hanover University of Music, Drama andMedia, Germany, E-mail: [email protected] Schneider, Professor for Communication and Media Studies at the Hanover Universityof Music, Drama and Media, Germany, E-mail: [email protected] Ueberheide, Research Associate at the Hanover University of Music, Drama andMedia, Germany, E-mail: [email protected]

Page 2: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON130 C. Buschow, B. Schneider and S. Ueberheide

sion is generally considered to be one of the most important issues facing to-day’s media industry. Such developments create a challenge for the players inthe market and raise the question as to how they should react. As one strategyfor legacy media companies, practitioners (e.g., Bulkeley, 2010; Ericsson Con-sumerLab, 2011; Nielsen, 2011; Richardson, 2010) and researchers (e.g., Baillie,Fröhlich, and Schatz, 2007; Chorianopoulos and Lekakos, 2008; Geerts, 2009;Schatz, Baillie, Fröhlich, and Egger, 2008; Summa 2011) propose merging tele-vision and social media technologies into ‘Social TV’. Social TV opens up newpossibilities of value creation for TV producers and writers (Andrejevic, 2008),networks and channels (HMR International, 2011), advertisers (Zigmond andStipp, 2010) as well as agencies (Nielsen, 2011). Practitioners argue that net-works and channels in particular can benefit from Social TV because userscould return to linear TV driven by simultaneous communication in social me-dia. Ynon Kreiz, the former CEO of Endemol Group, stressed the importance ofthis development when he said: “We think that social media meets televisionis the next big thing […] Whoever figures it out will be the next Steve Jobs ofthis generation” (Bergman, 2011a).

Despite this practical relevance, only a handful of studies addresses inter-personal and group communication while watching TV (e.g., Barker, 1997;Hepp, 1998; Klemm, 2000; Lee and Lee, 1995; Westerik, 2009). Research onsimultaneous communication emerged prior to the development of the Internet:Often based on qualitative research, this literature emphasizes communicationwhen groups watch together at a single location. Traditionally, the researchhas centered on family members’ face-to-face-interactions in front of a singleTV screen (e.g., Morley, 1986; Westerik, 2009). From this point of view, “TV hasalways been social”, as Christy Tanner, CEO of TVGuide.com, puts it (Bergman,2011b).

Today however, new technologies and social media have extended inter-personal communication beyond the boundaries of family, friends, and geo-graphical regions. On social networking sites such as Facebook and microblog-ging services like Twitter, users chat in real time parallel to TV reception. Thisadditional function of television creates a particularly interesting challenge forcommunication research. TV is no longer just the linear “lean-back medium”(Dewdney and Ride, 2006, p. 289), but rather actively involves its users andstimulates interpersonal communication in social media – including amongstrangers. Social TV has transformed television into an ‘active medium’. Forexample, Twitter’s relevance for TV audience research has only recently beenhighlighted by academic research (e.g., Deller, 2011; Wohn and Na, 2011). Twit-ter offers the possibility to not only focus on particular shows, but on all typesof television programs (Deller, 2011). This is in direct contrast to the investiga-

Page 3: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 131

tion of specific online fan communities in earlier research (e.g., Franz, 2008 onAlly McBeal; Godlewski and Perse, 2010 on reality television shows) and enablescomparative analysis of different TV programs.

This paper focuses on the phenomenon of Social TV and empirically ex-plores users’ communication activities in social media while watching TV. Weconducted a study where we examined communication on Twitter in Germanyduring the broadcast of different German television programs. In the following,we introduce Social TV and explain its relevance in Germany. We review theliterature with regard to different communication activities and describe themethods used to explore our research questions.

2 Social TV as a phenomenonSocial TV relates to the communicative exchange about linear television con-tent, or that which is at least stimulated by it. Based on the several availabledefinitions (e.g., Cesar and Chorianopoulos, 2008; Chorianopoulos and Leka-kos, 2008; Research and Markets, 2010), it is not always clear whether pre-communication and follow-up communication are included (see Fig. 1).

Figure 1: Differentiation of communication activities by time of occurrence.Source: own research

In our research we focused only on the parallel communication about cer-tain TV programs during their broadcast. To be relevant for our investigation,communication had to be transferred by technology: here, via the social mediaplatform Twitter (Deller, 2011; Wohn and Na, 2011).

2.1 Twitter as a communication platform

The microblogging service Twitter allows users to post updates up to 140 char-acters (so-called tweets), answering the question “What’s happening?” About340 million tweets are sent (tweeted) daily. People also use Twitter to retweet

Page 4: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON132 C. Buschow, B. Schneider and S. Ueberheide

(RT) messages, thereby recommending content they like or wish to promote.Tweets are commonly marked using a hashtag (#) indicating the issue theyreference. Every word that is complemented by # is a hashtag that identifiesthe tweet’s topic. With more than 500 million active users worldwide and 4.1million users in Germany (Schmidt, 2012), Twitter is one of the most popularsocial networking sites. Although in Germany the microblogging service is no-where near as popular as in the US or in the UK, its popularity is growing fast:The number of unique visitors from Germany rose by 37 % between March 2011and March 2012 (ComScore, 2012). Many of the issues discussed on Twitter arerelated to TV programs (Deller, 2011; Wohn and Na, 2011). For example, GermanTV networks have more than 1.2 million followers on Twitter (BITKOM, 2012).Nevertheless, compared to the average viewing behavior of the wider public,Social TV remains a niche phenomenon in Germany. A recent online survey byTNS Infratest (2013) reports that only 28 % of 14- to 64-year-old Germans arefrequently online while watching TV; moreover, their activities are not alwaysconnected to TV. Only one-third of these users are involved in activities directlyrelated to a particular program (TNS Infratest, 2013).

3 Conceptual framework: Categories ofcommunication activities

Research on face-to-face-communication prior to the advent of Internet technol-ogies has shown that different types of communication activities can occurduring TV reception (e.g., Hepp, 1998; Holly, Püschel, and Bergmann, 2001;Klemm, 2000; Weisz et al., 2007; Wohn and Na, 2011). Thus, communicationsactivities are categorized in terms of (1) Evaluation of shows and actors, (2)Emotions, (3) Abstraction, and (4) Connecting with the community.

3.1 Evaluation of shows and actors

Positive and negative evaluations of specific programs and actors have beenidentified as one of the most common communication activities (Hepp, 1998;Klemm, 2000). Wohn and Na (2011) found “opinion (a subjective message aboutthe program)” to account for 38 % of users’ tweets concerning a speech byBarack Obama and for 32 % concerning an episode of So You Think You CanDance. In a content analysis of communication activities with regard to a car-toon, Weisz et al. (2007) classified 14 % of activities as evaluations of the shown

Page 5: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 133

video content. However, the main weakness in their study is that program-related activities account for 42 % and also include ratings and reviews of plots,characters, and actors (Weisz et al., 2007, p. 884).

3.2 Emotions

Studies show that viewers react emotionally to TV content (Godlewski andPerse, 2010; Wohn and Na, 2011). Therefore, emotions also play a major role inSocial TV. For example, in Wohn and Na’s (2011) study, emotions account forabout one quarter of all tweets in each investigated case. Godlewski and Perse(2010) show that negative emotion while watching a favorite program is a sig-nificant predictor of participation in online activity. Emotions also play a rele-vant role for Weisz et al. (2007): The authors point out that the emotion “joy”alone (operationalized as “laughter”) accounts for about 9 % of all messages.

3.3 Abstraction

Viewers associate TV content with their own lives. The television experience isintegrated into the immediate environment and reflected either on a personalor societal level (Matthewson, 1992; Holly, Püschel, and Bergmann, 2001). Onthe personal level, abstraction focuses on the meaning of a plot, a discussion,or an actor for the viewer’s own life. On the societal level, it highlights the roleof the content for society as a whole.

3.4 Connecting with the community

In the study by Weisz et al. (2007), about 23 % of activities were not directlyrelated to the content, but rather were part of private conversations. Two per-cent were “blurtings” (Goffman, 1978). Blurtings are minimal expressions, suchas “uh” or “wow”, releasing emotions in a simplistic way (Klemm, 2000, p.153). Examples from Twitter are “Blablablablablablabla. #illner” or “laaaaaaa-me #tatort”.

3.5 Genres of TV programs

Several studies have investigated the effect of different genres on Social TVactivities (e.g., Ducheneaut, Moore, Oehlberg, Thornton, and Nickell, 2008;

Page 6: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON134 C. Buschow, B. Schneider and S. Ueberheide

Ericsson ConsumerLab, 2011; Geerts, Cesar, and Bulterman, 2008; Geerts,2009): “Having exposed participants to various content genres, we found that[…] the type of TV content has a significant influence on communication”(Schatz et al., 2008, p. 1). Despite that shared conception, researchers disagreeon the genres and formats deemed suitable for Social TV. On the one hand, forexample, sporting events and reality shows are considered appropriate formatsby many authors (Ducheneaut et al., 2008; Ericsson ConsumerLab, 2011; Geerts,2009). On the other hand, there is little consensus: Ducheneaut et al. (2008)name cartoons, documentary films, and action and adventure films, while theauthors of the Ericsson ConsumerLab (2011) list news magazines, debates/talkshows, and TV events. In contrast to Ducheneaut et al. (2008), Ericsson Con-sumerLab (2011) argues that movies and series are rather irrelevant for SocialTV. In this regard there is still considerable ambiguity.

3.6 Overview of the research questions

Having presented Social TV as a phenomenon and Twitter as a communityfor Social TV as well as having theoretically differentiated the communicationactivities of TV viewers during reception, we will address the following researchquestions:

[RQ1] What communication activities do Twitter users engage in on the mi-croblogging service while watching TV in parallel?

From the conceptual framework we derive the following research sub-questionsfor RQ1:

[SRQ1] How relevant are evaluations of shows and actors?

[SRQ2] How relevant are emotions?

[SRQ3] How relevant are abstractions?

[SRQ4] How relevant are connections to the community?

A review of the literature has also shown that further investigation of TV genresand their respective influence on communication activities is still required:

[RQ2] How do different TV programs influence the communication activitiesof Twitter users?

Page 7: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 135

Our design therefore focuses on a wider sample of different TV shows. Theseshows differ with respect to such basic conditions as themes, story lines, oractors. Our research does not focus on analyzing such shows in detail, buton gathering varieties of possible communication activities. Using the samestandardized instrument, we aimed at comparing a diverse variety of shows touncover similarities or differences of these communications.

4 Method

A quantitative content analysis seems an appropriate method for investigatingthe proposed research questions (Krippendorff, 2004; Neuendorf, 2001). Thematerial collected consisted of communication activities on Twitter about cer-tain TV programs during broadcast. The unit of analysis employed was everysingle tweet relating to a certain program and within the broadcasting time.Fortunately, tweets are public and contain a timestamp. Tweets were identifiedby their hashtag. Hashtags seemed to represent an ideal applicability criterionfor this investigation, because they enable the easy identification of relevanttweets (Deller, 2011). In the case of the selected TV shows (see Table 1), theofficial hashtags were determined by the broadcasters and announced on TV,on the websites, or on broadcasters’/actors’ Twitter accounts. Overall, the sam-ple consisted of 31,388 tweets relating to seven different TV programs. A code-book was developed and a group of 45 students was trained by the researchersto code the tweets. After a pretest was conducted, Holsti’s coefficient was calcu-lated before the field phase (Holsti, 1969; Neuendorf, 2001). The codebook’spretest produced an overall coefficient of .916, with formal categories .958 andcontentual categories .908, and can thus be interpreted as very good reliabilitybetween coders.

4.1 Selection of TV programs

We chose to apply different criteria in the selection of TV programs:1. The programs needed to be of an appropriate genre for Social TV, such as

reality shows, talk shows, TV events, and sports (e.g., Ducheneaut et al.,2008; Ericsson ConsumerLab, 2011; Geerts, Cesar, and Bulterman, 2008;Geerts, 2009). Every genre had to be represented in the sample.

2. The programs had to be broadcast in German Free TV within the inquiryperiod which ran from January to April 2012.

Page 8: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON136 C. Buschow, B. Schneider and S. Ueberheide

Table 1: Selection of TV programs.

Genre Program Channel Hashtag

Talent show Deutschland sucht den Superstar RTL* #dsds(Pop Idol)The Voice of Germany Pro7* / Sat.1* #voiceIch bin ein Star, holt mich hier RTL* #ibesraus! (I’m a Celebrity, Get Me Outof Here)

TV event Schlag den Raab (Beat the Star) Pro7* #sdrTatort (Crime series) ARD** #tatort

Talk show / Debate Maybritt Illner (Political talk) ZDF** #illnerGünter Jauch (Political talk) ARD** #jauch

Source: own research; *private broadcast; **public service broadcast

3. The programs had to have public relevance so that active communicationon Twitter could be expected. As an indicator of public relevance, the rat-ings by audience research GfK were used.

4. The programs had to more or less reflect the German TV market (dual sys-tem: private broadcast / public service broadcast).

5. The programs were partnered up to allow for pair-wise comparisons be-tween shows of the same genre (except for Tatort and Schlag den Raab).

According to these criteria the following programs, presented in Table 1, wereselected. Sports were excluded from the sample because it is unclear whethercommunication can be identified as a Social TV activity or as communicationevoked by the live experience or other media.

Deutschland sucht den Superstar is the German equivalent of the Britishprogram Pop Idol or the American program American Idol. The Voice of Germanyis based on the international reality-singing competition The Voice. Ich bin einStar, holt mich hier raus is based on the British program I’m a Celebrity, Get MeOut of Here. While it is not a typical talent show, viewers are asked to vote forcandidates. As a TV event, Schlag den Raab (Beat the Star) was chosen. It is atypical live event where a candidate challenges the host Stefan Raab in quizzesand sport games. Tatort is a crime series that has been televised in Germanyregularly on Sunday evenings since 1970. Because of its character as a regularTV event and its role in the history of German television along with its consist-ently high ratings, it was selected for this investigation. However, for both ofthese latter shows, no match for comparison could be found. For the categorytalk shows, we selected the popular political debates of Maybritt Illner andGünter Jauch.

Page 9: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 137

Analysis of the communication showed that none of these shows – expectfor The Voice – had a Social TV strategy at the time of investigation. Germanbroadcasters did not stimulate online discussion via, for example, referring totwitter hashtags within the program or creating their own apps. Compared tothe US or other European countries, Social TV strategies are still underdevel-oped in Germany (Buschow, Schneider, Carstensen, Heuer, and Schoft, 2013).

4.2 Variables and operationalization

For the investigation of what communication activities occurred on Twitter dur-ing the broadcast of TV programs, we used categories identified in earlier stud-ies. Table 2 presents a summary of the investigated communication activitiesas well as their operationalization, a textual example, and the reliability of theconstructs.

The category Evaluation was operationalized by four different variables(positive/negative, show/actor) representing the findings from the literature. Toidentify Emotions within a tweet they had to be explicitly expressed. For exam-ple, “joy” was present when a tweet contained a phrase like “I’m happy” or“I’m glad” (“I’m happily looking forward to Florence! @TheVoiceGermany #Voiceit’s gonna be awesome!”). Abstraction was – as proposed in the literature –differentiated into two forms: “personal abstraction” and “societal abstraction”.Personal abstraction occurred when the users connected topics from the pro-gram to their personal life or personal experiences, for example: “I could easilylive without chancellor Merkel. #illner”. Societal abstraction takes place when atopic is deemed to generally refer to society as a whole, for example: “A prob-lem with politics is that you don’t make enough money. Without bribery! #illner”.Connecting with the community summarized four types of interaction betweenthe members of the Twitter community. (1) “Reference” means explicit address-ing a specific user. Such an explicit address can be identified by @ and thename of another user (“@_holger I didn’t say she is bad”). If this combinationappeared within a tweet, a reference was identified. (2) A “retweet” is a Twitter-specific way to quote another user. Retweets were identified by the prefix “RT”.(3) “Contact” refers to addressing Twitter users in general, and not specificusers. This is the case when the tweet uses “you” as a general pronoun, orwhen there is a question without a @-specification, for example: “#sdr the lastgame! Who’s going to win?!” (4) A “call to action” occurs in game and talentshows where viewers are an important part of the show because they are askedto vote for candidates. Such calls can be identified by imperatives like “vote”/“call [for XYZ]” (“Vote for Ivy!! #voice”). We also addressed some stylistic means

Page 10: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON138 C. Buschow, B. Schneider and S. Ueberheide

Tabl

e2:

Oper

atio

naliz

atio

nan

dda

tam

anag

emen

t.

Cate

gory

Varia

ble

nam

eTe

xtua

lexa

mpl

eRe

liabi

lity*

Eval

uatio

nNe

gativ

eev

alua

tion

ofth

esh

ow“d

isab

ility

asa

puni

shm

ento

fGod

...M

edie

valT

Vin

2012

..7

7#f

acep

alm

#Jau

ch”

Posi

tive

eval

uatio

nof

the

show

“the

Fran

kfur

t#Ta

tort

Ilik

em

ost”

Nega

tive

eval

uatio

nof

anac

tor

“Hol

geri

n#T

ator

tis

apo

orfe

llow

”Po

sitiv

eev

alua

tion

ofan

acto

r“S

tefa

nRa

abis

just

awes

ome!

#sdr

”Em

otio

nsJo

y“I

’mha

ppily

look

ing

forw

ard

toFl

oren

ce!@

TheV

oice

Germ

any

.96

#Voi

ceit’

sgo

nna

beaw

esom

e!”

Ange

r“A

t#SD

RIa

lway

sge

tang

ry!R

aab

can

neve

rkee

pto

the

rule

san

dm

ustd

iscu

ssab

oute

very

thin

g”Fe

ar“#

Tato

rtto

day

scar

esm

e.So

dark

...”

Sadn

ess

“#ib

esso

sad

that

mic

aela

isou

tD’:”

Disg

ust

“The

hook

erw

asso

disg

ustin

g!#T

ator

t”Pi

ty“I

just

hope

that

Luca

will

notc

ryif

Dani

ele

win

s,Ic

anno

tsee

him

cry

agai

n:’(

Ican

nots

tand

it...

#dsd

s”Ab

stra

ctio

nPe

rson

alab

stra

ctio

n“I

coul

dea

sily

live

with

outc

hanc

ello

rMer

kel.

#illn

er”

.90

Soci

etal

abst

ract

ion

“Apr

oble

mw

ithpo

litic

sis

that

you

don’

tmak

een

ough

mon

ey.

With

outb

riber

y!#i

llner

”Co

nnec

ting

Refe

renc

e“@

[…]I

don’

tsay

she

isba

d#i

bes”

.9w

ithth

eRe

twee

t**

“RT

@[…

]The

#FDP

will

chan

geno

thin

g#I

llner

”co

mm

unity

Cont

act

“#sd

rthe

last

gam

e!W

ho’s

goin

gto

win

?!”

Call

toac

tion

“Vot

efo

rIvy

!!#v

oice

”St

ylis

ticSw

earw

ords

“Boo

ooah

F*CK

#dsd

s”.9

3m

eans

Spam

**“G

o,vi

sitm

yw

ebsi

te[..

.]#d

sds”

Typi

cala

bbre

viat

ions

onth

ein

tern

et**

“#vo

ice

:-)#x

avie

rnai

doo

<3<3

<3”

Rest

ricte

dsp

eech

code

**“A

in’t

nobo

dygo

into

f*ck

wit

raab

#sdr

Sour

ce:o

wn

rese

arch

;*Ho

lsti’

sco

effic

ient

(Hol

sti,

1969

)**n

otin

clud

edin

clus

tera

naly

sis.

Page 11: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 139

such as swear words, spam, specific abbreviations and emoticons, and othersirrelevant to this investigation. All variables used in the study were binary cod-ed with “1 = present” and “0 = not present”.

4.3 Data management: Cluster analysis

31,388 cases were gathered during the inquiry period in accordance with theapplicability criteria. To reduce the complexity within such a volume of caseswe used a cluster analysis to identify groupings of tweets in data. Cluster analy-sis allows cases to be typified according to certain variables, resulting in inter-nally homogeneous groups that are as dissimilar as possible from each other(Everitt, Landau, Leese, and Stahl, 2011; Mooi and Sarstedt, 2011). Most of thevariables described above (see Table 2), except for some of the stylistic aspects,were used in the cluster analysis. Pearson’s correlation coefficient was calculat-ed for each variable pair to check for a correlation. The highest value of Pear-son’s correlation coefficient between a pair was approximately r = .2. Thus,every variable was considered in the cluster analysis.

A hierarchical cluster analysis with the nearest-neighbor method was real-ized to identify any unexpected spikes (Everitt et al., 2011). The distances weremeasured by the squared Euclidean distance. As no spikes were identified,every case was considered in the cluster analysis. Subsequently, a hierarchicalcluster analysis using Ward’s method was conducted to identify a reasonablenumber of clusters (Everitt et al., 2011). The distances were measured by thesquared Euclidean distance. The number of clusters was determined by thevariance criterion. An erratic gain of the sum of squares at a certain step ofclustering leads to a reasonable number of clusters. A scree plot helped tovisualize this gain, indicating that a solution with either three or seven clusterswas reasonable. For each of the solutions a k-means cluster analysis was gener-ated, allowing a comparison of the interpretability. Finally, a solution with sev-en clusters was chosen.

5 Results

[RQ1] Communication activities and their specific relevance (SRQ 1–4)

The first research question was answered by descriptive statistics as well ascluster analysis. The dataset included the cases of all TV programs investigated.

Page 12: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON140 C. Buschow, B. Schneider and S. Ueberheide

Table 3: Relevance of specific communication activities by category.

SRQ Category %

1 Evaluation of shows and actors 24*Positive about show 3Positive about actor 9Negative about show 6Negative about actor 7

2 Emotions 4*Joy 1Anger 1Fear .2Sadness .3Disgust .3Pity .4

3 Abstraction 5*Personal 4Societal 1

4 Connecting with the community 49*Reference 31Contact 15Call to action 3

Source: own research; N = 31,388; *differences due to rounding errors; multiple categoriesper tweet possible.

Table 3 shows the distribution of communication activities differentiated bycategory.

The data shows that 24 % of all tweets are from the category “Evaluationof shows and actors” (SRQ1). In the case of actors we see a ratio of 56 % positiveto 44 % negative ratings. Programs were more often criticized than actors (67 %negative ratings). The “Emotion” category (SRQ2) was a very rare phenomenon.Only 4 % of all tweets contained explicit expressions of emotions. “Abstraction”(SRQ3) was also relatively infrequent – only 5 % of all tweets contained person-al or societal abstraction. 31 % of the tweets contained a reference to anotheruser, while 15 % sought to establish contact with the entire community and 3 %called for the community to act in a certain way. In sum, 49 % of all tweetswere categorized as “Connecting with the community” (SRQ4).

Furthermore, the data challenges some popular clichés about Twitter oftenarticulated in the public discourse. Firstly, the debates on Twitter are not regu-larly ‘vulgar’. Only 5 % of all analyzed tweets contained swear words. We col-lected a very marginal amount of spam: Only 0.4 % of the tweets could beidentified as such. The level of language is higher than expected. Only 11 % of

Page 13: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 141

Table 4: Description and size of the clusters.

Cluster no.* Cluster N %

1 Debate about the program 14,142 45.12 Fandom 1,686 5.43 Critique 1,548 4.94 Insult 1,143 3.65 Retweet 8,294 26.46 Contacting the community 3,088 9.87 Discourse 1,487 4.7

31,388 99.9*

Source: own research; N = 31,388; *differences due to rounding errors.

tweets could be identified as employing ‘restricted speech code’. The identifica-tion depended on the correct use of grammar and the diction of the tweet.

Overall, the tweets seem to be motivated by two desires: to interact withthe community and/or to engage with the program. The interpretation of thecluster analysis confirms this assumption and allows a more detailed look atthe patterns within our sample. Table 4 lists the clusters and the number oftweets assigned to each one.

The first cluster accounts for 45 % of all tweets. It is characterized by posi-tive as well as negative evaluations of actors and shows, and is hence referredto as “Debate about the program”. Cluster 2 is characterized by positive evalu-ations of actors and is designated as “Fandom”. Clusters 3 and 4 constitute theopposite: Both are dominated by negative evaluations of actors. While “Insults”(Cluster 4) is characterized by swear words, “Critiques” (Cluster 3) is more well-mannered. All of these clusters (representing 59 % of all tweets) primarily focuson the TV content itself.

Clusters 5 to 7 mainly relate to interactions between users. The fifth clusteris the second largest, accounting for 26.4 % of all tweets. It is dominated byretweets and also includes communication that addresses specific users; wehave designated it “Retweets”. Cluster 6 comprises communication addressingthe whole community, for example, calls to action; we refer to it as “Contactingthe community”. Cluster 7 is characterized by direct communication betweencertain users (@-tag) and by communication that addresses the entire commu-nity, and is referred to as “Discourse”.

Page 14: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON142 C. Buschow, B. Schneider and S. Ueberheide

[RQ2] Influence of TV programs on communication activities

Investigating the second research question entails highlighting the significanceof TV program-related tweets that are assigned to a specific cluster. Table 5shows the TV programs by row and the clusters by column. The standardizedresiduals presented here are an expression of the comparison of the empiricalcount with the expected count and provide information about the significanceof differences (Bakeman and Robinson, 1994). The bold numbers show whetherthe number of tweets relating to a certain TV program within a cluster is signifi-cantly high (+) or low (−).

The results show that different types of TV programs evoke different kindsof tweets. Moreover, there are also differences between TV shows within a sin-gle genre. Every TV show has its own specific structure of tweets. The reality-singing competition The Voice of Germany mainly produces tweets that arerelated to the show’s candidates. “Fandom” – meaning positive ratings of ac-tors – is an important issue within the communication about The Voice. How-ever, there are also a larger number of tweets which can be considered ascritique toward the show’s actors. The regular appearance of actors in the showleads to a strong base of fans that cheer for their candidate and criticize com-peting actors. However, the program itself is a less popular topic.

Significantly, Deutschland sucht den Superstar (Pop Idol) often evoked in-sulting tweets and “Retweets” although it is of the same genre as The Voice.Overall, swear words appeared rarely (5 % of all tweets; for an example tweet,see Table 2), but a significant portion of these tweets are related to this showand its actors. In comparison to The Voice, the tweets are much more polemicor even insulting. It would seem that Deutschland sucht den Superstar has beenunable to generate a stable fan base that cheers for certain actors.

Tweets related to Ich bin ein Star, holt mich hier raus (I’m a Celebrity, GetMe Out of Here) mostly deal with the program itself and comprise critiques aswell as expressions of “Fandom”. The significantly high amount of tweets in“Contacting the community” reflects the viewers’ special relationship to theactors competing on the program: The community is mobilized when it is timeto vote and is requested to vote for a certain candidate.

Schlag den Raab (Beat the Star), the live game show, evokes tweets thatinclude debates about the show, comprise critiques, and create a discoursewithin the community. The number of tweets in “Fandom” is significantly low,which may be explained by the fact that the candidates only compete in asingle episode. This inhibits the proper development of a relationship betweenviewers and candidates; rather, the actual competitions and games on the showare in focus.

Page 15: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 143

Tabl

e5:

Stan

dard

ized

resi

dual

sof

clus

tera

ssig

nmen

tsof

TVpr

ogra

m-re

late

dtw

eets

.

TVpr

ogra

mCl

uste

r

1:2:

3:4:

5:6:

7:De

batin

gFa

ndom

Criti

que

Insu

ltRe

twee

tCo

ntac

ting

Disc

ours

eth

epr

ogra

m(N

=1,

686)

(N=

1,54

8)(N

=1,

143)

(N=

8,29

4)co

mm

unity

(N=

1,48

7)(N

=14

,142

)(N

=3,

088)

The

Voic

eof

Germ

any

−7.1

16.9

5−2

2.8

−4.4

0.2

Deut

schl

and

such

tden

−4.8

1.8

−7.1

4.2

8.1

−2.7

1,3

Supe

rsta

r(P

opId

ol)

Ich

bin

ein

Star

,hol

t2.

89.

99.

6−0

.2−1

4.8

8.2

−5.6

mic

hhi

erra

us!

(I’m

aCe

lebr

ity,

GetM

eOu

tofH

ere)

Schl

agde

nRa

ab6.

2−1

3.2

2.5

−0.2

−4.7

0.6

2.7

(Bea

tthe

Star

)Gü

nter

Jauc

h−0

.3−1

.8−0

.5−1

.36.

4−4

.6−3

.8(P

oliti

calt

alk)

May

britt

Illne

r−4

.7−3

.4−0

.1−3

.58.

6−1

2.5

(Pol

itica

ltal

k)Ta

tort

8−5

.6−7

.7−0

.7−5

.95.

4−4

.1(C

rime

serie

s)

Sour

ce:o

wn

rese

arch

;sta

ndar

dize

dre

sidu

als:

valu

es≥

|2|i

ndic

ate

asi

gnifi

cant

diffe

renc

eon

a5

%le

vel(

p<

.05)

,val

ues

≥|2

.6|o

na

1%

leve

l(p

<.0

1)(B

akem

anan

dRo

bins

on,1

994,

p.22

–23)

.Diff

eren

ces

are

mar

ked

inbo

ld.

Page 16: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON144 C. Buschow, B. Schneider and S. Ueberheide

Tweets about the political talk shows Maybritt Illner and Günter Jauch areoften “Retweets”. Those related to Maybritt Illner also partly comprise a dis-course. Users generalize the topics of the political talk show and reflect themon a societal level. In this case it is interesting to see that even programs withthe same concept evoke different kinds of tweets

The crime series Tatort evokes a debate about the plot which can be charac-terized mainly as a quality discourse: Users typically discuss the story, cast,and production value of the current episode.

6 Conclusion and outlookThe main goal of this study was to explore the communication activities occur-ring on Twitter while people watch TV simultaneously. The results show howTwitter, as a relatively new communication technology, as well as different TVprograms encourage specific (and new) types of communication. The two keyfindings of our study are:1. Social TV on Twitter comprises a communicative exchange about linear

television content. Nevertheless, Social TV also relates to communicationwithin the community which is not necessarily textually connected to thecontent.

2. We explored those TV programs for which former research expected thebiggest relevance for Social TV. We can now state which types of communi-cation these programs evoke – at least in Germany.

The following discussion focuses primarily on these aspects.

6.1 Communication activities relating to TV content – andbeyond

The data explore the communicative use of Twitter for Social TV in Germany.The finding that evaluations play a major role in the data is in line with as-sumptions from the literature (Hepp, 1998; Klemm, 2000; Weisz et al., 2007;Wohn and Na, 2011). Based on earlier studies, we expected evaluations to ac-count for around one-third of all activities. The investigation confirmed thisassumption. However, emotions are less important than predicted in the litera-ture (Godlewski and Perse, 2010; Weisz et al., 2007). The limited relevance ofemotions need not imply that German Twitter users are less emotional. Weiszet al.’s (2007) study that classified 9 % of communication activities as “joy”

Page 17: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 145

used a cartoon as stimulus. Our study did not include such humorous material.The “Abstraction” category was rare and very much dependent on specific gen-res (see below).

The most striking finding seems to be the large number of community-oriented tweets (clusters 5 to 7). In one-third of all communication activities wesee some interaction with specific individuals. Today, these people also includestrangers and not only family members and friends, as was the case prior tothe advent of the social media technologies. As suggested by Klemm’s (2000)observations of families, joint TV reception stimulates interpersonal communi-cation beyond the program’s broadcast. A large number of activities also consti-tute retweets of other users’ posts along with recommendations to spread them.However, this practice also highlights the self-referentiality of tweeting.

6.2 Influence of specific TV content

One of the few consistencies between researchers in this field is the assumptionthat reality shows are suited for Social TV (Duchenaut et al., 2008; EricssonConsumerLab, 2011; Geerts et al., 2008). Our investigation confirms this hypoth-esis: The Voice of Germany, Deutschland sucht den Superstar (Pop Idol), and Ichbin ein Star, holt mich hier raus (I’m a Celebrity, Get Me Out of Here) stimulatecommunication that is particularly characterized by an evaluation of the partic-ipants on the show – positive as well as negative. Candidates polarize the com-munity and arouse expressions of fandom as well as critiques. Upon closerinspection, a special kind of interaction plays an important role: In the lastminutes of the show the number of tweets calling for users to take action in-creases significantly. This is caused by users who call on other users to votefor their favorite candidate.

From the literature we have seen that there is also some consensus thatdebates and talk shows are appropriate for Social TV. The political talk showsMaybritt Illner and Günter Jauch received less attention than reality shows.However, the tweets referring to them focused more on interaction than onevaluation of the show and its participants. The users refer to each other andstart a discourse. Looking at the objects of this discourse it is obvious that usersabstract topics from the show and elevate them to a higher societal level. Withrespect to political talk shows, we see a potential for activating debates and a“critical discourse” (Habermas, 1989) accompanying the broadcast. However,closer inspection reveals that there are indeed differences between both politi-cal talk shows. Even though the show’s concepts are very similar, the influenceof the discussed topic as well as the show’s host should not be underestimated.

Page 18: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON146 C. Buschow, B. Schneider and S. Ueberheide

Tatort (crime series) stimulated a debate within the community in relationto the quality of the episode or story or the production value. As a crime series(Duchenaut et al., 2008) and public broadcast TV program with a momentoushistory, viewers seem to have certain expectations and feelings relating to theseries. Therefore, for Social TV in Germany Tatort has specific potential, al-though discussions are not always positive. Similar to Tatort, tweets on Schlagden Raab (Beat the Star) debated the show and represent a discourse withinthe community. In contrast to Tatort, the discourse here is not primarily aboutquality. Communication activities focus on the games. From the data it is clearthat the rate of tweets per minute rises significantly during the games comparedto the rest of the broadcast.

6.3 Limitations of the study

If Social TV is part of the future of traditional linear television, as suggestedby practitioners and researchers, further studies are needed to understand thephenomenon more precisely.

Our results have been exclusively obtained from consciously selected TVprograms. These programs are not necessarily representative of all German TVprograms or of international formats. Data has only been collected from Twitter.However, we also see a lively exchange about TV on other social networkingsites such as Facebook. In our study we excluded communication activitiesoccurring before and after broadcast. There is however some research – forexample by Godlewski and Perse (2010) – that would help to examine pre- andfollow-up communication in further investigations.

On the one hand, more research focusing on international comparisonsand/or the comparison of different technologies/infrastructure is needed. Com-paring face-to-face communication in one location and communication trans-ferred by technology could also be an interesting perspective. Such researchwould additionally stimulate some viral cooperation of intergroup communica-tion research and communication technology research. On the other hand,(qualitative) case study research should investigate certain shows in more de-tail. Against this background, data from this study could be juxtaposed withviewers’ social demographic data. This would allow for a more detailed inter-pretation of the communication activities of the shows’ respective audiences.

Furthermore, we would advocate survey research: Questionnaires of ‘heavyusers’ or non-users could explore modes of usage as well as motivations for‘tweeting television’.

Page 19: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 147

ReferencesAndrejevic, M. 2008. Watching television without pity: The productivity of online fans.

Television & New Media (TVNM), 1(9), 24–46.Baillie, L., Fröhlich, P., & Schatz, R. 2007. Exploring social TV. In ITI (Eds.), 29th International

Conference on Information Technology Interfaces, 25–28 June 2007 (pp. 215–220).Cavtat. doi:10.1109/ITI.2007.4283773

Bakeman, R. & Robinson, B. F. 1994. Understanding log-linear analysis with ilog: Aninteractive approach. New Jersey: Lawrence Erlbaum.

Barker, C. 1997. Television and the reflexive project of the self: Soaps, teenage talk andhybrid identities. The British Journal of Sociology, 48(4), 611–628.

Bergman, C. 2011a. Why social TV will be a multi-billion dollar business, June 27. RetrievedJanuary 15, 2013 from http://lostremote.com/why-social-tv-will-be-a-multi-billion-dollar-business_b19841.

Bergman, C. 2011b. TVGuide debuts revamped ‘Watchlist’ with social ties, August 2.Retrieved January 15, 2013 from http://lostremote.com/tvguide-debuts-revamped-watchlist-with-social-ties_b20734.

BITKOM (Eds.) 2012. Soziale Netzwerke erweitern die Fernsehwelt [Social networks broadenthe world of television]. Retrieved January 15, 2013 from http://www.bitkom.org/73785_73776.aspx.

Bulkeley, W. M. 2010. Relying on relationships to rebuild TV audiences – Social TV.Technology Review, (5/6), 55–56.

Buschow, C., Schneider, B., Carstensen, L., Heuer, M., & Schoft, A. 2013. Social TV inDeutschland – Rettet soziale Interaktion das lineare Fernsehen? [Social TV inGermany – Is linear television saved by social interaction?] MedienWirtschaft, 10(1),24–32.

Cesar, P., & Chorianopoulos, K. 2008. Interactivity and user participation in the televisionlifecycle: Creating, sharing, and controlling content. In UXTV ’08 (Eds.), Proceedings ofthe 1st international conference on designing interactive user experiences for TV andvideo (pp. 125–128). New York: ACM. doi:10.1145/1453805.1453830

Chorianopoulos, K., & Lekakos, G. 2008. Introduction to social TV: Enhancing the sharedexperience with interactive TV. International Journal of Human-Computer Interaction,24(2), 113–120. doi:10.1080/10447310701821574

ComScore (Eds.) 2012. Unique visitors of Twitter.com in Germany. Retrieved January 15, 2013from http://de.statista.com/statistik/daten/studie/223174/umfrage/unique-visitors-von-twittercom-in-deutschland/.

Deller, R. 2011. Twittering on: Audience research and participation using Twitter.Participations: Journal of Audience & Reception Studies, 8(1), 216–245. RetrievedJanuary 15, 2013 from http://www.participations.org/Volume%208/Issue%201/PDF/deller.pdf.

Dewdney, A., & Ride, P. 2006. New Media Handbook. Abingdon, New York: Routledge.Ducheneaut, N., Moore, R. J., Oehlberg, L., Thornton, J. D., & Nickell, E. 2008. Social TV:

Designing for distributed, sociable television viewing. International Journal of Human-Computer Interaction, 24(2), 136–154. doi:10.1080/10447310701821426

Ericsson ConsumerLab (Eds.) 2011. TV & video 2011. Consumer trends. Global version.Retrieved January 15, 2013 from http://www.ericsson.com/res/docs/2011/11_1650_RevB_TV_Video_Consumer_Trends_2011_Global_Version.pdf.

Page 20: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON148 C. Buschow, B. Schneider and S. Ueberheide

Everitt, B. S., Landau, S., Leese, M., & Stahl, D. 2011). Cluster analysis (5th ed.). Chichesteret al.: John Wiley & Sons.

Franz, K. 2008. Kommunikative Aneignung von Fernsehserien und medialeKommunikationskultur: [email protected] [Communicative appropriation of TVsoaps and media-based communication culture: [email protected]]. Dissertation,University of Marburg. Retrieved January 15, 2013 from http://archiv.ub.uni-marburg.de/diss/z2008/0664/.

Geerts, D. 2009. Sociability heuristics for interactive TV. Supporting the social uses oftelevision. Dissertation, Katholieke Universiteit Leuven. Retrieved January 15, 2013 fromhttps://lirias.kuleuven.be/bitstream/1979/2607/5/Doctoraatsproefschrift_Geerts.pdf.

Geerts, D., Cesar, P., & Bulterman, D. 2008. The implications of program genres for thedesign of social television systems. In UXTV ’08 (Eds.), Proceedings of the 1st

international conference on designing interactive user experiences for TV and video (pp.71–80). New York: ACM. doi:10.1145/1453805.1453822

Godlewski, L. R., & Perse, E. M. 2010. Audience activity and reality television: Identification,online activity, and satisfaction. Communication Quarterly, 2(58), 148–169.doi:10.1080/01463371003773358

Goffman, E. 1978. Response cries. Language, 54(4), 787–815.Habermas, J. 1989. The structural transformation of the public sphere: An inquiry into a

category of bourgeois society. Chichester et al.: John Wiley & Sons.Hepp, A. 1998. Fernsehaneignung und Alltagsgespräche. Fernsehnutzung aus der

Perspektive der Cultural Studies [Television appropriation and everyday conversation.Television use from the perspective of Cultural Studies]. Wiesbaden: Westdeutscher.

HMR International (Eds.) 2011. More than TV – Handbuch für eine Branche, die sich neuerfindet [More than TV – Handbook for an industry which is reinventing itself]. Cologne.

Holly, W., Püschel, U., & Bergmann, J. (Eds.) 2001. Der sprechende Zuschauer. Wie wir unsFernsehen kommunikativ aneignen [The speaking audience. How we communicativelyappropriate television]. Wiesbaden: Westdeutscher.

Holsti, O. R. 1969. Content analysis for the social sciences and humanities. Reading:Addison-Wesley.

Klemm, M. 2000. Zuschauerkommunikation. Formen und Funktionen der alltäglichenkommunikativen Fernsehaneignung [Audience communication. Forms and functions ofeveryday communicative television appropriation]. Frankfurt am Main: Peter Lang.

Krippendorff, K. 2004. Content analysis: An introduction to its methodology. Thousand Oakset al.: Sage.

Lee, B., & Lee, R. S. 1995. How and why people watch TV: Implications for the future ofinteractive television. Journal of Advertising Research, 35(6) 9–18.

Matthewson, L. 1992. Talking to the TV. The conversational behavior of television viewers.Wellington Working Papers in Linguistics, (5), 17–35.

Mooi, E., & Sarstedt, M. 2011. A concise guide to market research. The process, data, andmethods using IBM SPSS Statistics. Heidelberg et al.: Springer.

Morley, D. 1986. Family television. Cultural power and domestic leisure. London: Comedia.Neuendorf, K. A. 2001. The content analysis guidebook. Thousand Oaks et al.: Sage.Nielsen (Eds.) 2011. The relationship between social media buzz and TV ratings, October 6.

Retrieved January 15, 2013 from http://blog.nielsen.com/nielsenwire/online_mobile/the-relationship-between-social-media-buzz-and-tv-ratings/.

Research and Markets (Eds.) 2010. Social TV and the emergence of interactive TV.

Page 21: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

DE GRUYTER MOUTON Tweeting television 149

Richardson, A. 2010. Young viewers lead the take-up of social TV as more go online to catchup. New Media Age, 15(10), 26.

Schatz, R., Baillie, L., Fröhlich, P., & Egger, S. 2008. Getting the couch potato to engage inconversation: Social TV in a converging media environment. In A. Lugmayr, S. Kemper,M. Obrist, T. Mirlacher & M. Tscheligi (Eds.), Adjunct proceedings of EuroITV 2008 (pp.25–26). Tampere: Tampere University of Technology.

Schmidt, H. 2012. Twitter durchbricht die 4-Millionen-Marke in Deutschland [Twitter hits thefour million mark in Germany]. FOCUS Online, April 22. Retrieved January 15, 2013 fromhttp://www.focus.de/digital/internet/netzoekonomie-blog/social-media-twitter-durchbricht-die-4-millionen-marke-in-deutschland_aid_740627.html.

Summa, G. 2011. Social TV: The future of television in the internet age. Unpublished masterthesis at the Massachusetts Institute of Technology (MIT), Sloan School ofManagement. Retrieved January 15, 2013 from http://dspace.mit.edu/handle/1721.1/65819.

TNS Infratest (Eds.) 2013. TNS Convergence Monitor 2013. Retrieved from: http://www.tns-infratest.com/presse/pdf/Presse/2013-08-27_TNS_Infratest_ConvergenceMonitor_Charts_Parallel-Nutzung.pdf [30. 10. 2013].

Weisz, J. D., Kiesler, S., Zhang, H., Ren, Y., Kraut, R. E., & Konstan, J. A. (2007). Watchingtogether: Integrating text chat with video. In CHI ’07 (Eds.), Proceedings of the SIGCHIconference on human factors in computing systems (pp. 877–886). New York: ACM.doi:10.1145/1240624.1240756

Westerik, H. 2009. The social embeddedness of media use. Action theoretical contributionsto the study of TV use in everyday life. Berlin, New York: De Gruyter Mouton.

Wohn, D. Y., & Na, E.-K. 2011. Tweeting about TV: Sharing television viewing experiences viasocial media message streams. First Monday [Online], 16(3). Retrieved January 15, 2013from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/3368/2779.

Zigmond, D. & Stipp, H. 2010. Assessing a new advertising effect: Measurement of theimpact of television commercials on internet search queries. Journal of AdvertisingResearch, 50(2), 162–168.

Page 22: Tweetingtelevision: Exploringcommunicationactivitieson ... · Studies show that viewers react emotionally to TV content (Godlewski and Perse,2010;WohnandNa,2011).Therefore,emotionsalsoplayamajorrolein

Copyright of Communications: The European Journal of Communication Research is theproperty of De Gruyter and its content may not be copied or emailed to multiple sites orposted to a listserv without the copyright holder's express written permission. However, usersmay print, download, or email articles for individual use.