• 沒有找到結果。

Hardaker’s  Model  of  Trolls

Chapter  3   -­‐  Methodology

3.1   Hardaker’s  Model  of  Trolls

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

YouTube  users  incited  PewDiePie’s  fan  community  by  posting  insults  towards  PewDiePie   and  his  fans  such  as  “PewDiePie  sucks,”  or  “PewDiePie’s  fans  are  all  8-­‐year-­‐olds.”  

There  were  many  of  these  negative  and  hateful  comments  but  I  excluded  comments   which  attracted  little  to  no  respondents.  Instead,  I  selected  comments  which  contained  an   ongoing  conversation  chain  of  at  least  50  or  more  responses  for  two  reasons:  1)  I  could   weed  out  “haters”  who  posts  negative  messages  but  do  not  respond  to  anyone  who  has   replied  to  their  original  comment  and  2)  to  identify  successful  trolls  who  are  committed  to   and  have  succeeded  in  engaging  their  victims  for  longer  durations.  In  these  comparatively   lengthy  conversations,  the  suspected  troll  must  continue  to  interact  with  his  so-­‐called   victims,  extending  the  spiral  of  disruption  and  pursuit  of  lulz.    

Other  red  flags,  which  indicated  the  presence  of  a  troll,  included  hostile  behaviors   such  as  ‘flaming’  and  ‘hating,’  as  described  in  chapter  2.7;  this  may  include  insults,  negative   and/or  vulgar  language,  and  strings  of  capitalized  text.    

Furthermore,  I  was  the  only  coder  working  on  all  analysis  for  this  study  so  there  was   no  need  for  inter-­‐reliability  tests  and  all  results  are  consistent  in  that  they  were  analyzed  by   one  person.  There  are,  however,  issues  that  may  arise  from  single-­‐researcher  analysis  as   explained  further  in  the  limitations  section.  

 

3.1  Hardaker’s  Model  of  Trolls  

I  have  chosen  to  analyze  the  interactions  of  YouTube  trolls  with  other  users  

according  to  Hardaker’s  (2010)  characteristics  of  trolls  in  asynchronous  computer-­‐mediated-­‐

communication  (CMC).  Hardaker’s  study  originally  sought  to  redress  the  lack  of  a  proper   academic  framework  to  describe  the  phenomenon  of  trolling.  As  a  corpus  which  analyzed   data  from  several  year’s  worth  of  unmoderated  Usenet  newsgroup  posts,  Hardaker’s  study  

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

looked  primarily  at  how  trolls  were  being  described  by  community  members,  sometimes   including  the  trolls  themselves,  and  ultimately  identified  four  recurrent  themes  regarding   trolls/trolling:  deception,  aggression,  disruption,  and  success.  

Given  Hardaker’s  classification  of  Usenet’s  user  interaction  system  as  asynchronous   CMC,  it  is  reasonable  to  infer  that  her  findings  are  also  applicable  to  other  online  forms  of   asynchronous  CMC  such  as  YouTube’s  commenting  system.  YouTube  users  interact  through   computer-­‐mediated  commenting  areas  provided  under  each  YouTube  video  and  

asynchronously  by  posting  messages  to  one  another  whenever  they  are  online  rather  than   in  real-­‐time.  Therefore,  responses  may  be  posted  within  hours  and  days  to  months  or  even   years  in  between.    

Looking  more  in-­‐depth  at  Hardaker’s  characteristics  of  trolls,  ‘deception’  refers  to   users  who  attempt  to  pass  themselves  off  as  sincere  community  members  instead  of  overtly   identifying  themselves  as  a  malicious  troll.  ‘Aggression’  is  described  by  Hardaker  as  

“aggressive,  malicious  behaviour  undertaken  with  the  aim  of  annoying  or  goading  others   into  retaliating,”  while  ‘disruption’  includes  causing  aggravation  without  necessarily   attacking  specific  individuals.  Finally,  ‘success’  entails  trolls  being  appraised  by  users  

regarding  a  troll’s  quality  and  how  others  respond  to  the  troll.  I  posit  that  ‘success’  is  tied  to   the  notion  of  lulz,  which  remains  the  primary  motivator  for  trolls  and  the  entire  basis  for   troll  culture.  I  hypothesize  that  trolls  themselves  may  also  speak  about  success  or  defeat  in   their  communication  with  others.  

I  intend  to  test  Hardaker’s  model  of  trollish  characteristics,  aiming  to  seek  if  and  how   these  four  characteristics  manifest  themselves  in  the  interactions  of  YouTube  trolls  and   other  users.  Acknowledging  that  Hardaker’s  model  was  created  from  ground  level  research,   and  may  not  be  fully  compatible  with  study  on  YouTube  trolls  as  she  analyzed  UseNet  trolls,  

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

I  will  also  attempt  to  account  for  any  differences  and  identify  any  additional  characteristics   prevalent  within  YouTube  trolls  that  were  not  mentioned  in  Hardaker’s  study.  For  example,   Sierra’s  (2014)  notion  of  “Kool-­‐Aid  Points”  for  trolls  who  attack  others  because  they  feel   their  attention  and  status  is  unwarranted,  might  be  a  motivator  for  trolls  to  target   PewDiePie’s  page.  Thus,  I  propose  the  following  research  questions:  

 

RQ1.  Do  YouTube  trolls  exhibit  the  characteristics  of  'deception,'  'aggression,'   'disruption'  and  'success'?  

RQ2.  Are  there  any  other  characteristics  that  YouTube  trolls  exhibit?  

RQ3.  If  so,  how  do  trolls  exhibit  these  characteristics?  

 

There  are  a  few  points  of  contention  that  require  clarification  regarding  the  use  of   Hardaker’s  model.  Hardaker’s  formulation  of  trollish  characteristics  is  based  on  analysis  of   the  Usenet  community.  Although  her  model  of  “trolls  in  asynchronous  CMC”  may  be  a   reasonable  starting  point  for  analyzing  YouTube  trolls,  there  are  bound  to  be  certain   idiosyncratic  differences  between  YouTube  trolls  and  Usenet  trolls.  Thus  it  is  reasonable  to   expect  areas  in  which  this  study’s  findings  diverge  and/or  contribute  to  Hardaker’s  model.  

Additionally,  Hardaker  evaluated  descriptions  of  trolls  within  the  Usenet  community  in  order   to  construct  a  model  of  trollish  attributes,  whereas  I  aim  to  utilize  her  model  to  directly   analyze  trolls  and  their  comments.  While  it  is  true  that  I  may  not  be  able  to  completely   isolate  trollish  comments  from  other  users’  comments,  this  study  places  higher  emphasis  on   directly  examining  characteristics  of  trolls  as  exhibited  by  the  trolls  themselves,  more  so   than  indirectly  examining  descriptions  of  trolls  by  other  community  members.  

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

Additionally,  there  might  be  a  few  hidden  themes  regarding  the  nature  of  trolls,  as   mentioned  in  previous  sections.  Existing  literature  has  theorized  that  many  trolls  are  indeed   white  young  males  who  many  times  seek  to  direct  their  trollish  behaviors  towards  women,   minority  groups,  and  homosexuals  (A.  Chen,  2014;  Herring  et  al.,  2002;  Milner,  2013;  Phillips,   2011,  2013).  Ironically,  “maleness  and  whiteness”  have  also  traditionally  been  used  to  

describe  gamer  identity  (A.  Shaw,  2010)  which  belongs  to  the  very  culture  which  PewDiePie   espouses  through  his  videos.  I  will  attempt  to  confirm  or  disconfirm  if  such  themes  do  rise   up  during  my  analysis  of  YouTube  trolls:  

RQ4.  What  are  the  social  implications  of  YouTube  trolls?  

 

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y