Search Bias

Learn about how algorithms contain bias and affect search results even in trusted information sources.

Today  you  feel  ready  to  search. You  are  confident  that  you  know  what  it  means  when  a  source  is  Scholarly, That  is,  it's  found  in  a  peer  reviewed  journal  or  other  publication. It  has  a  list  of  references  and  it  offers  a  good  argument. You've  been  to  a  library  search  session, a  timer  to  know  that  the  best  way  to  start  searching  broadly  is  to  use  a  database. These  assumptions  are  all  correct, but  they're  not  the  end  of  the  story. Unfortunately,  databases  are  actively  hiding important  materials  from  you  using  something  called  algorithmic  bias. But  there  are  tools  and  strategies  you  can  use  to  make  sure you  combat  this  unintentional  but  pervasive  problem. What  is  algorithmic  bias? In  his  book  masked  by  trust  bias  and  library  discovery, Matthew  reads  my  explains  that  algorithms  are examples  of  pattern  recognition  that  show  up  when  many, many  searches  are  run  in  a  database  or  search  engine. Algorithms  are  deeply  woven  into  most  aspects  of  our  lives. From  Netflix  suggesting  what  you'd  like  to  watch  next, to  Facebook  promoting  posts  that  are  more  prone  to  listening  and  emotional  response. Yet  these  algorithms  promoted  as  curating perfectly  customised  recommendations  are  thoroughly  human. Software  engineers  constantly  update  their  programs  to  improve  performance. Sometimes  experiments  are  run  without  users  knowing. Remember  when  Facebook  prone  and  upsetting  material  to  one  group  of  users  and uplifting  material  to  others  to  gauge  which  type  of  content  promoted  the  most  clicks. And  if  it's  really  totally  customized, Can  you  and  I  even  say  we  use  the  same  algorithm  will  Googling  best  restaurant  and  Lawrence. But  you  might  argue  databases  claim  that  they're  unbiased, influenced  only  by  the  search  terms  and  other  parameters  that  you  select. And  it  is  true  that  library  databases  are  free  of certain  kinds  of  external  influences  like  advertising, the  harvesting  of  all  your  browsing  data  to  cure  a  resources. The  lower  of  designing  material  to  show  up  first  on  a  search  engine, otherwise  known  as  SEO. But  the  tools  you  use  to  access  materials  on  campus  is  selected, arranged,  catalogued,  and  made  accessible  by  the  same  folks  who  designed  algorithms. Human  beings,  complete  with  all  their  internal  biases. The  use  of  keywords  as  neutral  is  highly  contested  by  librarians. Consider  the  constant  review  and  revision  of offensive  keywords  that  is  undertaken  regularly  by  the  Library  of  Congress. In  addition,  as  reads  my  notes, they  assumed  that  broad  topics  can  be  easily  identified  by a  few  keywords  and  reference  material  is  always  reliable  knowledge  source. Aside  from  the  problems  of  assuming  the  material  gathered  is  neutral. What  about  the  nature  of  lived  experience  and  the  information  that  grants? What  about  traditional  knowledges? What  about  oral  histories,  scenes,  blogs, all  kinds  of  places  where  knowledge  making  occurs  outside  the  academic  framework. Of  course,  knowing  these  pitfalls  doesn't  mean  that  you can  or  should  discard  these  resources. But  it's  good  to  be  aware  that  a  database  search does  not  provide  all  the  information  that  there  is. Ask  yourself  when  searching. Who  is  the  audience, who  benefits  from  this  information  being  selected  to  promote  who  is  silenced? Now  that  you've  answered  those  questions, consider  those  authors  and  keywords. Try  and  make  sure  you're  looking  for  underrepresented  authors  and  perspectives. Use  multiple  databases,  really  large  databases  like  Academic  Search  Premier, or  ones  that  contain  a  vast  amount  of  information. But  other  databases  are  available  that  offer  a  small  but  extremely  thorough  collection. In  particular,  checkout  ones  that  offer  alternative  perspectives  on  a  topic. Read  more  about  algorithmic  bias, along  with  masked  by  trust  checkout, algorithms  of  oppression  by  Sophia  Noble, Race  After  Technology  by  rehab  Benjamin, and  the  intersectional  Internet  edited  by  Sophia  Noble  and  Brenda  showtimes. Most  importantly,  stay  alert  and  stay  critical  when  searching. Never  trust  that  the  perfect  resource  exist. Use  the  same  skills  you've  cultivated  and  evaluating  resources to  consider  the  biases  and  strengths  of  how  these  resources  get  it  to  you. Talk  about  it  with  your  professor  and  other  students. Know  that  a  librarian  is  always  ready  and  excited  to  talk  more  with  you  about  it.