Automatic web accessibility metrics: Where we are and where we can go

 The  fact  that  several  web  accessibility  metrics  exist  may  be  evidence  of  a  lack  of  a  comparison  framework   that  highlights  how  well  they  work  and  for  what  purposes  they  are  appropriate.  In  this  paper  we  aim  at  formulating   such  a  framework,  demonstrating  that  it  is  feasible,  and  showing  the  findings  we  obtained  when  we  applied  it  to   seven  existing  automatic  accessibility  metrics.  The  framework  encompasses  validity,  reliability,  sensitivity,  adequacy   and   complexity   of   metrics   in   the   context   of   four   scenarios   where   the   metrics   can   be   used.   The   experimental   demonstration  of  the  viability  of  the  framework  is  based  on  applying  seven  published  metrics  to  more  than  1500  web   pages  and  then  operationalizing  the  notions  of  validity-­as-­conformance,...

• Type: Journal
• Year: 2011
• Where: IWC
• Authors: Markel Vigo, Giorgio Brajnik
• Keywords: IWC 2011, Neural Networks, Page Measure, Research Avenues, Validity Reliability