How can we improve learning

how classroom assessments improve learning thomas guskey and how to improve collaborative learning with video tools in the classroom
JaydenGibbs Profile Pic
JaydenGibbs,United States,Teacher
Published Date:19-07-2017
Your Website URL(Optional)
Comment
VideoDoc:  Combining  Videos  and  Lecture  Notes   for  a  Better  Learning  Experience     by     Rebecca  P.  Krosnick     S.B.,  Massachusetts  Institute  of  Technology,  2014     Submitted  to  the   Department  of  Electrical  Engineering  and  Computer  Science   in  Partial  Fulfillment  of  the  Requirements  for  the  Degree  of       Master  of  Engineering  in  Electrical  Engineering  and  Computer  Science     at  the     Massachusetts  Institute  of  Technology     September  2015     ©  2015  Rebecca  P.  Krosnick.  All  rights  reserved.     The  author  hereby  grants  to  MIT  permission  to  reproduce  and  to  distribute  publicly   paper  and  electronic  copies  of  this  thesis  document  in  whole  or  in  part  in  any   medium  now  known  or  hereafter  created.         Author:    ________________________________________________________________________________       Department  of  Electrical  Engineering  and  Computer  Science       June  12,  2015     Certified  by:    ________________________________________________________________________________       Robert  C.  Miller       Professor  of  Electrical  Engineering  and  Computer  Science       Thesis  Supervisor     Accepted  by:    ________________________________________________________________________________       Albert  R.  Meyer   Professor  of  Electrical  Engineering  and  Computer  Science       Chairman,  Masters  of  Engineering  Thesis  Committee           1                                                                                                 2    VideoDoc:  Combining  Videos  and  Lecture  Notes   for  a  Better  Learning  Experience     by     Rebecca  P.  Krosnick     Submitted  to  the  Department  of  Electrical  Engineering  and  Computer  Science  on   June  12,  2015  in  Partial  Fulfillment  of  the  Requirements  for  the  Degree  of  Master  of   Engineering  in  Electrical  Engineering  and  Computer  Science       Abstract     Videos  provide  learners  an  engaging  way  to  learn  material,  but  they  are  not  easy  to   navigate.  Electronic  textbooks  are  easy  to  navigate  and  help  learners  review   material  they  have  already  seen,  but  they  are  not  very  engaging.  VideoDoc  combines   videos  and  textbooks  to  provide  learners  with  a  single  resource  that  engages  them   and  is  easy  to  navigate.  The  interface  can  be  played  like  a  video  or  read  like  a   textbook.  Lecture  videos  and  their  corresponding  transcripts  are  broken  into   sections  by  topic,  and  each  section  also  has  screenshots  of  representative  video   frames.  A  user  can  navigate  the  interface  by  scrolling  through  the  sections  or   clicking  on  section  titles  in  an  interactive  table  of  contents.  A  VideoDoc  lecture  is   automatically  generated  from  a  time-­‐annotated  text  transcript  and  a  labeling  of   talking-­‐head  video  frames,  and  an  instructor  can  fine-­‐tune  section  boundaries  and   add  section  titles  using  an  editing  interface.  Through  a  user  study  we  found  that   VideoDoc  helped  users  more  easily  navigate  lecture  videos,  but  some  users  had   trouble  learning  how  to  use  features  of  the  editing  interface.       Thesis  Supervisor:  Robert  C.  Miller   Title:  Professor  of  Electrical  Engineering  and  Computer  Science                   3    Acknowledgements     I  would  like  to  thank:     • My  outstanding  advisor,  Rob  Miller,  for  his  guidance,  support,  and  interest.  I   appreciate  the  substantial  amount  of  time  he  invested  in  me  and  this  project   even  though  I  was  a  short-­‐term  member  of  his  research  group.     • Rob  Miller  and  Elena  Agapie  for  conceiving  of  the  original  idea  of  VideoDoc.     • Michele  Pratusevich  for  working  alongside  me  this  year  to  build  the   computer  vision  parser  that  provides  input  to  VideoDoc  about  people  and   lecture  material.  Thank  you  for  your  quick  responses  to  my  emails   requesting  JSON  files.     • Juho  Kim  for  offering  advice  and  resources  as  I  wrote  my  thesis  proposal  and   designed  my  user  study.     • My  roommate  and  fellow  UID  member  Elena  Glassman  for  thoughtful   conversation  about  research  and  life.     • The  rest  of  the  User  Interface  Design  group,  including  Max  Goldman,  Carrie   Cai,  Amy  Zhang,  Abby  Klein,  and  Lyla  Fischer,  for  playtesting  VideoDoc  and   offering  suggestions  and  support.     • My  friends  and  family  for  their  support  throughout  this  process.  I  would   especially  like  to  thank  my  parents,  Lisa  and  Steven,  and  my  sister,  Sarah,  for   their  endless  love  and  support  this  year  and  always.                             4    Table  of  Contents   Abstract  .......................................................................................................................................  3   Acknowledgements  .................................................................................................................  4   1   Introduction  .......................................................................................................................  7   2   Related  Work  ..................................................................................................................  12   2.1   Making  Videos  Engaging  ...................................................................................................  12   2.2   Navigating  Videos  ................................................................................................................  13   2.3   Combining  Video  and  Text  ...............................................................................................  16   2.4   Massive  Open  Online  Courses  .........................................................................................  17   2.5   Video  Digests  .........................................................................................................................  21   3   User  Interface  ..................................................................................................................  23   3.1   Student  Interface  .................................................................................................................  23   3.1.1   Definitions  ........................................................................................................................................  24   3.1.2   Layout  ................................................................................................................................................  25   3.1.3   Getting  an  Overview  ....................................................................................................................  26   3.1.4   Watching  ...........................................................................................................................................  27   3.1.5   Navigating  ........................................................................................................................................  31   3.1.6   Searching  ..........................................................................................................................................  32   3.1.7   Reviewing  .........................................................................................................................................  33   3.2   Author  Interface  ..................................................................................................................  34   3.2.1   Section  Boundaries  .......................................................................................................................  35   3.2.2   Titles  ...................................................................................................................................................  41   3.2.3   Text  Styling  ......................................................................................................................................  41   3.2.4   Text  Editing  ......................................................................................................................................  44   3.2.5   Original  Transcript  With  Edits  ................................................................................................  45   4   Implementation  .............................................................................................................  47   4.1   Section  JSON  File  Format  ...................................................................................................  47   4.2   Generating  a  Section  JSON  File  ........................................................................................  54   4.3   Displaying  the  Interface  ....................................................................................................  57   4.4   Playback  .................................................................................................................................  58   4.5   Author  Interface  ..................................................................................................................  60   4.5.1   Text  Editing  and  Styling  .............................................................................................................  60   4.5.2   Indicating  Text  Styling  in  the  Toolbar  ..................................................................................  61   4.5.3   Section  Boundary  Changes  .......................................................................................................  62   4.5.4   Determining  the  Selected  Sentence  ......................................................................................  62   4.5.5   Displaying  Transcript  Additions  and  Deletions  ...............................................................  63   5   Evaluation  ........................................................................................................................  65   5.1   Research  Questions  ............................................................................................................  65   5.2   Participants  ...........................................................................................................................  66   5.3   Student  Interface  .................................................................................................................  66   5.3.1   Procedure  .........................................................................................................................................  66   5.3.2   Lecture  Videos  ................................................................................................................................  69   5.3.3   VideoDoc  Preparation  .................................................................................................................  70   5.3.4   User  Tasks  ........................................................................................................................................  72   5.3.5   Satisfaction  and  Usability  Questions  ....................................................................................  73   5    5.3.6   Results  ................................................................................................................................................  74   5.3.6.1   First  Time  Watching  Experience  .....................................................................................................  74   5.3.6.2   Navigation  .................................................................................................................................................  77   5.3.6.3   Other  Observations  ...............................................................................................................................  78   5.4   Author  Interface  ..................................................................................................................  79   5.4.1   Procedure  .........................................................................................................................................  79   5.4.2   Lecture  Videos  ................................................................................................................................  80   5.4.3   VideoDoc  Preparation  .................................................................................................................  81   5.4.4   Satisfaction  and  Usability  Questions  ....................................................................................  81   5.4.5   Results  ................................................................................................................................................  82   5.4.5.1   General  Editing  Approaches  .............................................................................................................  82   5.4.5.2   Changing  Section  Boundaries  ...........................................................................................................  83   5.4.5.3   Titles  for  Merged  and  Split  Sections  ..............................................................................................  84   5.4.5.4   Lecture  Slides  ..........................................................................................................................................  84   5.4.5.5   Representative  Frames  .......................................................................................................................  85   5.4.5.6   Styling  Text  ...............................................................................................................................................  86   5.4.5.7   Editing  Text  ..............................................................................................................................................  86   5.4.5.8   Other  Observations  ...............................................................................................................................  87   6   Discussion  and  Future  Work  .....................................................................................  88   6.1   User  Study  ..............................................................................................................................  88   6.2   Other  Future  Work  ..............................................................................................................  89   7   Conclusion  ........................................................................................................................  91   8   References  ........................................................................................................................  92                                                   6                 1         Introduction     Videos  and  electronic  textbooks  are  two  important  resources  used  in  online   education,  and  each  has  its  advantages  and  disadvantages.  Videos  are  highly   engaging  due  to  their  dynamic  nature  and  the  presence  of  a  human  voice  and  often  a   human  face,  whereas  textbooks  are  less  engaging  due  to  their  static  content  and  lack   of  human  interaction.  On  the  other  hand,  textbooks  are  easier  to  skim  than  videos.  A   user  can  see  hundreds  of  words  on  a  page  of  a  textbook  at  time,  allowing  them  to  get   an  overview  of  the  content  included,  whereas  in  a  video  a  user  must  scrub  through   to  see  multiple  frames  of  content.  Since  no  audio  is  played  during  scrubbing,  a  user   is  also  missing  out  on  these  important  spoken  words,  diminishing  the  quality  of   their  skimming.  The  quality  of  skimming  is  closely  related  to  the  quality  of   navigation.  A  user  can  navigate  a  textbook  relatively  easily  by  using  the  table  of   contents  and  the  index  to  identify  a  set  of  relevant  pages  and  then  using  the  Ctrl-­‐f   keyboard  shortcut  and  skimming  to  narrow  in  on  the  desired  content.  In  order  to   navigate  a  video,  the  user  must  scrub  until  they  find  the  desired  content,  since  they   do  not  know  ahead  of  time  at  what  time  which  content  occurs.  They  must  rely  on   visual  cues  during  scrubbing,  and  if  visual  cues  are  not  informative  enough,  the  user   will  use  trial  and  error  as  they  pause  and  play  the  video  to  determine  if  they  have   found  the  desired  content.     Users  take  advantage  of  both  videos  and  electronic  textbooks  in  a  single  course  1   so  they  have  an  effective  tool  for  each  of  their  learning  needs.  They  can  use  video  the   first  time  they  learn  the  material  so  they  are  engaged  and  motivated.  They  can  use   the  textbook  for  subsequent  times  they  review  the  material  because  it  is  easier  to   7    navigate.  Users  can  also  use  the  textbook  for  the  very  first  time  they  encounter  the   material  to  get  an  overview  of  the  material  contained.  However,  if  the  video  and   textbook  for  a  class  do  not  appear  on  the  same  webpage,  it  is  difficult  for  students  to   use  them  at  the  same  time.  Additionally,  if  the  video  and  textbook  content  do  not   cover  identical  content,  it  is  difficult  for  users  to  find  the  corresponding  parts  in  the   two  resources,  and  it  is  also  difficult  for  instructors  to  maintain  the  two  resources.   An  example  of  non-­‐identical  content  is  the  video  and  textbook  using  different   example  problems  to  explain  the  same  concept.     To  address  these  problems,  we  propose  VideoDoc  (Figure  1),  an  online  lecture   interface  that  combines  videos  and  a  textbook  layout  into  one  resource  in  order  to   improve  online  lecture  delivery.  VideoDoc  can  be  played  like  a  video  or  read  like  a   textbook.  These  are  VideoDoc’s  key  features:     • Sections:  A  lecture  video  is  broken  up  into  sections  by  topic,  shown  with  a   title,  the  corresponding  text  transcript,  and  representative  video  frames.   • Table  of  Contents:  A  list  of  section  titles  appears  on  the  left  of  the  page,   serving  as  a  table  of  contents  for  the  lecture.   • Static  Content:  A  user  can  quickly  skim  the  text,  representative  video   frames,  and  table  of  contents  to  get  an  overview  of  the  material  on  their  first   encounter.  They  can  later  read  the  text  to  review  the  material.   • Playback:  A  user  can  watch  the  engaging  video  to  learn  the  material.   • Navigation:  A  user  can  easily  navigate  the  lecture  by  scrolling  through  the   page  or  by  clicking  on  a  table  of  contents  title  to  be  taken  to  the   corresponding  section.   • Talking-­‐head  vs.  Content:  VideoDoc  extracts  talking-­‐head  clips  from  a  video   and  then  displays  content  and  talking-­‐head  video  streams  in  parallel,   allowing  a  user  to  view  relevant  content  frames  while  watching  the  talking-­‐ head.  The  content  video  stream  is  played  in  the  currently  selected  section’s   8    content  area,  and  the  talking-­‐head  stream  is  always  played  in  the  bottom  left   corner  of  the  viewport.     Combining  videos  and  text,  segmenting  videos  into  sections  and  streams,  showing   representative  video  frames,  and  offering  an  interactive  table  of  contents  provides  a   user  with  greater  control  over  the  content  they  view.           Figure  1:  VideoDoc  student  interface,  showing  a  6.00.1x  edX  lecture  on  selection  sort.  Here   the  lecture  is  in  play  mode  and  currently  at  the  last  sentence  of  the  section  titled  “Amortized   cost  of  sorting”,  as  indicated  by  the  sentence’s  yellow  highlighting  and  the  section’s  blue   border.         Creating  a  VideoDoc  lecture  requires  as  input  the  lecture  video,  a  time-­‐annotated   text  transcript,  and  a  labeling  of  which  parts  of  the  video  display  a  talking-­‐head  and   which  parts  do  not.  With  this  input  we  are  able  to  create  a  basic  VideoDoc  with   default  section  boundaries  and  no  section  titles.  A  course  instructor  can  then  edit   their  VideoDoc  using  the  author  interface  (Figure  2),  tailoring  section  boundaries,   creating  section  titles,  and  adding  text  styling.   9       Figure  2:  VideoDoc  author  interface,  showing  a  6.00.1x  edX  lecture  on  iteration  2.  Titles   and  the  text  transcript  are  editable  text  areas.  The  toolbar  at  the  top  of  the  screen  allows  for   styling  text,  viewing  text  additions  and  deletions  as  compared  to  the  input  transcript,  and   changing  section  boundaries  by  splitting  and  merging  sections.         We  evaluated  how  the  VideoDoc  student  interface  affects  learners’  watching  and   reviewing  experiences  as  well  as  how  usable  the  interface  is.  We  conducted  a   controlled  study  that  compared  the  VideoDoc  student  interface  with  the  edX  video   viewing  interface.  Participants  watched  and  then  answered  questions  about  short   6.00.1x  lectures  in  each  interface.  We  found  that  while  watching  a  lecture  in   VideoDoc,  participants  enjoyed  being  able  to  preview  upcoming  lecture  material  via   the  table  of  contents  and  scrolling  through  video  content  frames.  When  answering   questions  about  the  lecture  material,  participants  found  the  overview  provided  by   the  table  of  contents  to  be  helpful  in  guiding  their  search  for  material  and  found   clicking  on  table  of  contents  titles  and  scrolling  through  the  page  to  improve   navigation.  Participants  generally  found  the  interface  easy  to  learn.       We  evaluated  the  usability  of  the  VideoDoc  author  interface  by  presenting   participants  with  a  raw,  unedited  VideoDoc  and  asking  them  to  edit  it  as  if  they  were   an  instructor  for  the  course.  Participants  were  easily  able  to  add  and  edit  section   10    titles,  but  about  half  of  the  participants  had  trouble  changing  section  boundaries,   attempting  to  do  so  by  copying  and  pasting  text  rather  than  by  using  the  merge  and   split  buttons  in  the  toolbar.  After  coaching  them  on  the  features  they  had  trouble   learning,  participants  felt  that  the  author  interface  allowed  them  to  make  all  desired   edits  to  the  VideoDoc  lecture.     The  main  contributions  of  this  thesis  are:   • An  interface  that  presents  a  lecture  video  in  a  textbook-­‐like  layout  to  provide   an  overview  of  material,  improve  navigation,  and  reduce  disruptions  caused   by  media  transitions   • A  method  for  generating  a  bare-­‐bones  lecture  provided  the  lecture  video,   time-­‐annotated  text  transcript,  and  labeling  of  talking-­‐head  video  frames   • An  editing  interface  for  modifying  section  boundaries,  adding  titles,  editing   text,  and  styling  text                                       11                 2         Related  Work     In  this  chapter  we  present  existing  lecture  delivery  interfaces  and  how  users  learn   with  them.  We  explore  properties  that  affect  viewer  engagement  and  navigation  in   videos,  and  we  discuss  the  effect  of  combining  videos  and  text.       2.1     Making  Videos  Engaging   Because  videos  are  more  engaging  than  text,  we  wish  to  make  the  video  component   of  VideoDoc  as  engaging  as  possible.  Engagement  is  defined  as  learner  attention,   motivation,  and  satisfaction.  Previous  work  has  shown  three  ways  to  make  videos   more  engaging:  include  talking-­‐heads,  make  videos  short,  and  include  tablet-­‐writing.     Kizilcec  et  al.  3  study  how  picture-­‐in-­‐picture  lecture  videos  that  include  a  talking-­‐ head  in  the  corner  of  the  screen  impact  learners’  information  retention,  visual   attention,  and  affect  as  compared  to  lecture  videos  without  a  talking-­‐head.  They   found  that  the  presence  of  the  talking-­‐head  did  not  have  a  significant  impact  on   information  retention,  that  “on  average,  learners  spent  41%  of  time  looking  at  the   face  when  it  was  present  and  transitioned  between  looking  at  the  face  and  slides   every  3.7  seconds”,  and  that  learners  greatly  prefer  the  picture-­‐in-­‐picture  lecture   videos.  Although  there  is  no  evidence  that  learners  learn  better  with  the  presence  of   12    a  talking-­‐head,  it  appears  that  learners  are  more  engaged  and  motivated  to  learn   when  the  talking-­‐head  is  present.  In  VideoDoc  we  therefore  include  a  talking-­‐head   when  footage  exists  in  order  to  maximize  engagement.     Guo  et  al.  4  also  explore  properties  of  videos  that  are  particularly  engaging.  They   measure  engagement  “by  how  long  students  are  watching  each  video,  and  whether   they  attempt  to  answer  post-­‐video  assessment  problems”.  They  analyze  the  user’s   level  of  engagement  with  respect  to  video  length,  the  instructor’s  speaking  rate,   whether  the  video  was  a  lecture  or  tutorial  recording,  and  the  form  of  delivery.  The   most  significant  findings  are  that  the  most  engaging  videos  for  learners  are  short   videos  of  less  than  6  minutes  in  length,  videos  that  include  talking-­‐heads,  and  videos   that  are  Khan-­‐style.  As  a  result,  VideoDoc  presents  lecture  videos  as  a  series  of   shorter  clips,  and  when  there  is  appropriate  footage,  we  include  talking-­‐heads  and   Khan-­‐style  drawings.       2.2     Navigating  Videos   Kim  et  al.  5  study  click  data  from  edX  MOOCs  to  analyze  when  users  drop  out  of   videos  and  which  parts  of  videos  users  interact  with  the  most.  They  find  that  users   tend  to  drop  out  of  videos  that  are  longer,  re-­‐watched,  or  tutorial-­‐style.  They  also   find  that  high  levels  of  user  interaction  often  occur  at  visual  transitions.  One  such   example  is  users  re-­‐watching  a  video  at  the  point  where  the  slide  has  just  changed   and  the  instructor  is  starting  to  explain  a  new  concept.  Since  this  is  a  common  place   for  users  to  want  to  start  re-­‐watching  the  video,  we  would  like  for  it  to  be  easy  for   them  to  navigate  to.  However,  it  currently  is  not  easy  for  them  to  navigate  to,  since   they  must  scrub.  VideoDoc  eases  navigation  to  the  beginning  of  distinct  concepts,   where  interaction  peaks  are  likely  to  occur,  by  segmenting  videos  by  concept  and   displaying  all  video  clips  on  the  screen  at  the  same  time.  In  fact,  Kim  et  al.  6   suggest  “displaying  a  representative  frame  for  each  interaction  peak  to  visually   13    summarize  a  video”,  and  this  is  what  VideoDoc  does,  not  only  to  summarize  the   video  but  also  to  make  navigation  easier.  The  representative  frame  for  a  video  clip  is   the  slide  explained  inside  of  it.  As  a  result,  the  user  is  able  to  see,  at  the  same  time,   all  the  slides  contained  inside  the  video  and  is  able  to  easily  skim  the  slides  to   determine  which  video  clip  contains  their  topic  of  interest.     Another  example  of  high  levels  of  user  interaction  occurring  at  a  visual  transition  is   users  re-­‐watching  the  slide  that  occurred  right  before  the  video  transitioned  to  the   talking-­‐head.  This  may  happen  when  the  video  transitions  to  the  talking-­‐head  before   the  user  finishes  digesting  the  previous  slide.  VideoDoc  tries  to  prevent  these  abrupt   transitions  by  showing  the  user  two  video  streams  in  parallel:  a  talking-­‐head  stream   and  a  content  stream.  Instead  of  a  slide  disappearing  when  the  video  transitions  to  a   talking-­‐head,  the  slide  remains  visible  in  the  content  stream  while  the  talking-­‐head   is  shown  in  the  talking-­‐head  stream,  allowing  the  user  to  continue  digesting  the   slide  while  the  talking-­‐head  speaks.     Li  et  al.  7  built  and  tested  a  video  delivery  system  that  speeds  up  viewing  and   improves  navigation.  The  system  speeds  up  viewing  by  removing  pauses  in  speech   and  by  allowing  the  user  to  change  the  playback  speed.  The  system  improves   navigation  primarily  with  a  table  of  contents  and  shot  boundary  frames.  For  a   conference  presentation  or  classroom  video,  the  video  is  segmented  into  smaller   clips  based  on  slide  boundaries,  and  each  clip  is  represented  in  a  ribbon  of  frames   by  a  screenshot  of  its  corresponding  slide  or  talking-­‐head.  The  user  can  then  easily   scroll  through  the  ribbon  to  see  the  contents  of  the  video  and  click  on  a  particular   frame  to  play  the  corresponding  clip.  Each  clip  has  a  corresponding  line  in  the  table   of  contents,  which  can  also  be  clicked  to  play  the  clip.  Li  et  al.  found  that  users   enjoyed  the  video  delivery  system  because  it  saved  them  time  and  they  felt  they  had   more  “control  over  what  content  they  watched”  as  compared  to  a  traditional  system.   Users  likely  felt  they  had  more  control  because  they  could  easily  scan  through  the   set  of  slides  to  get  an  overview  of  the  content  as  well  as  to  navigate  to  a  desired   topic.  They  could  also  watch  the  video  continuously  and  anticipate  when  the  next   14    visual  transition  would  occur  and  what  content  the  next  slide  would  contain.   VideoDoc  uses  a  similar  concept  of  slide  screenshots  in  order  to  ease  navigation  and   give  users  more  control,  but  shows  content  and  talking-­‐heads  in  two  parallel   streams.     Li  et  al.  tested  their  system  on  a  variety  of  scenarios  including  classroom  lectures,   conference  presentations,  sports  games,  shows,  news  programs,  and  travel  videos.   For  scenarios  that  did  not  include  slides,  clips  were  segmented  based  on  significant   boundaries  as  according  to  the  system.  Users  particularly  enjoyed  the  ribbon  of   screenshots  in  baseball  games  because  they  could  choose  to  watch  only  clips  where   their  team  of  choice  was  up  to  bat.  They  also  enjoyed  the  screenshots  in  news   programs  because  they  could  get  a  quick  overview  of  the  stories  included  and   decide  which  ones  they  wanted  to  watch.     Monserrat  et  al.  8  built  NoteVideo,  a  system  that  improves  navigation  and  content-­‐ awareness  in  tablet-­‐writing  videos.  In  NoteVideo,  each  distinct  blackboard  character   in  the  original  video  is  linked  to  the  corresponding  video  clip  that  starts  at  that   character,  and  a  user  can  click  on  a  character  to  start  playing  the  corresponding   video  clip.  This  is  easier  than  scrubbing  the  video  to  the  location  where  the  letter  of   interest  first  appears,  as  is  necessary  in  a  traditional  video  interface.  Additionally,  at   all  times  all  characters  from  the  original  video  appear  on  the  screen,  with  future   characters  faded.  This  feature  allows  users  to  have  a  better  idea  of  the  overall   content  of  the  video.     Monserrat  et  al.  next  built  an  improved  version  of  the  software  called  NoteVideo+   that  adds  a  hovering  text  transcript  for  each  character  and  adds  a  scrubber.  The   transcript  for  each  character  allows  a  user  to  see  more  information  about  a   particular  part  of  the  video  before  playing  it.  Scrubbing  with  the  scrubber  helps  a   user  to  understand  the  order  in  which  certain  characters  appear  in  the  video.  In  the   first  version  the  user  could  only  easily  determine  which  characters  appeared  before   and  after  the  character  currently  being  written.  The  user  could  guess  the  order  of   15    characters  based  on  their  location  on  the  screen  but  when  multiple  columns  of   writing  started  appearing,  this  became  more  difficult.  NoteVideo+  is  an  interesting   solution  for  improving  video  navigation  and  content-­‐awareness,  and  VideoDoc   should  allow  similar  video  formats.  However,  we  cannot  rely  solely  on  the   NoteVideo+  video  format  to  improve  content  delivery  because  not  all  lectures  will   contain  tablet  writing.       2.3     Combining  Video  and  Text   Breslow  et  al.  1  conduct  a  broad  study  on  edX’s  first  MOOC,  6.002x.  They  analyze   how  and  when  students  use  each  MOOC  resource,  discuss  student  demographics,   and  analyze  student  success.  They  found  that  students  overall  spend  more  time   watching  the  lecture  videos  than  they  spend  using  any  other  single  resource.  They   also  found  that  textbook  usage  peaks  around  exam  time  and  that  during  exams   students  spend  more  time  using  the  textbook  than  any  other  single  resource.  These   usage  patterns  suggest  that  students  use  video  as  their  primary  source  of  learning   and  that  they  use  textbooks  to  study  for  exams  and  to  search  for  particular  content   for  answering  exam  questions.  These  observations  indicate  that  it  is  important  to   have  engaging  videos  for  the  first  stage  of  learning  and  that  it  is  important  to  have   easily-­‐skimmable  text  for  later  stages,  like  reference.  Since  videos  and  textbooks   accommodate  different  and  essential  aspects  of  learning,  it  is  important  to  include   both  of  these  resources  in  online  course  delivery.     An  early,  but  high  quality,  video-­‐text  hybrid  for  education  is  eClass,  built  by  Abowd   et  al.  9  in  the  late  1990s.  The  system  provides  a  page  of  vertically  scrollable  slides   and  a  continuous  video  of  the  professor  and  classroom  from  lecture.  There  is  also  a   timeline  of  the  lecture  that  includes  slide  and  website  references.  The  user  can  get   an  overview  of  the  content  by  scrolling  through  the  slides  and  can  navigate  to  a   particular  part  of  the  video  by  selecting  the  appropriate  slide,  similar  to  navigation   16    in  the  system  created  by  Li  et  al.  7.  VideoDoc  provides  similar  functionality  to   eClass  with  the  addition  of  a  text  transcript  of  the  audio.  We  believe  a  transcript  will   help  the  learner  review  explanations  without  having  to  re-­‐watch  the  video.  The   transcript  will  also  allow  a  learner  to  read  along  watching  the  video  if  they  wish  to   do  so.  Adding  in  the  transcripts  requires  us  to  modify  the  layout,  but  we  keep  the   vertically  scrollable  slides.     In  2012,  Miller  et  al.  10  built  an  interactive  electronic  textbook  for  teaching  novice   programmers.  The  textbook  includes  traditional  text,  short  videos  that  cover  the   most  important  concepts,  and  a  code  visualizer.  The  visualizer  allows  students  to   run  code  written  by  the  instructors,  write  and  run  their  own  code,  and  see  the  state   of  their  variables  and  the  program  output  at  each  time  step.  Miller  et  al.  found  that   for  students  in  a  blended  class  using  the  textbook,  the  students  found  classroom   lectures  to  be  the  most  helpful  to  learning,  followed  by  reading  the  textbook  and   then  using  the  code  visualizer.  Students  also  enjoyed  having  all  of  the  course   material  in  one  place,  which  supports  the  hypothesis  that  a  hybrid  video-­‐textbook   system  like  VideoDoc  will  improve  the  learning  experience.  VideoDoc  does  not,   however,  show  short  video  clips  that  cover  material  non-­‐identical  to  the  text   because  we  believe  this  is  confusing  for  students  and  difficult  for  instructors  to   maintain.  It  is  interesting  that  students  did  not  mention  the  short  videos  in  the   textbook  as  one  of  the  more  helpful  learning  tools.  Perhaps  the  fact  that  students   ranked  classroom  lectures  and  textbook  reading  high  indicates  that  students  find   audiovisual  lectures  given  by  humans  to  be  helpful  when  introducing  the  material   and  they  find  text  helpful  when  reviewing  the  material.       2.4     Massive  Open  Online  Courses   We  must  of  course  also  explore  MOOCs  because  they  are  a  large  multimedia   educational  resource  used  today.  We  will  see  that  current  MOOCs  do  not  provide  the   17    quality  of  navigation  and  content-­‐awareness  that  we  would  like.  The  major  MOOC   platforms  are  edX  11,  Coursera  12,  and  Udacity  13.  Their  MOOCs  have  videos  as   the  centerpiece  and  they  also  have  PDF  course  notes,  homework  assignments,  and   discussion  forums.  In  Figures  3,  4,  and  5  we  show  screenshots  of  the  video-­‐viewing   interface  on  each  MOOC  platform.     The  MOOCs  show  spoken  words  to  varying  degrees.  edX  MOOCs  have  a  full  text   transcript  next  to  the  playing  video,  with  the  line  currently  being  spoken  highlighted   in  the  transcript.  Coursera  MOOCs  provide  a  transcript  that  can  be  downloaded  as  a   text  file  from  a  different  page  on  the  site  than  where  the  video  is  played.  A  user   could  download  the  transcript  to  view  side  by  side  with  the  video,  but  this  is  not   convenient.  Udacity  MOOCs  provide  no  transcript  at  all.  In  Coursera  and  Udacity  a   user  can  turn  on  closed  captioning  to  view  spoken  words  as  text  on  the  screen,  but   closed  captioning  is  not  as  informative  as  a  full  transcript.     Additionally  these  MOOCs  are  not  as  navigationally  friendly  as  the  systems  built  by   Li  et  al.  7  and  Abowd  et  al.  9.  Coursera  videos  are  long,  most  longer  than  10   minutes  and  many  longer  than  15  minutes.  Yet  Coursera  gives  no  indication  of  the   topics  within  a  video  besides  one  very  general  topic  for  the  entire  video,  making  it   difficult  for  a  user  to  navigate  to  a  specific  concept  when  re-­‐watching  the  video.   Udacity  has  short  videos,  most  less  than  5  minutes  long,  and  each  video  is  labeled   with  a  topic  on  the  ribbon  of  boxes  at  the  top  of  the  page.  However,  the  user  must   hover  over  a  particular  box  to  see  the  topic  for  the  corresponding  video,  so  the  user   still  cannot  get  an  overview  of  all  the  short  videos  by  just  glancing  at  the  screen.  edX   has  a  similar  ribbon  to  Udacity  and  longer  videos  than  Udacity,  so  it  is  also  difficult   to  navigate  and  to  get  an  overview  of  the  lesson.     Udacity  videos  are  always  the  tablet-­‐writing  style,  but  note  that  edX  and  Coursera   videos  can  be  a  variety  of  production  styles,  not  just  the  ones  shown  in  the  Figures  3   and  4.  Possible  production  styles  include  slides,  tablet-­‐writing,  picture-­‐in-­‐picture,   18    studio  whiteboard,  green-­‐screen  instructor,  and  traditional  lecture  hall,  among   others.                   Figure  3:  An  edX  Introduction  to  Solid  State  Chemistry  lecture  14.  To  the  right  of  the  video   appears  an  interactive  text  transcript.  Directly  above  the  video  appears  a  ribbon  of  other   lectures  and  exercises  in  the  Covalent  Bonding  learning  sequence.  At  the  top  of  the   screenshot  are  links  to  other  class  resources  including  the  textbook  and  discussion  forum.       19       Figure  4:  A  Coursera  Natural  Language  Processing  lecture  15.  Coursera  lectures  do  not   provide  a  text  transcript  on  the  same  page  as  the  video.         Figure  5:  An  Udacity  Applied  Cryptography  lecture  16.  Above  the  video  appears  a  ribbon   of  other  lectures  and  exercises  in  Lesson  1.  Below  the  video  are  downloadable  videos  and   transcripts  for  Lesson  1.  On  the  left  of  the  page  are  links  to  other  class  resources  including   written  materials  and  the  discussion  forum.   20    

Advise: Why You Wasting Money in Costly SEO Tools, Use World's Best Free SEO Tool Ubersuggest.