Perceptual fluency and duration judgments An experimental study regarding the effect of fluency on perception of time

Erik Holmlund

Spring term 2016 Bachelor Thesis in Psychology, 15 hp Supervisor: Ulrich Olofsson, Phd; Department of Psychology, Umeå University

   

PERCEPTUAL  FLUENCY  AND  DURATION  JUDGMENTS     Erik  Holmlund  

This   study   investigated   whether  perceptual   fluency  could  affect  duration  judgments.  Fluency  refers   to  levels  of  subjective  ease,  in  which  stimuli  can  be  processed  (Lanska,  Olds,  &  Westerman,  2014).   The   study   was   conducted   with   experimental   within   group   factorial   design.   Visual   stimuli   were   selected   from   Snodgrass   and   Vanderwarts   (1980)   standardized   set   of   260   pictures.   Pairs   were   made   with   low   and   high   levels   of   complexity.   Duration   was   about   1000   milliseconds   with   .10   variations.  1/3  of  pairs  were  without  variation.  Participants  were  asked  to  judge  which  image  was   presented   for   longest   time.   Total   amount   of   participants   was   37.   Main   hypothesis   was   that   low   levels  of  complexity  would  be  judged,  to  a  greater  frequency,  as  having  been  presented  for  longer   duration.   Observed   mean   (M=   20.27,   SD   =   2.90)   was   slightly   lower   than   level   of   chance   (M   =   21)   and  the  difference  was  non  significant,  t(36)  =  -­‐1.53,  p  >  .13.  The  null  hypothesis  was  not  rejected.         Denna   studie   undersökte   huruvida   grad   av   perceptuell   fluency   kan   påverka   bedömningar   av   duration.  Fluency  syftar  till  nivå  av  subjektivt  lätthet  varpå  stimuli  kan  bearbetas  (Lanska,  Olds  &   Westerman,  2014).  Studien  genomfördes  med  experimentell  inom  grupp  faktoriell  design.  Visuella   stimuli   valdes   från   Snodgrass   och   Vanderwarts   (1980)   standardiserade   set   av   260   bilder.   Par   gjordes  med  låga  och  höga  nivåer  av  komplexitet.  Durationer  var  runt  1000  millisekunder  med  .10  i   variation.   1/3   av   paren   var   utan   variation.   Deltagarna   ombads   att   bedöma   vilken   bild   som   presenterades   under   längst   tid.   Totala   antalet   deltagare   vad   37.   Huvud   hypotesen   var   att   låga   nivåer   av   komplexitet   skulle   bedömas,   till   en   högre   frekvens,   att   ha   presenterats   under   längre   duration.   Observerad   medelvärde   (M   =   20.27,   SD   =   2.90)   var   något   lägre   än   nivån   av   slumpen   (M   =   21)  och  skillnaden  var  inte  signifikant,  t(36)  =  -­‐1,53,  p  >  .13.  Noll  hypotesen  förkastades  inte.  

    The  perception  of  time  is  regarded  as  a  subjective  phenomenon  (Hansen  &  Trope,   2013).   The   same   situation   can   elicit   different   experiences   of   time   depending   on   different   situational   factors,   for   example   the   attractiveness   and   richness   of   the   situation  (Ahn,  Liu,  &  Soman,  2009).  A  highly  familiar  claim  is  that  the  impression   of   time   depends   on   how   busy   you   are.   Many   know   the   experience   of   a   busy   schedule  eliciting  feelings  of  how  “time  flies”  compared  to,  for  example,  watching   paint  dry.  This  type  of  claim  is  actually  supported  by  the  scientific  community.  The   more  changes  (Hansen  &  Trope,  2013)  and  the  amount  of  richness  (Avni-­‐Babad  &   Ritov,   2003)   in   given   situation   leads   to   an   experience   of   time   passing   faster   due   to   the   fact   that   less   amount   of   attention   directed   to   time.   On   the   other   hand,   Ahn,   Liu,   and   Soman   (2009)   found   results   in   their   study   indicating   that   more   richness   in   given  situation,  being  characterized  by  vivid  stimuli,  seems  to  elicit  a  notion  of  time   passing  quicker  yet  feel  as  longer  at  recall.       Time  duration  judgments  are  thought  to  be  affected  by  cognitive  load,  working  in   opposite   directions   depending   on   the   duration   judgment   paradigm   (Block,   Hancock,   &   Zakay,   2010).   This   refers   to   the   fact   that   if   participants   know   that   duration   judgments   are   of   importance   (prospective   paradigm)   or   if   participant   are   not  aware  of  time  judgments  until  after  duration  ended  (retrospective  paradigm),   results  in  different  judgments  of  duration.  The  prospective  paradigm  is  thought  to   regard   attention-­‐based   models   of   explanation   and   the   retrospective   paradigm   memory-­‐based   models   of   explanation   (Block   &   Zakay,   1997).   The   relationship  

2

between  perception  of  time  and  duration  judgments  is  summarized  as:  the  faster   time  seems  to  pass  the  smaller  experience  of  duration  (Ahn,  et  al.  2009).       Within  the  domain  of  the  retrospective  paradigm,  processing  difficulty  is  known  to   affect   duration   judgment   (Block   &   Zakay,   1997),   but   significant   only   for   easy   processing.  Block  and  Zakay  stresses  the  importance  of  processing  difficulty  during   target  durations  as  an  important  mediator  variable.  Direction  of  relationship  is;  as   difficulty  increases,  the  experience  of  duration  decreases.  This  study  was  aimed  at   investigating   whether   a   similar   variable   to   processing   difficulty,   within   the   retrospective  paradigm  and  stimuli  processing,  could  affect  duration  judgments.       The  variable  of  interest  is  known  as  perceptual  fluency  and  refers  to  the  speed  and   ease   in   which   stimuli   can   be   processed   (Lanska,   Olds,   &   Westerman,   2014).   A   high   level  of  fluency  refers  to  easy  processing  and  shorter  recognition  speed  compared   to   low   level   of   fluency   with   more   extensive   processing   and   longer   recognition   speed.  Underlying  assumptions  are  that  stimuli  can  vary  in  degree  of  ease  in  which   processing  can  proceed  (Reber,  Schwarz  &  Winkielman,  2004).  Thus  can  different   people   process   the   same   stimuli   with   different   levels   of   subjective   ease/fluency   (Brinol,  Petty  &  Tormala,  2006).  Brinol,  Petty  and  Tormala  point  out  that  this  could   be  of  importance  within  the  domain  of  judgments  and  attitudes.  Two  phenomena   are   especially   relevant:   ease   of   retrieval   and   general   processing   fluency.   This   is   regarded   as   related   phenomena   that   tap   in   to   different   sources   of   fluency.   Exemplified   as   the   fact   that   feelings   of   ease   and   fluency   can   arise   due   to   the   fact   that   thoughts   are   easy   to   retrieve   and   generate   or   that   the   stimuli   are   easy   to   process  or  perceive.  This  adds  to  previous  research  indicating  that  the  underlying   reasons   for   variation   in   fluent   processing   are   found   in   both   subjective   and   objective   paradigms   (Reber,   et   al.   2004).   This   also   reflects   the   distinction   between   conceptual   and   perceptual   fluency   research   pointed   out   by   Lanska,   Olds,   and   Westerman  (2014).       The  distinction  between  ease  of  retrieval  and  general  processing  fluency  (Brinol,  et   al.   2006)   is   also   interpreted   as   to   concern   the   duration   judgment   paradigm   as   discussed   earlier.   Duration   judgment   within   the   prospective   paradigm   is   considered   to   be   explained   by   attention-­‐based   models   (Block   &   Zakay,   1997),   which   is   interpreted   to   result   from   perceptual   processing,   fluency   and   objective   variables.   This   study   aims   at   perceptual   processing   fluency,   thus   focusing   mainly   on  objective  variables  although  subjective  variables  will  be  discussed  briefly.       The   subjective   experience   is   regarded   as   a   central   part   in   the   concept   of   fluency   (Reber,   et   al.   2004).   Studies   show   that   the   easier   given   stimuli   can   be   processed,   the  more  appealing  stimulus  is  experienced  to  be.  Reber,  Schwarz  and  Winkielman   (2004)   also   point   out   the   fact   that   this   principle   can   be   linked   to   aesthetic   judgments.  More  fluent  processing  is  regarded  as  being  linked  to  higher  aesthetic   responses,  which  is  being  defined  as  an  enjoyable  subjective  experience  directed  at   an  object  without  previous  reasoning.      

3

Subjective   variables   (Reber,   et   al.   2004)   known   to   affect   levels   of   fluency   processing   identified   by   previous   research   include   prior   experiences,   repetition   (mere   exposure   effect),   implicit   evaluations   regarding   regularity   and   prototypically.   These   variables   links   fluent   processing   through   aesthetic   evaluations   of   beauty.   Previous   research   indicates   that   the   mere   exposure   effect   possibly  could   enhance   objects   perceptual   fluency   (Seamon,   Brody   &   Kauff,   1983).   More   amount   of   exposure   to   same   stimuli,   or   similar   stimuli,   is   interpreted   to   reduce  the  amount  of  resources  of  perceptual  processing.       Objective   variables,   identified   by   previous   research,   known   to   affect   fluent   processing   through   aesthetic   judgments   include   amount   of   information,   contrast,   clarity   and   symmetry   (Reber,   et   al.   2004).   Less   amount   of   information   is   considered   to   correlate   with   higher   aesthetic   evaluations,   thus   linking   to   higher   perceptual   fluency.   Furthermore,   less   amount   of   information   is   also   thought   to   result   in   more   effective   cognitive   processing,   measured   with   recognition   time   (Checkosky   &   Whitlock,   1973).   Interpreted   as   higher   perceptual   fluency.   This   principle  extends  to  contrast  (Reber,  Winkielman  &  Schwarz,  1998).  More  contrast   is   linked   to   higher   perceptual   fluent   processing,   thus   faster   recognition   speed   (Reber,   et   al.   2004),   a   phenomenon   specifically   current   during   short   presentations   (Reber,   et   al.   2004,   referenced   in   Reber,   &   Schwarz,   2001).   Reber,   Schwarz   and   Winkielman  (2004)  also  links  clarity  as  a  related  variable  that  has  been  shown  to   influence   perceptual   fluency.   Studies   show   that   contrast   and   clarity   affect   perceptual   fluency   through   the   aesthetic   evaluations.   More   contrast   and   clarity   lead  to  more  positive  aesthetic  evaluations  (Reber,  et  al.  1998).  Symmetric  shapes   are   considered   to   contain   less   amount   of   information   compared   to   asymmetric   shapes  (Garner,  1974).  Thus  linking  level  of  symmetry  as  a  variable  of  importance   to  perceptual  fluency  processing.       Previous   study   found   results   indicating   that   longer   presentation   time   was   correlated   with   higher   aesthetic   evaluations   compared   to   shorter   presentation   times   (Reber,   et   al.   1998).   An   independent   pilot   study   at   Umeå   University,   presented   as   a   poster   only,   found   result   indicating   significant   differences   in   duration  judgments  due  to  manipulation  in  level  of  fluency  (U.  Olofsson,  personal   communication,  Feb  2016).  Whereas  more  fluent  visual  stimuli  where  perceived  to   be   presented   with   longer   duration   compared   with   less   fluent   stimuli,   when   in   fact,   duration   was   the   same.   The   goal   for   present   study   was   to   build   further   knowledge   on   the   question   if   level   of   fluency   can   affect   perception   of   time.   Speed   of   recognition   is   considered   to   be   a   standardized   tool   for   measure   of   perceptual   fluency  (Reber,  et  al.  2004).  In  this  study,  recognition  speed  is  not  measured  per  se,   but   underlying   assumption   is   that   more   fluent   stimuli   is   processed   with   faster   recognition  speed  compared  to  less  fluent  stimuli.     Snodgrass   and   Vanderwart   (1980)   have   developed   a   standardized   set   of   260   pictures  with  control  for  four  major  variables;  norms  for  name  agreement,  image   agreement,  familiarity  and  visual  complexity.  The  drawings  are  black  a  white  line   art   drawings,   presented   in   a   realistic   way,   of   typical   real   life   objects   (kitchen   materials,   animal,   body   parts,   insects   and   furniture   for   example).   Variables   of  

4

relevance  for  this  study  are  familiarity  (M  =  3.29,  SD  =  .95)  and  visual  complexity   (M  =  2.96,  SD  =  .89)  both  measured  on  a  five-­‐point  scale.  These  two  concepts  are   significant   negatively   correlated   (-­‐.466)   because   complex   drawings   are   considered   more  unfamiliar  than  more  simple  drawings.       Fluency  was  operationalized  as  visual  complexity  on  Snodgrass  and  Vanderwarts   (1980)  set  of  standardized  pictures.  Pairs  were  made  where  one  drawing  had  low   complexity   (LC)   and   the   other   had   high   complexity   (HC).   Perception   of   time   was   operationalized  as  the  forced  choice  selection  of  which  drawing  that  was  presented   for   longest   duration   within   each   pair.   Dependent   variable   was   the   selection   of   image  being  perceived  as  having  been  presented  during  longest  duration.       LC  >  HC;  LC  =  HC;  LC  <  HC       Exposure   duration   kept   to   maximum   1100   milliseconds   to   avoid   counting   strategies   (Lavoie,   &   Grondin,   2004).   Weber's   law   used   to   compute   and   for   temporal   discrimination.   Differences   in   duration   were   .10.   Duration   presentation   for   short   presentation   was   1000   ms.   Long   presentation   was   1100   ms.   For   the   category  with  same  presentation  time  duration  were  1050  ms.  For  half  of  the  pairs,   LC   is   presented   before   HC   with   the   order   reversed   for   the   other   half.   This   was   made   to   compensate   for   recency   and   primacy   effects.   Familiarity   was   matched   within  pairs.       The   alternative   main   hypotheses   was   that   LC   images   were   to   be   selected   more   often  than  HC:  LC  >  HC,  this  in  line  with  previous  discussed  theories.       LC  >  HC:  LC  selected  more  frequently   LC  =  HC:  LC  selected  more  frequently   LC  <  HC:  LC  selected  less  frequently     Drawings   presented   for   longer   durations   were   thought   to   be   selected   more   frequently:   (HC+   +   LC+)   >   (HC   +   LC).   There   was   thought   to   be   an   interaction   between  duration  and  complexity:  (LC+  -­‐  HC)  >  (LC  -­‐  HC)  >  (HC+  -­‐  LC).     Null  hypothesis  was  that  there  is  no  main  effect/interaction  between  fluency  and   perceived   duration   judgments.   Response   frequency   for   values   picked   low   complexity  images  would  then  be  equal  to  level  of  chance  (M  =  21).       (LC  >  HC;  LC  =  HC;  LC  <  HC)  =  21         Method     The   study   was   conducted   with   an   experimental   within   group   factorial   design.   Independent   variables   were   level   of   fluency   and   duration.   Dependent   variable   was   perceived   duration   judgment.   Forty-­‐two   trails   were   divided   to   three   conditions   with  14  trails  each.  For  14  of  the  pairs  low  complexity  was  presented  longer  than  

5

high  complexity:  LC  >  HC.  For  14  of  the  pairs  HC  was  presented  longer  than  LC:  HC   >   LC.  Finally,  for  14  of   the  pairs  LC   and   HC   was   presented   with   the   same   duration:   LC   =   HC.   These   three   conditions   were   split   in   half   with   reverse   order   of   presentation.   Thus   compensating   for   recency   and   primacy   effects.   Furthermore   was   order   of   pair   presentation   randomized.   Hence   compensating   for   learning   effects  due  to  variation  of   between  pair  complexity  (M  =  2.96,  SD  =  .46,  Minimum  =   2.13,   Maximum   =   3.78).   The   pictures   used   for   this   study   was   neutral   by   nature,   therefore   not   considered   to   risk   psychological   distress   to   participants.   Participation   was   voluntary   and   informed   consent   was   made.   Participants   were   aware  of  their  rights  to  discontinue  the  experiment  without  declaration.       Materials     Fluency  was  operationalized  as  visual  complexity  on  Snodgrass  and  Vanderwarts   (1980)  standardized  set  of  260  pictures  measured  on  a  five-­‐point  scale.  Pairs  were   fixed,  but  each  pair’s  condition  was  randomized.  Fluency  was  categorized  as  either   low  with  high  visual  complexity  (M  =  4.04,  SD  =  .43)  or  high  fluency  with  low  visual   complexity  (M  =  1.89,  SD  =  .50).  Pairs  were  fixed  with  a  difference  within  pairs  at   minimum   2   scale   points   (M   =   2.14,   SD   =   .16)   for   visual   complexity.   Familiarity   variation   was   maximum   .54   points   on   a   five-­‐point   scale   (M   =   .23,   SD   =   .14),   to   ensure   that   the   same   amount   of   processing   was   directed   to   image   recognition   within   pairs.   Participants   was   asked   under   a   forced   choice   a   condition   to   choose   which   picture   was   perceived   as   having   been   presented   for   the   longest   duration.   Long  duration  was  1100  ms  and  short  duration  was  1000  ms.  Condition  with  same   duration  was  both  pictures  presented  for  1050  ms.  Software  used  to  program  the   experiment   was   Inquisit   Lab   5.   The   experiment   was   conducted   on   the   same   MacBook  computer  with  a  13-­‐inch  screen.  Screen  light  was  set  at  maximum  for  all   participants.       Participants     Participants   were   students   sampled   with   convenience   sampling   in   the   area   of   Umeå  University’s  library.  Participant  was  asked,  where  seated,  to  take  part  in  the   experiment,   mainly   within   the   study   hall   and   nearby   coffee   shop   under   quiet   conditions.   The   total   amount   of   participants   was   37.   Of   these   were   19   male   and   18   female.  Mean  age  was  about  24  years  (M  =  23.78,  SD  =  2.72).     Procedure     Participants   were   asked   to   take   part   of   an   experimental   study   regarding   visual   images  and  perception  of  time.  Instructions  explained  the  experiments  procedure   and   were   given   both   in   text   and   verbally   when   needed.   The   task   was   to   choose   which   picture   was   presented   under   longest   amount   of   time.   The   experiment   was   designed  with  a  fixation  point  presented  for  1000  ms  followed  by  two  pictures  in   sequence.   After   both   pictures   were   presented   a   question   mark   appeared   and   participant   chose   key   1   for   the   first   picture   and   key   2   for   the   second   picture,   followed  by  next  trail.  The  experiment  took  about  five  minutes  to  complete.  

6

  Analysis     Data   was   initially   plotted   and   considered   normally   distributed   for   all   categories.   Significance   level   was   set   at   .05   and   analysis   was   two   tailed.   One   samples   t-­‐test   was   made   for   main   effect   analysis   and   a   2   x   3   ANOVA   specific   analysis   of   interaction  effects.         Results     First,  to  establish  to  what  extent  subjects  were  able  to  reliably  detect  the  correct   durations,   the   mean   number   of   correct   classifications   (M   =   17.46,   SD   =   3.02)   was   compared  to  level  of  chance  (M  =  14).  A  one  sample  t-­‐test  found  this  difference  to   be  significant,  t(36)  =  6.96,  p  <  .00.       The   main   question   was   whether   low   complexity   items   would   be   classified   as   having   the   longer   duration   to   a   greater   frequency   level   than   chance.   Contrary   to   the  hypothesis  the  observed  mean  of  20.27  (SD  =  2.90)  was  slightly  lower  than  level   of  chance  (equal  probability  of  selecting  low  complexity  items  and  complex  items).   A  one-­‐sample  t-­‐test  revealed  that  the  difference  was  non  significant,  t(36)  =  -­‐1.53,  p   >  .13.       The   effects   of   presentation   order   and   duration   were   tested   with   a   2   (Order)   x   3   (Duration)   ANOVA.   It   yielded   significant   a   main   effect   of   presentation   order,   F(1,36)  =  21.60,  p  <  .00,  eta2  =  .37,  due  to  subjects  selecting  the  low  complexity  item   more  often  if  it  was  presented  second  than  first  in  the  pair.  There  was  also  a  main   effect   of   exposure   duration,   F(1,36)   =   28.08,   p   <   .00,   eta2   =   .44,   reflecting   that   subjects   were   more   likely   to   select   the   low   complexity   item   when   it   was   actually   presented  at  a  longer  duration.  Pairwise  comparisons  between  the  three  durations   were  all  significant  (p  <  .05).  There  was  no  interaction  between  presentation  order   and  duration  however,  F(1,36)  =  1.28,  p  >  .28.       Looking   specifically   on   the   pairs   that   were   presented   for   equal   duration,   low   complexity  items  were  significantly  judged  as  having  a  longer  duration  than  high   complexity   items   only   if   they   were   presented   second   in   the   pair   (M   =   4.03,   SD   =   1.42),   t(36)   =   2.25,   p   <   .05.   When   they   were   presented   first   they   were   instead   judged  as  having  the  shorter  duration  (M  =  2.95,  SD  =  1.22),  t(36)  =  2.75,  p  <  .01.       Discussion     The  main  hypothesis  for  this  study  was  that  duration  judgments  would  be  higher   in   frequency   for   low   complexity   images   compared   to   high   complexity   images.   Intention  was  to  build  further  knowledge  to  previous  pilot  study,  which  indicated   that   fluency   significantly   affected   duration   judgments   (U.   Olofsson,   personal   communication,  Feb  2016).  Higher  levels  of  fluency  seemed  to  enhance  perceived  

7

duration  judgments.  Contradictory  to  previous  research  the  main  effect  from  this   study   was   in   fact   working   in   opposite   direction,   although   non   significant   (M   =   20.27,   SD   =   2.90)   t(36)   =   -­‐1.53,   p   >   .13.   Statistical   power   calculations   showed   low   levels  of  power  (33.4%).  The  null  hypothesis  was  not  rejected.     Subjects  were  thought  to  be   able  to  reliably  detect  the  correct  durations  compared   to  level  of  chance  (14),  which  was  correctly  predicted  (M  =  17.46,  SD  =  3.02)  t(36)  =   6.96,   p   <   .00.   Statistical   power   was   high,   calculated   to   100%.   Furthermore,   there   was   thought   to   be   an   interaction   between   duration   and   complexity:   (LC+   -­‐   HC)   >   (LC   -­‐   HC)   >   (HC+   -­‐   LC).   The   overall   effect   from   data   analysis   was   not   supporting   this   hypothesis,   although   parts   of   it   could   be   regarded   as   to   support   the   hypothesis.  For  example  the  main  effect  of  exposure  duration,  F(1,36)  =  28.08,  p  <   .00,  eta2  =  .44,  which  reflected  that  subjects  were  more  likely  to  select  the  simple   item  when  it  was  actually  presented  at  a  longer  duration.       The   absence   of   main   effect   can   be   due   to   two   possible   explanations.   Either   there   is   no   effect,   that   fluency   does   not   affect   duration   judgments,   or   the   operational   definition   of   independent   variable   of   fluency   was   not   appropriate.   Third   possible   reason  could  be  due  to  weak  manipulation,  which  will  be  discussed  subsequently.   Reber,  Schwarz  and  Winkielman  (2004)  stresses  objective  variables  identified  by   previous  research,  known  to  affect  fluent  processing  through  aesthetic  judgments   such   as;   amount   of   information,   contrast,   clarity   and   symmetry.   In   this   study,   fluency   was   operationalized   as   visual   complexity   on   Snodgrass   and   Vanderwarts   (1980)   set   of   standardized   pictures.   Although   high   control   of   variation   for   fixed   pairs   regarding   visual   complexity,   no   other   previous   known   variable   was   manipulated  in  this  study.  Visual  complexity  is  regarded  as  similar  manipulation  as   amount  of  information  and,  to  a  degree,  symmetry.  Some  of  selected  images  were   regarded  as  symmetric,  wheel  for  example,  although  not  controlled  for.  However,   possible  explanation  for  lack  of  main  effect  might  be  due  to  low  construct  validity   and  week  operational  definition  of  fluency.  Maybe  contrast  and  clarity  also  should   have   been   manipulated.   On   the   other   hand,   these   variables   were   controlled   for   because  all  images  were  presented  with  same  contrast  and  clarity.         The   independent   pilot   study   regarding   fluencys   effect   on   duration   judgments   did   not   control   for   amount   of   fluency   manipulation.   Because   ordinary   photographic   images   were   used   there   was   little   control   for   either   which   variables   was   manipulated   and   how   much   manipulation   was   made,   except   for   duration   manipulation.   Which   in   fact   was   the   same   amount   .10   as   for   present   study   although  durations  was  longer  (1800  and  2000  ms).  Possible  explanation  for  this   difference  between  results  may  also  be  that  the  two-­‐point  manipulation  (M  =  2.14,   SD   =   .16)   on   Snodgrass   and   Vanderwarts   (1980)   five-­‐point   scale   for   visual   complexity   was   too   weak.   Thus   resulting   from   insensitive   measures,   that   the   operationalization   of   the   independent   variable   fluency   was   possibly   not   enough   sensitive.   On   the   other   hand,   a   larger   difference   on   Snodgrass   and   Vanderwarts   five-­‐point   scale   for   visual   complexity   would   possibly   appear   too   obvious   for   participants.   Reasons   for   choosing   images   from   this   set   of   pictures   were   high   control  for  amount  of  manipulation.    

8

  Results   that   did   show   significant   differences   for   example   when   low   complexity   (LC)   images   was   presented   second   in   pair   and   was   significantly   more   frequent   selected   as   to   be   presented   with   longer   duration   (M   =   4.03,   SD   =   1.42),   t(36)   =   2.25,   p  <  .05  and  the  other  way  around  for  first  presentation  within  pairs   (M  =  2.95,  SD  =   1.22),   t(36)   =   2.75,   p   <   .01,   is   interpreted   as   cancelling   each   other   out.   This   is   regarded  to  be  due  to  recency  and  primacy  effects.  Statistical  power  calculated  to   62%   when   LC   images   were   presented   second   and   78%   when   LC   images   was   presented  first  in  pairs.     External  validity  was  not  prioritized  for  this  study  because  the  main  purpose  was   more   directed   towards   theoretical   testing.   The   goal   was   not   to   generalize   which   motivated   convenience   sampling.   The   sample   is   regarded   as   non-­‐representative   and  results  not  able  to  be  generalized  beyond,  in  this  case,  the  setting  and  situation   of   students   at   Umeå   University.   On   the   other   hand,   there   is   no   particular   hypothesis   stating   a   theory   of   differences   between   students   and   a   broader   population,   except   maybe   targeting   age   as   a   variable   of   interest.   For   example,   it   could   be   a   difference   in   old   age   due   to   cognitive   decline.   The   experiment   was   conducted  in  a  setting  with  little  control.  This  could  possibly  affect  the  dependent   variable   if   the   experiment   had   been   designed   with   a   between   groups   design.   A   within   group   design   is   regarded   to   be   less   sensitive   to   situational   influences   and   individual   differences.   Situational   settings   were   aimed   at   a   calm   and   discrete   experimental   situation   and   the   variation   was   not   grand.   Another   specific   consideration   is   whether   half   of   the   participants   should   have   been   randomly   assigned   to   be   asked   about   which   image   was   presented   during   shortest   duration   and   thus   creating   a   negative   result   which   could   be   transformed   in   analysis.   This   could  possibly  have  affected  the  dependent  variable.     Temporal   precedence   was   fulfilled   and   conditions   were   randomly   created   for   all   participants.  Internal  validity  was  regard  as  good.  A  within  group  design,  measure   was   done   at   one   occasion   which   avoids   threats   to   internal   validity   such   as   maturation,   history,   regression   to   the   mean   and   instrumentation   threats.   Observer   bias  is  thought  to  be  low  due  to  design.  The  participants  selected  which  image  they   perceived   to   be   shown   for   longest   duration   by   themselves.   However,   demand   characteristics  could  have  affected  participant  responses.  The  participant  was  not   aware  of  the  study’s  purpose  but  could  possibly  guess  that  there  was  a  difference   between  images  within  pairs.  On  the  other  hand,  the  question  was  precise  and  no   other  evaluation  was  made  beyond  perceived  duration  judgments.     Durations   were   tested   before   experiment   was   conducted   to   establish   suitable   durations.   No   ceiling   and   floor   effects   were   identified   for   durations   around   1000   ms.   For   example,   if   durations   had   been   longer,   the   high   complex   image   would   possibly  also  be  perceived  as  fluent,  thus  creating  a  floor  effect.  Previous  research   has   shown   that   longer   duration   is   correlated   with   higher   aesthetic   evaluations   (Reber,   et   al.   1998),   which   is   linked   to   higher   perceptual   fluency.   The   mere   exposure  effect  is  also  thought  to  enhance  objects  perceptual  fluency  (Seamon,  et   al.   1983).   This   result   is   interpreted   as   that   longer   durations   could   enhance  

9

perceptual   fluency.   One   possible   weakness   regarding   the   experiments   design   is   that  the  duration  for  the  fixation  point  was  closely  to  target  durations.  This  could   possibly  have  affected  duration  judgments.         In   this   study   the   dependent   variable   duration   judgments   was   measured   in   relation   to   two   items.   The   question   was,   which   was   presented   for   longest   duration.   This   measurement   design   was   regarded   as   sufficiently   sensitive   to   measure   the   dependent   variable.   It   is   regarded   as   a   precise   and   reliable   measurement,   which   is   indicated   by   participant’s   ability   to   correctly   identify   longest   duration   images   (M   =   17.46,  SD  =  3.02)  compared  to  levels  of  chance  (M  =  14).       This   study   was   working   within   the   domain   of   the   retrospective   duration   judgment   paradigm  (Block  &  Zakay,  1997),  when  participants  know  that  duration  judgments   are   of   importance   before   duration   has   started.   Block   and   Zakay   (1997)   found   significant  results  indicating  that  processing  difficulty  affected  duration  judgments,   as   processing   difficulty   increases,   experience   of   duration   decreases.   This   result   was   in   line   with   the   main   hypothesis   that   more   fluency   would   affect   duration   judgment   to   be   perceived   as   longer.   Or   phrased,   more   complexity   (less   fluency)   would  decrease  the  perceived  duration.         There  is  also  research  with  variables  that  could  be  interpreted  as  similar,  to  some   extent   with   perceptual   fluency,   and   processing   difficulty,   showing   results   in   opposite   direction.   Ahn,   Liu,   and   Soman   (2009)   found   results   in   their   study   indicating   that   more   richness   in   given   situation,   being   characterized   by   vivid   stimuli,  seems  to  elicit  a  notion  of  time  passing  quicker  yet  feel  as  longer  at  recall.   This   is   thought   to   be   due   to   the   number   of   memory   markers   are   large.   However,   recall  was  made  after  a  much  larger  amount  of  time.       Practical   implications   for   knowledge   regarding   processing   fluency   and   duration   judgments   in   particular,   is   considered   to   be   of   importance   for   example   advertisement   when   durations   are   limited,   or   for   commercial   needs   when   the   perception   of   time   wishes   to   be   manipulated,   during   queuing   for   example.   Other   practical   implications   could   be   for   traffic   signs   when   duration   is   mostly   short.   Further  research  building  on  this  work  could  investigate  whether  other  variables,   such  as  contrast,  clarity  and  symmetry  could  affect  duration  judgments.       The   overall   conclusion   from   this   study   is   that   fluency,   operationalized   as   visual   complexity,   does   not   affect   duration   judgments   during   short   presentations   around   1000   ms.   On   the   other   hand,   it   cannot   be   excluded   that   a   larger   manipulation   of   fluency,  with  more  variables  manipulated,  would  not  affect  duration  judgments.              

10

References     Ahn,   H.,   Liu,   M.,   &   Soman,   D.   (2009).   Memory   markers:   How   consumers   recall   the   duration   of   experiences.   Journal   Of   Consumer   Psychology,   19(3),   508-­‐516.   HYPERKINK   “http://dx.doi.org/10.1016/j.jcps.2009.05.002”     Avni-­‐Babad,   D.   &   Ritov,   I.   (2003).   Routine   and   the   Perception   of   Time.   Journal   Of   Experimental   Psychology:   General,   132(4),   543-­‐550.   HYPERLINK   “http://dx.doi.org/10.1037/0096-­‐ 3445.132.4.543”     Block,  R.,  Hancock,  P.,  &  Zakay,  D.  (2010).  How  cognitive  load  affects  duration  judgments:  A  meta-­‐ analytic   review.   Acta   Psychologica,   134(3),   330-­‐343.   HYPERLINK   “http://dx.doi.org/10.1016/j.actpsy.2010.03.006”     Block,   R.   &   Zakay,   D.   (1997).   Prospective   and   retrospective   duration   judgments:   A   meta-­‐analytic   review.   Psychonomic   Bulletin   &   Review,   4(2),   184-­‐197.   HYPERLINK   “http://dx.doi.org/10.3758/bf03209393”     Brinol,  P.,  Petty,  R.,  &  Tormala,  Z.  (2006).  The  Malleable  Meaning  of  Subjective  Ease.  Psychological   Science,  17(3),  200-­‐206.  HYPERLINK  “http://dx.doi.org/10.1111/j.1467-­‐9280.2006.01686.x”     Checkosky,  S.,  &  Whitlock,  D.  (1973).  Effects  of  pattern  goodness  on  recognition  time  in  a  memory   search   task.   Journal   Of   Experimental   Psychology,   100(2),   341-­‐348.   HYPERLINK   “http://dx.doi.org/10.1037/h0035692”     Garner,   W.   (1974).   The   processing   of   information   and   structure.   Potomac,   Md.:   L.   Erlbaum   Associates;  distributed  by  Halsted  Press,  New  York.       Hansen,  J.  &  Trope,  Y.  (2013).  When  time  flies:  How  abstract  and  concrete  mental  construal  affect   the  perception  of  time.   Journal  Of  Experimental  Psychology:  General,  142(2),  336-­‐347.  HYPERLINK   “http://dx.doi.org/10.1037/a0029283”     Lanska,   M.,   Olds,   J.,   &   Westerman,   D.   (2014).   Fluency   effects   in   recognition   memory:   Are   perceptual   fluency   and   conceptual   fluency   interchangeable?.   Journal   Of   Experimental   Psychology:   Learning,   Memory,  And  Cognition,  40(1),  1-­‐11.  HYPERLINK  “http://dx.doi.org/10.1037/a0034309”     Lavoie,   P.,   &   Grondin,   S.   (2004).   Information   processing   limitations   as   revealed   by   temporal   discrimination.   Brain   And   Cognition,   54(3),   198-­‐200.   HYPERLINK   “http://dx.doi.org/10.1016/j.bandc.2004.02.039”     Reber,   R.,   &   Schwarz,   N.   (2001).   The   hot   fringes   of   consciousness:   Perceptual   fluency   and   affect.   Consciousness  &  Emotion,  2(2),  223-­‐231.  HYPERLINK  “http://dx.doi.org/10.1075/ce.2.2.03reb”     Reber,   R.,   Schwarz,   N.,   &   Winkielman,   P.   (2004).   Processing   Fluency   and   Aesthetic   Pleasure:   Is   Beauty   in   the   Perceiver's   Processing   Experience?.   Personality   And   Social   Psychology   Review,   8(4),   364-­‐382.  HYPERLINK  “http://dx.doi.org/10.1207/s15327957pspr0804_3”     Reber,   R.,   Winkielman,   P.,   &   Schwarz,   N.   (1998).   Effects   of   Perceptual   Fluency   on   Affective   Judgments.   Psychological   Science,   9(1),   45-­‐48.   HYPERLINK   “http://dx.doi.org/10.1111/1467-­‐ 9280.00008”     Seamon,   J.,   Brody,   N.,   &   Kauff,   D.   (1983).   Affective   discrimination   of   stimuli   that   are   not   recognized:   Effects  of  shadowing,  masking,  and  cerebral  laterality.  Journal  Of  Experimental  Psychology:  Learning,   Memory,   And   Cognition,   9(3),   544-­‐555.   HYPERLINK   “http://dx.doi.org/10.1037/0278-­‐ 7393.9.3.544”    

11

Snodgrass,   J.,   &   Vanderwart,   M.   (1980).   A   standardized   set   of   260   pictures:   Norms   for   name   agreement,   image   agreement,   familiarity,   and   visual   complexity.   Journal   Of   Experimental   Psychology:   Human   Learning   &   Memory,   6(2),   174-­‐215.   HYPERLINK   “http://dx.doi.org/10.1037//0278-­‐7393.6.2.174”

12

13