Inspection

2
Interac(ve*form*1
Interac(ve*Form:*
Inspec(on*methods
• Select$an$informa2on$kiosk$or$machine
• Describe$it´s$purpose,$content$and$form
• Perform$a$cogni(ve*walkthrough
– describe$the$result
Eva$Ragnemalm,$IDA
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
3
4
Interac(ve*form*2
Goals*of*usability*evalua(on
• Select$a$webGbased$map$service
• Perform$am$analysis$of$it´s$quali2es$in$use$(according$to$
lecture$3)$and$describe$the$result
• Perform$a$heuris(c*evalua(on$and$describe$the$result
• Are$we$evalua2ng$in$order$to$improve$the$design?
– Forma2ve$evalua2on
– Focus$on$describing$problems$and$sugges2ng$improvements
• Are$we$evalua2ng$in$order$to$determine$if$the$prototype/
system$fulfills$the$customer’s$requirements$and$can$be$
delivered?
– Summa2ve$evalua2on
– Focus$on$measuring$proper2es$that$appear$in$the$requirements
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
5
6
Type*of*evalua(on
When*to*use*what?
• Inspec2on
• Applicable$tes2ng$method$vs$system$representa2on
– no$user$involved,$employ$usability$experts
• User$tes2ng
Forma2ve
Forma2ve
Sketches,
scenarios,
storyboards
Paper$prototypes
Forma2ve
Summa2ve$
(Forma2ve)
– set$users$to$interact$with$the$ar2fact$and$study
• Both$can$be$used$forma2vely$and$summa2vely
Inspec2on$
methods
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Computer$
prototypes
Working$system
User
tes2ng
User
tes2ng
User
tes2ng
Inspec2on$
methods
Inspec2on$
methods
Inspec2on$
methods
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
7
8
Inspec(on*methods
Cogni(ve*walkthrough
• Cogni2ve$walkthrough
• Heuris2c$Evalua2on
• Originally$designed$for$walkGupGandGuse$systems$(such$as$
informa2on$kiosks,$teller$machines$etc).
• Based$on$the$idea$that$the$user$has$a$goal,$looks$for$an$
ac2on$that$appears$to$lead$toward$that$goal,$performs$the$
ac2on$and$evaluates$if$the$result$did$bring$him/her$closer$to$
that$goal$(repeats$un2l$goal$is$achieved.
• Proposed$1990$(C.$Lewis,$P.$Polson,$C.$Wharton$and$J.$
Rieman)$revised$and$extended$into$several$versions
• Done$in$a$group$(developers$and$possibly$users),$or$
informally,$alone.
• Ac2on$Analysis$(backGofGtheGenvelope$version)
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
9
Cogni(ve*walkthrough*method
10
For*each*step*in*the*correct*sequence*of*
ac(ons,*tell*the*story*and*ask:
• Prepara2ons:
– Who$are$the$users$(what$is$familiar$to$them?$What$do$they$
expect?)
– Detailed$descrip2on$of$the$task$(or$tasks)$to$focus
– List$of$the$correct$sequence$of$ac2ons$the$task$requires
• For$each$step$in$the$sequence$of$ac2ons,$analyse$if$the$user$
is$going$to$find$and$execute$that$correct$step$(and$
understand$that$it$was$correct).
– tell$a$believable$story$(check$to$see$it$is$believable)
– if$no$believable$story$is$found,$you$have$a$problem$in$the$design.
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
• Will$the$user$be$trying$to$achieve$the$effect$of$the$step?
• Will$the$user$no2ce$the$control$(bucon/ac2on/...)$is$
available?
• If$the$user$no2ces$the$control,$will$they$understand$that$it$
will$lead$to$the$desired$effect?
• If$the$right$ac2on$is$performed,$will$the$user$understand$
that$progress$is$made$toward$the$goal?
• For$any$ques2on$that$is$answered$in$the$nega2ve,$focus$on$
why$not.
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
11
12
Result*of*cogni(ve*walkthrough
• Unbelievable$story,$doubdul$assump2on$or$ques2on$
answered$in$nega2ve$indicate$problems.$
• List$or$problems$at$specified$steps$in$the$sequence.$The$why$
should$provide$clues$to$how$to$solve$it.
• Examine$the$assump2ons$underlying$the$story$told.$Having$
several$experts$in$a$group$makes$finding$assump2ons$
easier.
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
13
14
Important
For*Interac(ve*Form*1:
• Don’t$combine$the$finding$of$the$correct$sequence$of$
ac2ons$into$the$walkthrough$itself!$
• Decide$on$a$target$to$inspect.
• Determine$your$users,$describe$them.
• Select$and$describe$(in$detail)$the$task$you’re$looking$at.
• Then$divide$into$pairs$or$triads,$perform$the$analysis$(using$
the$ques2ons$listed)$separately
• Combine$your$results$for$the$group$report
– If$the$walkthrough$shows$at$some$step$the$user$won’t$find$the$
correct$ac2on,$you$drop$that$problem$there$and$con2nue$to$the$
next$step$as$if$the$previous$step$had$been$correct.
• You$are$not$a$user!$You$may$not$find$everything,$and$you$
may$find$things$the$user$does$not$find$problema2c.
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
15
16
Heuris(c*Evalua(on
Heuris(c*evalua(on*method
• Developed$by$Jacob$Nielsen$and$Rolf$Molich$1990
• ”Discount$usability$engineering”
• Experts$are$presented$with$an$interface$design$and$inspect$
it.
• Requires$aggrega2ng$the$results$from$3$to$5$experts
• Based$on$Heuris2cs$G$guidelines$or$rules$of$thumb$for$how$
to$design$interac2ve$systems
• Select$set$of$heuris2cs$to$use
• The$usability$experts$receives$the$representa2on$of$the$
system$(prototype/specifica2on/scenario...)
• Expert$explores$system$to$become$familiar
• Expert$systema2cally$checks$system$against$heuris2cs$and$
notes$discrepancies$and$problems
• Collect$and$analyse$data
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
17
Nielsen’s*10*usability*heuris(cs
1. Simple$and$natural$
dialogue
2. Speak$the$user’s$language
3. Minimize$memory$load
4. Consistency$and$
standards
7. Efficiency$and$shortcuts
8. Good$error$messages
9. Prevent$errors
10.Provide$help$and$
documenta2on
5. Provide$feedback
6. Provide$clear$exits
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
18
Result:*uncovered*poten(al*problems
• Situa2ons$where$the$system$does$not$follow$the$principles
– ex:$loading$of$screen$3$might$take$2me,$give$feedback$to$show$
work$is$progressing
– ex:$screen$5$is$labeled$Wri2ng$while$the$associated$menue$is$called$
prin2ng.
• Best$results$if$experts$are$familiar$with$the$domain$as$well$
as$the$principles.
• Note:$all$iden2fied$problems$are$not$perceived$as$problems$
by$a$user
• Note:$The$experts$will$not$uncover$all$problems
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
19
20
How*many*evaluators*to*employ?
Collec(ng*and*analysing*data
• Not$just$number$of$errors$or$number$of$experts$finding$it
• Grade$problems$on:$
– Frequency$(how$ooen$will$the$problem$occur?)
– Effect$(how$difficult$is$it$for$the$user$to$cope$with)
– Persistency$(is$it$only$a$problem$the$first$2me$or$every$2me)
• Degree$of$seriousness$(summa2ve/forma2ve)
– 0:$It$is$not$a$problem$even$though$the$heuris2c$is$not$followed
– 1$:cosme2c$problem,$not$necessary$to$adress$unless$available$2me
– 2:$lesser$problem,$low$priorty
– 3:$bigger$problem,$high$priority
– 4:usability$catastrophe,$must$be$adressed$before$release
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
21
Number*of*evaluators
22
Who*should*do*the*inspec(on?
• Heuris2c$Evalua2on$literature:$don’t$evaluate$your$own$
design
– You$already$know$everything$about$your$own$design.$
– If$you$knew$the$principles$you$followed$them$and$if$you$didn’t$
know$them$you$are$not$an$expert.
• Cogni2ve$walkthrough$literature:$you$can$do$your$own$CW
– keep$in$mind$the$goal$is$to$develop$your$design,$not$validate:$focus$
on$finding$things$to$improve.
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
23
For*Interac(ve*Form*2
• Number$of$experts=number$of$group$members
• To$become$experts:$go$through$the$heuris2cs,$discuss$in$
groups$how$to$interpret.
• Agree$on$what$part$of$the$system$you$are$inspec2ng$
(exactly$what$screens$and$what$sequence)
• When$inspec2ng:$for$each$screen$(and$each$transi2on$
between$screens)$go$through$all$the$principles,$write$down$
problems$(refer$to$the$screen$and$the$principle)
• Aggregate$the$problems$found,$grade$and$report
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
24
Research*project:*Heuris(c*evalua(on*of*
persuasive*capacity*of*soLware
• Design$of$persuasive$systems$(behaviourGchange$systems)
– help$people$exercise$more,$eat$healthier,$stop$smoking$etc
• Guidelines$for$persuasive$systems$developed$by$BJ$Fogg$
(2005)$ex:
– be$trustworthy,$correct,$2mely,$provide$sugges2ons,$assess$the$
user’s$behaviour,$make$it$easy$to$do$the$right$thing$...
• Educa2onal$systems$that$people$do$learn$from$must$be$
persuasive!$
• Then:$Is$a$game$persuasive?$A$computer$tool$like$Word?
Interac2ve$Form:$inspec2on$methods/Eva$Ragnemalm
Ques%ons?
www.liu.se