Using Rationale to Assist Student Cognitive and Intellectual Development
Full text
An In e disciplina y Jou nal on Humans in ICT En i onmen s ISSN: 1795-6889
www.human echnology.jyu. i Volume 6 (1), May 2010, 106–128
106
USING RATIONALE TO ASSIST STUDENT COGNITIVE
AND INTELLECTUAL DEVELOPMENT
Abs ac : One o he ques ions posed a he Na ional Science Founda ion (NSF)-
sponso ed wo kshop on C ea i i y and Ra ionale in So wa e Design was on he ole o
a ionale in suppo ing idea gene a ion in he class oom. College s uden s o en s uggle
wi h p oblems whe e mo e han one possible solu ion exis s. Pa o he di icul y lies in
he need o s uden s o p og ess h ough di e en le els o de elopmen cogni i ely and
in ellec ually be o e hey can ackle c ea i e p oblem sol ing. A gumen a ion-based
a ionale p o ides a na u al mechanism o ep esen ing p oblems, candida e solu ions,
c i e ia, and a gumen s ela ing hose c i e ia o he candida e solu ions. Explici ly
exp essing a ionale o hei wo k encou ages s uden s o e lec on why hey made hei
choices, and o ac i ely conside mul iple al e na i es. We epo on an expe imen
pe o med du ing a Da a S uc u es cou se whe e s uden s cap u ed a ionale.
Keywo ds: design a ionale, c ea i i y, s uden cogni i e de elopmen .
RATIONALE AS A METHOD FOR BUILDING
CREATIVITY AND COGNITIVE MATURITY
One o he o ien ing ques ions o he Na ional Science Founda ion (NSF)-sponso ed
wo kshop on C ea i i y and Ra ionale in So wa e Design in 2009 was, “How can design
a ionale be used in he class oom o mo i a e and ins uc s uden s abou e lec ion, idea
gene a ion, and e alua ion?” (Daugh y, Bu ge, Ca oll, & Po s, 2009). A he hea o his
ques ion is an implici claim abou c ea i i y, ha is, “c ea i i y” in so wa e design seems o
in ol e no jus idea gene a ion i sel , bu also he i e a i e p ocess ha mo es he designe o
e lec , e alua e, and gene a e mo e ideas mul iple imes be o e commi ing o a inal design.
Ca oll’s (2009) wo kshop mani es o, “The Essen ial Tension o C ea i i y and Ra ionale
in So wa e Design,” emphasized his by poin ing o liminali y as a key aspec o he c ea i e
design p ocess. The mani es o desc ibed liminali y as “Thinking and ac ing on he bo de
be ween wo con as ing concep s o ules, such as apid swi ching be ween con e gen and
di e gen modes o hinking.”
© 2010 Jane E. Bu ge & Bo B inkman, and he Ago a Cen e , Uni e si y o Jy äskylä
URN:NBN: i:jyu-20105241909
Jane E. Bu ge
Depa men o Compu e Science
and So wa e Enginee ing
Miami Uni e si y, Ox o d, Ohio
USA
Bo B inkman
Depa men o Compu e Science
and So wa e Enginee ing
Miami Uni e si y, Ox o d, Ohio
USA
Ra ionale Assis ing S uden Cogni i e De elopmen
107
We see a di ec link be ween a s uden ’s cogni i e de elopmen and he abili y o engage in
c ea i e p ocesses. Pe y (1970) iden i ied nine posi ions o de elopmen s a ing wi h duali y,
whe e answe s exis o e e y hing and whe e hey can be igh o w ong, in o mul iplici y,
whe e all answe s a e alid, in o ela i ism, whe e hey begin o e alua e solu ions based on he
con ex , and con inuing h ough se e al le els o commi men , whe e s uden s can begin o
in eg a e knowledge and make hei own choices based on ha in o ma ion.
S uden s in he i s wo yea s o college end o display dualis ic and mul iplis ic
endencies. Though Pe y poin ed ou ha mos college s uden s a e no pu e S age 1 dualis s,
ew s uden s in his s udy eached e en he lowes le els o commi men un il hei junio yea (p.
155). Simila ly, Ma cia B. Bax e Magolda epo ed (1992, p. 71) ha mo e han 80% o junio s
a e “ ansi ional knowe s,” ( hose ha ecognize ela i ism in some knowledge domains, bu a e
s ill dualis ic in o he s), and mo e han 40% o sophomo es we e s ill mos ly dualis ic.
This unde s anding o he epis emic s yles o ou s uden s should in o m ou hinking
abou eaching design. Dualis ic cogni ion is inhe en ly opposed o he liminal s a e o mind
ha is so cha ac e is ic o c ea i i y in design. We belie e ha s uden s who come o a design
p oblem wi h he a i ude ha he e is a “ igh answe ” o be disco e ed by analysis will
commi o a design wi hou engaging in e lec ion o i e a ion. They will commi oo ea ly,
be o e hey ha e a chance o be c ea i e. S uden s in mul iplici y o he ea ly s ages o
ela i ism may be unable o dis inguish be ween he good designs and poo designs ha
eme ge in hei hinking. Some e idence o suppo hese claims can be ound in he wo k o
A man, Ca della, Tu ns, & Adams (2005), who showed ha senio enginee ing s uden s
spend 2 o 3 imes mo e ime on a design p oblem han eshman enginee ing s uden s. This
also co ela es highly wi h he quali y o he inal solu ion, hough hei esul s do no di ec ly
add ess ou claim ha hese e ec s a e due, in pa , o s uden epis emic s yles.
We p opose ha equi ing s uden s o gene a e design a ionale p io o implemen ing
hei solu ions is a mechanism o encou aging e lec ion and delaying commi men o hei
ini ial design choices. Design a ionale, he easons behind decisions made while designing,
is a way o ep esen design al e na i es and he delibe a ion ha p oduced hem. In a sense,
he a ionale can be conside ed a language o design (Dym, Agogino, E is, F ey, & Lie e ,
2005), much like ske ching (which cap u es s uc u al aspec s o design) o ma hema ics
(which exp esses cons ain s he design mus con o m o). In he case o a ionale, his is a
language ha cap u es he design in en and i s ela ionship o he design. The abili y o
analyze and e alua e design al e na i es in e ms o hei success a achie ing design goals
(in en ) equi es highe o de hinking skills.
In esponse o ou o ien ing ques ion, we claim ha design a ionale help o mo i a e and
ins uc s uden s in he c ea i e p ocess by pu ing o he momen o commi men o a
design. The ime spen in he liminal phase o design, i e a ing om idea o e alua ion and
back, can be leng hened by he use o design a ionale. A p ospec i e design a ionale ( ha is,
design a ionale buil be o e implemen a ion, as a me hod o explo ing possible designs)
se es as a way o documen ing he designe ’s p ocess o design.
This leng hening o he ime he s uden spends in ambigui y and e lec ion should also
lead o cogni i e de elopmen , by o cing he s uden o expe ience he kinds o e lec ion
and swi ching be ween modes o hough ha a e cha ac e is ic o highe le els o cogni ion.
Bu ge & B inkman
108
In he es o his pape we explo e mo e ully he ollowing wo ques ions:
1) Wha a e he links be ween c ea i i y in design and cogni i e le el, and how can
a ionale assis in de eloping a s uden ’s capaci ies o each?
2) How can we assess whe he o no use o a ionale has had he in ended e ec ?
In he balance o his i s sec ion we explo e he i s ques ion. Fi s we desc ibe he
mo i a ion o eaching c ea i i y in so wa e de elopmen , and hen expand on ou p oposi ion
ha “liminali y” links so wa e design wi h cogni i e de elopmen . Nex we discuss he use o
design a ionale as a pedagogical ool o encou aging cogni i e de elopmen h ough
e lec ion, and desc ibe some p io applica ions o a ionale o educa ion.
La e in he pape we desc ibe an expe imen al assignmen we designed based on ou
ideas, and p o ide some ini ial assessmen s o ou app oach. In he inal wo sec ions we
ou line a eas o u u e wo k and o he ways ha design a ionale may be used o s imula e
s uden cogni i e de elopmen .
C ea i i y in So wa e De elopmen
So wa e de elopmen is, a i s co e, a c ea i e en e p ise. Gi en a p oblem, he e a e many
possible solu ions. Fo some p ac i ione s, his is wha a ac s hem o he ield—so wa e
de elopmen as an exe cise in c ea i e design. Fo o he s, especially as college s uden s, he
mul i ude o solu ions, whe e he e is o en no clea “ igh ” answe , can be a sou ce o
us a ion. Wi h he many demands on hei ime, bo h cu icula and ex acu icula , he e is
signi ican p essu e o ind he, o a, co ec solu ion in as li le ime as possible. The skill o
being able o unde s and jus enough abou he ma e ial o come up wi h an answe se es
hem well in some o hei ea lie cou ses, whe e a p og am is co ec i i p oduces he
co ec se o ou pu s gi en a se o inpu s. Bu hey un in o di icul y in hei la e cou ses,
whe e solu ions need o be analyzed on mul iple dimensions. These di icul ies a e
exace ba ed in cou ses such as So wa e Enginee ing and Human–Compu e In e ac ion,
whe e he sys em design is in luenced no only by he echnology a ailable bu by how
people in end o use i .
I is essen ial ha compu e scien is s, and compu e science s uden s, hink c ea i ely in
o de o success ully de elop so wa e. Glass (1995) desc ibed se e al aspec s o so wa e
de elopmen whe e c ea i i y is c i ical: de e mining how o ansla e he cus ome /business
needs in o a p oblem ha he so wa e can sol e; esol ing s akeholde con lic s; designing
solu ions o new and complex p oblems; de e mining es cases; and enhancing exis ing
sys ems o mee needs ha we e no ini ially an icipa ed by he cus ome o he de elope s. A
s uden con inced o a single igh answe is likely o ei he insis ha he s akeholde (s)
p o ide his answe (when he s akeholde s may no be app oaching he p oblem wi h an
awa eness o wha is possible wi h he echnology a ailable) o insis ha hei solu ion is he
only one, o he bes one, e en i i may no be accep able o he clien .
Liminali y, C ea i i y, and Cogni i e De elopmen
The wo kshop mani es o (Ca oll, 2009) emphasized h ee majo cha ac e is ics o c ea i i y
in so wa e design: play ulness, empa hy, and liminali y. We ha e chosen o ocus on he
Ra ionale Assis ing S uden Cogni i e De elopmen
109
liminal aspec s o c ea i i y because i seems o be he mos na u al i o eshman- and
sophomo e-le el cou ses.
To be su e, many ins uc o s ha e g ea luck inco po a ing play ul o empa hic
app oaches in hei cou sewo k; many such assignmen s a e p esen ed e e y yea a he
SIGCSE (Special In e es G oup on Compu e Science Educa ion) con e ence. Bu , o mos
s uden s, Da a S uc u es is he i s equi ed cou se ha explici ly eaches a se o
ma hema ical ools ha can be used o compa e one solu ion o ano he . He e we a e speaking
o he use o asymp o ic analysis o compa e he ime and space equi emen s o da a
s uc u es. When applied o simple p oblems, like so ing, such analyses seem de ini i e: Fo
example, “Randomized Quickso is mo e e icien han Inse ion So .” Bu when designing
a da a s uc u e o a ealis ic p oblem, i is o en he case ha some ope a ions can only be
made as i o he ope a ions a e made slow, o i excessi e amoun s o memo y a e used, o
i auxilia y da a s uc u es a e used o bookkeeping. This means ha , as a da a s uc u e is
designed, he e a e many oppo uni ies o shi ocus om one ope a ion o ano he , and o
shi om analysis o idea gene a ion and back. The mani es o links his “ apid swi ching
be ween con e gen and di e gen modes o hinking” o c ea i i y.
A conc e e example will help cla i y ou poin . In he expe imen al assignmen desc ibed
mo e ully below, s uden s we e asked o design a lis -like da a s uc u e ha needed o
suppo dequeue ope a ions (adding and emo ing i ems a he ends) as well as sea ching by
key. One s uden , in his ini ial hinking, conside ed only a ays and linked-lis s as possible
designs, and selec ed linked-lis s because hey suppo dequeue ope a ions in cons an ime.
Upon e alua ion o he designs, howe e , he disco e ed ha sea ch would be e y slow, and
so he e u ned o idea gene a ion and added hash ables as a hi d design op ion. Upon
e alua ing hash ables he disco e ed ha he dequeue ope a ions would be icky o
implemen , and e u ned again o idea gene a ion.
Inspi ed by his example, we p opose ha a Da a S uc u es cou se is a na u al place o
look o he con as ing concep s ha gi e ise o liminal men al s a es. In Da a S uc u es we
each he heo y o algo i hm unning imes, bu also how o ac ually de e mine algo i hm
pe o mance h ough expe imen s o con i m (o no !) he heo y. We each he canonical da a
s uc u es, bu we also gi e s uden s p oblems o which he canonical da a s uc u es a e a poo
i . We p esen he ma e ial o he class using diag ams and pseudo-code, bu equi e s uden s o
ac ually w i e wo king p og ams using a eal language.
We do no wish o de ine c ea i i y only in e ms o liminali y, bu we eel ha much o
wha is c ea i e abou he wo k o s uden s eally a ises when hey a e able o syn hesize
seemingly incompa ible ideas om wo appa en ly opposing o un ela ed ways o hinking. In
e e ence o he mani es o (Ca oll, 2009), we claim ha s uden s a e bes able o “pu sue
su p ise and unexpec ed ou comes” when hey ac i ely emb ace and explo e he “bo de
be ween con as ing ideas.”
We belie e he abili y o emb ace liminal s a es and cogni i e de elopmen a e di ec ly
linked. Many use ul heo ies o cogni i e de elopmen migh in o m his discussion. We ha e
al eady desc ibed he key aspec s o Pe y’s (1970) model, which unde gi ds much o ou
hinking in hese ea ly sec ions. In ou inal sec ion, we also use Bloom’s Taxonomy (Bloom,
1956), which we ound help ul in iden i ying o he pedagogical applica ions o design
a ionale. The e idence o Pe y (1970, p. 55–56) and Bax e Magolda (1992, p. 71) sugges s
ha ou Da a S uc u es s uden s (who a e mos ly sophomo e compu e science majo s and
Bu ge & B inkman
110
junio enginee ing majo s) will s ill be in ansi ion owa ds ela i ism. Bax e Magolda’s
s udy showed ha mo e han 40% o sophomo es we e s ill no iceably absolu e in hei
hinking, and ha e y ew junio s (less han 10%) a e independen hinke s. In Pe y’s s udy
junio s we e a ed as being in “commi men ” (le els 7, 8 o 9) only abou 50% o he ime,
and o sophomo es i was less han 10%.
S uden s s uck in a dualis ic way o hinking a e unlikely o disco e c ea i e solu ions,
because hey will be sa is ied as soon as hey iden i y any “co ec ” solu ion. The aps o
s uden s in mul iplici y o naï e ela i ism a e sub le . A his le el, he s uden is awa e ha
he e a e many iable solu ions, bu ends o assume ha all a e equally good. This can again
block c ea i i y because he s uden chooses a solu ion somewha a bi a ily. When s uden s
a e “s uck” a he lowe le els o cogni i e de elopmen , we suspec ha he solu ion chosen
is likely o be ou ine, amilia , o a bi a y, a he han inno a i e and c ea i e.
So we p opose ha he e is a link be ween com o wi h liminal men al s a es and
cogni i e ma u i y, and ha design ac i i ies ha cause s uden s o expe ience apid swi ching
be ween con as ing ideas help s uden s o build up bo h cogni i e and c ea i e ma u i y.
Ra ionale, Re lec ion, and Liminali y
In he expe imen al assignmen sequence p esen ed in he nex sec ion, we used p ospec i e
design a ionale o encou age s uden c ea i i y in an indi idual design ask. As men ioned
abo e, a p ospec i e design a ionale is one ha is c ea ed be o e he design is implemen ed,
as pa o he design p ocess. Con as his wi h e ospec i e design a ionale, which a e
w i en a e he design is chosen, and may se e only o documen he chosen design.
P ospec i e design a ionale os e s bo h c ea i i y and cogni i e de elopmen by
encou aging, and cap u ing e idence o , e lec ion.
Re lec ion se es an impo an pu pose in bo h educa ion and in p ac ice. In educa ion,
many esea che s ha e p oposed a link be ween e lec ion and cogni i e/epis emic le el.
Dewey (1933, p. 9) de ined e lec i e hinking as “ac i e, pe sis en , and ca e ul conside a ion
o any belie o supposed o m o knowledge in he ligh o he g ounds ha suppo i and he
u he conclusions o which i ends.” Re lec ion guides he lea ning p ocess as e idence is
examined and conclusions d awn. Dewey’s claim ha e lec i e hinking is necessa y when i
is no possible o come up wi h “ce ain solu ions” was he eason why King and Ki chene
(2002) chose e lec ion as he basis o hei model o s uden epis emological de elopmen .
The e lec i e judgmen model (King & Ki chene , 1994) de ined se en s ages o s uden
epis emological de elopmen , b oken in o h ee ca ego ies: p e e lec i e hinking, quasi-
e lec i e hinking, and e lec i e hinking.
Schön, in his book The Re lec i e P ac i ione (1983), desc ibed he need o
p o essionals o mo e beyond echnical a ionali y, whe e p oblem sol ing is he applica ion
o heo y, o p ocesses ha allow o unce ain y and con lic . He desc ibed “knowing-in-
ac ion,” whe e p ac i ione s ac based on aci knowledge, and “ e lec ion-in-ac ion,” whe e
p ac i ione s e lec on wha hey a e doing as hey do i .
Fische , Lemke, McCall, & Mo ch (1991) desc ibed how design a ionale suppo s
e lec ion by cap u ing he designe ’s knowledge abou he si ua ion. Simila ly, he i e a ion
be ween idea gene a ion (di e gen hinking) and design selec ion (con e gen hinking) is a
e lec i e p ocess. Design a ionale suppo s bo h he cap u e o he al e na i es and hei
Ra ionale Assis ing S uden Cogni i e De elopmen
111
explo a ion by suppo ing he e alua ion o he mo e p omising al e na i es and any addi ional
decisions equi ed du ing hei elabo a ion. In he illus a i e example abo e we saw a s uden
using he p ocess o building design a ionale as an oppo uni y o c i ical e lec ion.
We should no e, as an aside, ha in his s udy we ocus only on he indi idual design
p ojec s, no on eamwo k. Though we g ea ly app ecia e he ole ha a ionale can play in
cap u ing and ans e ing knowledge in a eam se ing, we belie e ha he cap u e o
a ionale is bene icial e en o one indi idual engaged in an indi idual design p ojec .
So we claim ha design a ionale can be used o encou age c i ical e lec ion abou
so wa e design p oblems. Fu he , we claim ha such c i ical e lec ion, i emb aced by he
s uden , is likely o lead o g ea e c ea i i y. C i ical e lec ion and c ea i i y a e ce ainly
no he same hing; a he , c i ical e lec ion ends o p o ide g is o c ea i e ene gies o ac
upon. Inco ec assump ions end o ac as oadblocks o c ea i i y, bu c i ical e lec ion
helps us o challenge hese assump ions. We na u ally end o selec designs simila o olde
success ul designs wi h which we a e al eady com o able, bu c i ical e lec ion can cause us
o ejec amilia solu ions ha a e ac ually inapp op ia e.
P io Wo k on Ra ionale in Educa ion
Mo an and Ca oll’s (1996) book included wo app oaches o using a ionale in educa ion.
The i s was o p o ide a ionale in he o m o empla es o assis wi h use in e ace (UI)
design (Casaday, 1996). The empla es help designe s o “ask he igh ques ions” and assis
designe s wi h he p ocess by guiding hem owa d a solu ion. Ca ey, McKe lie, & Wilson
(1996, p. 375) buil a lib a y o “exempla y use -in e ace designs” along wi h hei a ionale
so hose examples could be used o each UI design. O he wo k using a ionale in UI design
includes using design space analysis (DSA; MacLean, Young, Bello i, & Mo an, 1991) as
pa o he FLUID ( amewo k o lea ning use in e ace design) in e ac i e media sys em
( an Aals , an de Mas , & Ca ey, 1995). The wo k p oposed he e uses a mo e gene al
app oach (no one aimed a a speci ic ype o design) and suppo s addi ional manipula ion
and e alua ion o design c i e ia, as well as using a ionale o assis wi h he de ini ion and
documen a ion o new designs.
Se e al so wa e enginee ing ex books ei he each a ionale (B uegge & Du oi , 2004) o
use a ionale as explana ion o design case s udies (Fox, 2006). Ra ionale is also p esen in he
o m o “consequences” in he ubiqui ous Gang o Fou (GoF) design pa e ns book (Gamma
Helm, Johnson, & Vlissides, 1995), used bo h as a e e ence and as a supplemen al ex book.
EXPLORING RATIONALE IN A DATA STRUCTURES COURSE
In he p e ious sec ion we claimed ha ca e ul use o design a ionale by dualis ic and
mul iplis ic hinke s should lead hem o inc eases in c ea i i y and cogni i e ma u i y. This
heo y has implica ions o how one s uc u es “design” p ojec s o lowe -le el cou ses. In
his sec ion we will p esen a i s a emp a such an assignmen o a Da a S uc u es cou se,
and con as i wi h he kinds o design assignmen s we had used in he pas .
Ou heo y also equi es some jus i ica ion h ough e idence. We ha e some ini ial esul s
based on ou e alua ion o he wo k p oduced by s uden s o ou expe imen al assignmen
Bu ge & B inkman
112
using design a ionale. While he expe imen was by no means a con olled expe imen , no
was i designed o alida e ou heo y (i was, ins ead, designed o help he s uden s lea n), we
s ill a e able o epo on some an alizing esul s ha poin he di ec ion o u u e wo k.
SEURAT and Pugh’s To al Design in Da a S uc u es
In he Da a S uc u es cou se, we chose o use wo di e en me hods o cap u ing p ospec i e
design a ionale. The i s was he a ionale managemen sys em SEURAT and he second was
based on examples om Pugh’s (1991) To al Design. In o de o he eade o unde s and how
we hink a ionale should be used in unde g adua e cou ses, we mus i s desc ibe wha da a
hese wo ypes o design a ionale cap u e, and how hey suppo decision making.
Le us s a wi h some gene al obse a ions. P oblem sol ing can be b oken in o ou
s ages: p oblem de ini ion and analysis, idea gene a ion, idea e alua ion and selec ion, and
implemen a ion o he selec ed idea (VanGundy, 1981). Ra ionale can suppo some idea
gene a ion echniques, such as b ains o ming, by ep esen ing al e na i es as gene a ed, and
a ibu e lis ing, a echnique de eloped by C aw o d (VanGundy, 1981), whe e a ibu es
lis ed would be al e na i es. Ra ionale cap u ed in he o m o a gumen a ion is especially
use ul, howe e , du ing he e alua ion and selec ion s age by cap u ing c i e ia, hei
ela ionship o he al e na i es, and suppo ing e alua ion. Some o he echniques desc ibed
by VanGundy (1981) ha could be suppo ed by a ionale a e (a) he ad an age–disad an age
app oach, enume a ing he ad an ages/disad an ages o each al e na i e wi h espec o a
p ede ined se o c i e ia; (b) he Ba elle me hod (Hamil on, 1974; VanGundy, 1981),
di iding c i e ia in o culling, a ing, and sco ing in o de o na ow he ield o al e na i es;
and (c) e e se b ains o ming (VanGundy, 1981; Whi ing, 1958), which is b ains o ming on
he disad an ages o each al e na i e. Ra ionale sys ems ha pe o m e alua ion, such as he
so wa e enginee ing using a ionale (SEURAT) sys em (Bu ge & B own 2004), can be
conside ed a ype o weigh ing sys em (VanGundy, 1981), by allowing weigh s o be
assigned o he c i e ia and using hose weigh s in e alua ion.
In his wo k, we use a gumen a ion-based a ionale o cap u e he idea gene a ion, idea
e alua ion, and selec ion s ages o p oblem sol ing. We used wo me hods o ep esen ing
a ionale, SEURAT ( o one expe imen al g oup) and w i en documen s p oposed in Pugh’s
o al design me hodology ( o he o he ). Bo h me hods equi e s uden s o lis many al e na i e
designs, de elop c i e ia by which o e alua e he designs, pe o m he e alua ion, and selec a
solu ion. Bo h me hods, u he mo e, equi e a gumen a ion o back up bo h he e alua ion
c i e ia and he inal decision. SEURAT adds he addi ional capabili ies o exp essing he
a ionale in a hie a chical o ma , showing decisions and subdecisions, as well as p o iding he
capabili y o calcula e a nume ical e alua ion o he suppo o each al e na i e.
So wa e Enginee ing Using RATionale (SEURAT)
The SEURAT sys em (Bu ge & B own, 2004) is a a ionale managemen sys em (RMS)
o iginally de eloped o assis wi h so wa e main enance by p o iding ways ha he a ionale
could be used beyond jus i s p esen a ion. SEURAT cap u es a ionale as s uc u ed
a gumen a ion (decision p oblems, decision al e na i es, and a gumen s) and uses bo h he
s uc u e o he a ionale (syn ax) and he con en (seman ics) o in e ence o e he a ionale
Ra ionale Assis ing S uden Cogni i e De elopmen
113
o de ec incomple eness (o he a ionale) and inconsis ency (o he design). The a gumen s
in SEURAT can e e o sys em equi emen s, desi ed quali ies, assump ions made, and
ela ionships be ween al e na i es. Figu e 1 shows some a ionale cap u ed in he SEURAT
Ra ionale Explo e . SEURAT s o es he a ionale in a ela ional da abase, allowing he
a ionales o be sha ed be ween mul iple use s du ing collabo a i e decision-making.
Figu e 1’s example shows h ee decisions, aken om he a ionale o a con e ence
oom scheduling sys em. The decisions a e displayed using a diamond-shaped icon
con aining a double-headed a ow. The second decision, “How do we know wha he
con e ence ooms a e?” has a wa ning icon o e laid on i . This is because he al e na i e
selec ed, “ascii ile gi ing a lis o oom names,” is no as well suppo ed as he o he
candida e al e na i e, “se ialized ec o o oom objec s.” The hi d decision, “How is he
use associa ed wi h he mee ing,” has an e o icon because none o he p oposed al e na i es
has been selec ed ye .
The s uden s who used SEURAT in he expe imen we e gi en a u o ial on how a ionale
a e en e ed in o SEURAT and how hey could use SEURAT’s abili y o e alua e al e na i es o
assis hem in hei decision-making. The s uden s we e ins uc ed o en e he unc ional and
non unc ional equi emen s ha applied o he p oblem hey we e sol ing and hen o en e he
decisions, al e na i es, and a gumen s. They we e ins uc ed o use hei equi emen s in
a gumen s, a he han u ilizing he o he ypes o a gumen s suppo ed by SEURAT.
Figu e 1. An image o he SEURAT Ra ionale Explo e , showing he hie a chical iew
o a s uc u ed design a ionale.
Bu ge & B inkman
114
Pugh’s To al Design
The o he sec ion o he class used a se o design documen s based on S ua Pugh’s (1991)
book To al Design. I is no p ope o say ha we used “his” documen s, because Pugh goes
o g ea leng hs o show many di e en ypes o documen s ha migh be use ul. In
pa icula , we ook ad an age o h ee main pa s o his app oach:
1. A p oduc design speci ica ion (PDS) lis ed he c i e ia by which designs should be
judged. Each s uden cons uc ed his o he own lis o c i e ia, esul ing in a
bulle ed lis wi h a gumen a ion ha was e y simila o he SEURAT g oup, bu
c ea ed in a wo d p ocesso ins ead o he SEURAT RMS.
2. A Pugh Ma ix was used o idea e alua ion. The s uden c ea ed a wo-
dimensional able in a wo d p ocesso . Each column co esponded o one o he
designs, and each ow o one o he c i e ia om he design speci ica ion. The
s uden selec ed one design o be he “baseline” design, and hen each design was
compa ed o he baseline in each c i e ion. A plus symbol was en e ed in he able
i he design in ques ion was supe io o he baseline on ha c i e ion, a minus
symbol i i was wo se. We also ins uc ed s uden s o en e hei a gumen s in
suppo o hei e alua ions in each cell.
3. A sho essay summa ized he idea selec ion phase and p o ided a gumen s in a o o
he s uden ’s solu ion, based on he e alua ion in he Pugh Ma ix. No e ha Pugh was
e y clea ha one should no jus coun up he numbe o plusses and minuses and hen
selec based on he nume ical answe ha esul s. We ins uc ed s uden s o use hei
e alua ion ma ix o suppo hei selec ion p ocess, bu o also use common sense.
Pugh’s p ocess has many commonali ies wi h SEURAT. Design c i e ia a e explici ly
lis ed, and equi e a gumen a ion. Designs a e e alua ed based on he c i e ia, and a gumen s
suppo ing hose e alua ions a e cap u ed. The Pugh p ocess suppo s a quasi-quan i a i e
app oach o selec ing he inal design.
The e a e majo di e ences as well. On he nega i e side, SEURAT equi es s uden s o
lea n a new ool ( he Ra ionale Explo e , which is a plug-in o he Eclipse In eg a ed
De elopmen En i onmen ) ins ead o using a amilia ool (Mic oso Wo d). On he
posi i e side, SEURAT o ces s uden s o be mo e ca e ul abou linking c i e ia wi h design
decisions. In he non-SEURAT g oup, some s uden s used a gumen s in hei Pugh Ma ix
ha had no ela ionship o hei c i e ia, some hing ha is much ha de o do in SEURAT.
SEURAT also na u ally leads one o ep esen subdecisions in a hie a chy unde he main
decisions, much like an ou line. The Pugh Ma ix places all decisions a he same le el.
Design in Da a S uc u es Be o e Design Ra ionale
Fo se e al yea s we ha e had design p ojec s in Da a S uc u es simila o he one desc ibed
he e. In he pas , howe e , s uden s simply submi ed a e ospec i e design jus i ica ion. These
documen s ook a a ie y o o ms, bu none o hem we e pa icula ly o mal, and only in e y
a e cases did he s uden s compose hem be o e implemen ing hei solu ion. We ound he
quali y o he esul ing p og ams w i en by he s uden s o be qui e disappoin ing. In pa icula ,
he e was some anecdo al e idence ha s uden s would no conside all o he impo an design
Ra ionale Assis ing S uden Cogni i e De elopmen
121
using SEURAT o s uden pe o mance when using he Pugh Ma ix me hod. I s uden s
using Pugh ma ices did subs an ially be e han s uden s using SEURAT, we migh
conclude ha SEURAT should no be used wi h younge s uden s. Howe e , his u ned ou
no o be he case: S uden s using SEURAT pe o med as well as, o be e han, s uden s
using Pugh ma ices in mos asks.
We used a wo-sided Mann-Whi ney U es o compa e expe imen al g oups (see Table 5).
Because we ha e no eason o belie e ha any o ou ub ics would co espond o a no mal
dis ibu ion, we el ha i would be unsound o use, o example, a - es , because i equi es
he sampled da a o be independen and no mally dis ibu ed. The Mann-Whi ney es does
no su e om his de ec because i wo ks by i s anking all samples, and hen e alua ing
he likelihood o he e being a ma ked di e ence in ank sum be ween he wo expe imen al
g oups. Mo e in o ma ion abou Mann-Whi ney U es s (also known as Wilcoxon ank sum
es s) may be ound in s a is ics ex books, such as Rice (1995, pp. 402-410). We se ou
h eshold o signi icance a he = 0.1 le el.
Table 5. Compa ison o SEURAT Use s o Ma ix Use s.
Al e na i e hypo hesis NS N
M S
S S
M Tes esul
S uden s using SEURAT a e mo e likely o
p esen all he ele an al e na i es han hose
using he Pugh me hod. (R1)
19 19 336 405 Null hypo hesis
accep ed, ≈ 0.32
S uden s using SEURAT a e mo e likely o
co ec ly map c i e ia o al e na i es han
hose using he Pugh me hod. (R2)
19 19 322 419 Null hypo hesis
accep ed, α ≈ 0.16
S uden s using SEURAT a e mo e likely o
selec an al e na i e based on hei a ionale
han hose using he Pugh me hod. (R3)
19 19 369.5 371.5 Null hypo hesis
accep ed, α ≈ 0.98
S uden s using SEURAT a e mo e likely o
change hei selec ed al e a i e a e c i e ia
change han hose using he Pugh me hod. (R4)
18 19 343.5 359.5 Null hypo hesis
accep ed, α ≈ 0.60
S uden s using SEURAT will ha e a mo e
comple e se o al e na i es han hose using
he Pugh me hod. (R5)
19 19 322.5 418.5 Null hypo hesis
accep ed, α ≈ 0.16
S uden s using SEURAT will ha e a mo e
c ea i e se o al e na i es han hose using
he Pugh me hod. (R6)
19 19 386 355 Null hypo hesis
accep ed. (No e : his
shows a nega i e
co ela ion) α ≈ 0.66
S uden s using SEURAT will do be e on
ins uc o -de ined pe o mance c i e ia han
hose using he Pugh me hod. (R7)
17 18 246.5 383.5 Null hypo hesis
ejec ed, α ≈0.05*
No es: Ns is he numbe o samples in he SEURAT g oup, and SS is hei scaled ank sum. NM is he numbe in
he Ma ix g oup, and SM is hei scaled ank sum. is a nume ical app oxima ion o he p obabili y o ejec ing
he null hypo hesis when i should be accep ed, based on a wo-sided Mann-Whi ney U es . No e ha since he
lowes ank is bes (1s place is be e han 38 h place), he smalle scaled ank sum indica es be e pe o mance.
* S a is ically signi ican
Bu ge & B inkman
122
We mus ake some ca e in in e p e ing hese esul s. In pa icula , look a ou esul o
R7. This measu es he speed o he s uden ’s solu ion: The ins uc o ga e s uden s some
speed- ela ed c i e ia a he s a . S uden s we e ee, howe e , o ejec hese c i e ia and
ins ead ocus on c i e ia such as ease o coding, ease o debugging, e-use o code, and o he
simila c i e ia ha a e con a y o high sco es in R7.
One majo h ea o he alidi y o his assessmen is due o he way he expe imen al
g oups we e assigned; he SEURAT g oup comp ised all s uden s om one sec ion o he
cou se, while he Ma ix g oup composed he o he sec ion. The SEURAT g oup was s onge
han he Ma ix g oup as measu ed by homewo k g ades on assignmen s o he han he
expe imen al assignmen . I is possible ha highe abili y le els o he SEURAT g oup masked
di icul ies wi h using SEURAT ha would ha e been iden i ied i expe imen al g oups we e
alloca ed in a mo e ca e ul way. Fu he mo e, some ha e sugges ed (e.g., Amabile, 1983) ha
echnical expe ise is a key ac o ha enables c ea i i y. This would mean ha he SEURAT
g oup migh be expec ed o be mo e c ea i e han he Ma ix g oup, on hese so s o asks,
simply due o hei inc eased echnical p o iciency.
To y o co ec o his p oblem we compu ed a bes - i line be ween each expe imen al
a iable (R1-R7) and inal homewo k g ade, and hen analyzed he esidual alues ha esul
when sub ac ing he p edic ed alues om he ac ual alues. The esidual alues essen ially
ell us how much he s uden was o e - o unde -pe o ming on his assignmen compa ed o
his o he usual pe o mance in he class. When compa ing he g oups using hese esiduals,
The g oups we e s ill no signi ican ly di e en o ub ics R1, R2 and R4,
The di e ences in R5 and R7 we e no longe signi ican , and
The Ma ix g oup ou pe o med he SEURAT g oup in R3 and R6, wi h le els o α
≈ 0.07 in bo h cases
Again, we a e no making any s ong claims abou he alidi y o his app oach, bu we
el ha in he in e es o comple eness, as well as ai ness o he Pugh me hod, we should
p esen a g ade-co ec ed e sion o he esul s.
Expe imen al Resul : Ra ionale and C ea i i y
One o he in e es ing end is he e y s ong nega i e co ela ion, -0.4906, be ween R1
(Conside all ele an al e na i es) and R6 (Conside a mo e c ea i e se o al e na i es).
While we canno make any claims abou s a is ical signi icance, his ela ionship seems an
in iguing opic o explo e in u u e esea ch.
We hypo hesize ha his esul s om some p econcep ions among he s udy pa icipan s
abou he numbe o al e na i es ha he ins uc o expec ed hem o gene a e. In pa icula ,
he ins uc o indica ed ha each s uden should ha e “a leas 4-5 al e na i es.”
Some commen a o s ha e sugges ed ha a ionale migh be inhe en ly con a y o
c ea i i y (see Ca oll, 2009). We do no iew ou esul s as suppo ing ha claim, because
we belie e he key p oblem was he p ocess by which s uden s decided whe he o no hey
had conside ed “a su icien a ie y o al e na i es.” S uden s may ha e been ushing o
escape he liminal/undecided s a e, and so simply s opped when hey had ou al e na i es.
We belie e ha his is no a p oblem inhe en in a ionale, bu ins ead a mis ake in he
ins uc o ’s design o he g ading ub ic o he assignmen .
Ra ionale Assis ing S uden Cogni i e De elopmen
123
Summa y o Claims
Because o he na u e o he expe imen , we wan o be e y ca e ul o p ecisely s a e wha we
hink ou expe imen shows. Fi s , ou expe imen gi es some e idence ha using a RMS
ins ead o a mo e adi ional (w i ing) assignmen does no nega i ely impac s uden lea ning.
Though i now seems ob ious ha his would be he case, he cou se ins uc o ini ially had
se ious ese a ions abou using SEURAT in class.
Second, he somewha weake sco es in e alua ing al e na i es seem o suppo ou claim
ha s uden s in he cou se ha e no eached he “commi men ” s ages (le els 7 o 9) o Pe y’s
(1970) scheme. On he o he hand, we would be able o ob ain much be e da a on his opic by
gi ing he subjec s app op ia e c i ical hinking in en o ies, and we should do his in he u u e.
Thi d, ou esul s sugges u u e expe imen s on a ionale and c ea i i y in educa ion mus
be mo e ca e ul abou he ins uc ions gi en o s uden s abou how o assess hei own idea
gene a ion p ocess. I appea s ha he use o a ionale did indeed inhibi c ea i i y, bu p obably
p ima ily as a esul o a poo ly designed g ading ub ic, which ocused on quan i y ins ead o
a ie y. I ou heo y, ha inc easing ime spen in a liminal s a e inc eases c ea i i y, is co ec ,
hen i migh make mo e sense o equi e s uden s o spend a p ede e mined amoun o ime on
idea gene a ion, ins ead o aiming o a p ede e mined numbe o ideas.
REFLECTIONS ON VERIFYING THE CONTRIBUTION OF RATIONALE
TO CREATIVITY IN THE CLASSROOM
In his pape we p oposed a heo y ha use o design a ionale wi h 2nd-yea compu e science
s uden s should lead o imp o ed c ea i i y as well as cogni i e de elopmen . We epo ed on
an expe imen al assignmen ha pu hese ideas in o p ac ice in a Da a S uc u es cou se, and
we epo ed on ou assessmen o he impac o his in e en ion on s uden lea ning. I is
clea , howe e , ha he e a e many lessons o lea n om he i s a emp abou how one
ough o y o e i y he e ec s o such an assignmen .
Fi s , we wished o make claims abou s uden cogni i e le els o epis emologies. A he
mos basic le el, we assumed ha s uden s in ou class would exhibi some dualis ic
endencies, and ha ew would be con ex ual o commi ed knowe s. We a e pa icula ly
in e es ed in how s uden s a di e en le els o cogni i e de elopmen espond o he
challenge o c ea ing design a ionale, bu he e was no way o us o assess his because we
did no collec his da a. In he u u e we a e conside ing applying epis emic in en o ies, such
as Bax e Magolda’s (1992) measu e o epis emological e lec ion, as well as a emp ing o
de elop an in en o y ha speci ically measu es he s uden ’s epis emology o knowledge as i
ela es o so wa e design. We hypo hesize ha s uden s in ansi ion o mul iplici y may iew
ela i ism as no mal in humani ies classes, bu no in enginee ing classes.
I is unlikely, howe e , ha such in en o ies can di ec ly assess he impac o ou
in e en ion on s uden s, e en i gi en as bo h p e es and pos es . Because s uden s ake
many cou ses, mos o which aim o inc ease s uden cogni i e le el, i seems unlikely ha
he e ec o ou in e en ion can be eased ou om such da a. In he u u e we should
di ec ly ask he s uden s ques ions abou why hey did wha hey did and how hey made
decisions on ou pa icula assignmen . The examples o s uden e lec ion p esen ed in
Bu ge & B inkman
124
ea lie in his pape we e sugges i e, bu we did no sys ema ically ask s uden s o commen
on hei p ocesses. So we only ha e such da a o a e y small numbe o s uden s.
Simila ly, gene al-pu pose measu es o idea ion simila o hose used by Shah e al.
(2003) should be adop ed o make ou esul s on c ea i i y mo e compa able wi h esul s
p esen ed by o he esea che s. This would s ill lea e us, howe e , wi h he p oblem o
disco e ing why he s uden p oduced he se o design ideas ha he/she p oduced. This again
equi es some quali a i e me hods ha we did no use in ou ini ial assessmen . We need o
ques ion s uden s, ei he on pape o h ough in e iews, as o why and when hey s opped
gene a ing ideas. I , o example, hey did no s a building hei design a ionale un il he
hou be o e i was due, we should no be su p ised ha he al e na i es discussed we e
canonical o amilia . An app oach simila o he e bal p o ocol analysis used by A man e
al. (2005) would be mos use ul.
FUTURE WORK—RATIONALE ACROSS THE CURRICULUM
The expe imen desc ibed in his pape ocused on one g oup o s uden s, hose aking he
Da a S uc u es cou se ( ypically sophomo e compu e science majo s and junio enginee ing
majo s). I he goal is o aid s uden cogni i e de elopmen , as shown by hei p og ession
h ough he le els o he Pe y (1970) scale, app op ia e exe cises and e alua ion measu es
need o be applied a mul iple poin s h oughou he cu iculum and, ideally, cogni i e
de elopmen e alua ed bo h as an agg ega e o e all s uden s and o indi idual s uden s as
hey pass h ough he p og am.
This would equi e ha he exe cises be a ge ed o speci ic s ages o de elopmen . A he
ea lie s ages o hei educa ion, s uden s can be p esen ed wi h he p oblems, candida e
solu ions, and he a ionale. Fo example, in a Da a S uc u es cou se, he s uden s lea n many
di e en ways o ep esen collec ions o objec s. The s uden ocus is o en on how o
implemen hese collec ions. The implemen a ion is ce ainly impo an , bu since many o
hese cons uc s a e o en supplied wi h he p og amming language (and do no equi e
implemen a ion), i is o en mo e impo an ha he s uden s unde s and why hey migh choose
a pa icula da a s uc u e o a p oblem, ha is, o analyze he possible solu ions by
de e mining which c i e ia a e ele an o he speci ic p oblem. The s uden s’ emphasis on he
implemen a ion a he han he selec ion becomes appa en in la e classes, whe e hey end o
s ick o one o wo a o i e s uc u es ha may o may no be he bes choices o he p oblem
a hand. P o iding a ionale in a o m ha can be easily unde s ood and manipula ed may be a
mo e e ec i e way o each he s uden s he adeo s in ol ed in selec ing be ween da a
s uc u es. The abili y o manipula e he a gumen c i e ia can also help he s uden s o explo e
how changing p io i ies esul in di e en p e e ed solu ions. Ra ionale can be p esen ed in a
o m whe e he c i e ia can be manipula ed by modi ying hei ela i e impo ance in o de o
demons a e how as c i e ia change, so should he ecommended solu ion.
When he s uden s a e com o able wi h he idea o mul iple al e na i e solu ions, he
nex s ep would be o in ol e hem in exe cises whe e he p oblems and c i e ia a e p o ided
bu whe e he s uden s need o iden i y (syn hesize) he candida e al e na i es based on wha
hey ha e lea ned in class and on hei own expe ience. An example o his would be i
Ra ionale Assis ing S uden Cogni i e De elopmen
125
s uden s we e asked o p o ide al e na i e me hods o da a en y o isualiza ion based on
usabili y c i e ia ha hey ha e lea ned in an HCI cou se.
The abili y o iden i y he p oblems hemsel es, p opose solu ions, and de ine c i e ia
equi es e alua ion—iden i ying wha aspec s o he solu ions a e impo an o he p oblem
and i s con ex . This is an essen ial skill in bo h so wa e equi emen s analysis and in design.
The equi emen s elici a ion p ocess is one o de ining he p oblem and he c i e ia unde
which he solu ion will be e alua ed, while he design p ocess in ol es iden i ying and
selec ing solu ions o ha p oblem.
The mo emen om dualism h ough mul iplici y and in o ela i ism and commi men is
mo e o a challenge. Kloss (1994) ecommends se e al s a egies o mo e s uden s om
dualism owa ds ela i ism ha s ess he impo ance o analyzing and s uc u ing di e en
poin s o iew. This equi es looking a he al e na i es and e idence, including unde s anding
he ole o assump ions. As s uden s mo e om wo king wi h he a ionale o o he s o
p oducing a ionale hemsel es, he a ionale can se e as bo h an ins uc ional ool and as a
means o assessing hei in ellec ual de elopmen .
Table 6 lis s he le els o he Bloom Taxonomy (1956), how he e lec ion and a ionale
app oach should suppo hose le els, and how s uden s a di e en le els o de elopmen , as
measu ed by he Pe y (1970) scale, would pe o m on he a ionale-suppo ed asks.
Th ee cou ses in ou depa men cu iculum now con ain explici cou se ou comes
ega ding analysis o mul iple al e na i es: CS2 (1s yea ), Da a S uc u es (2nd yea ), and
Senio Design P ojec (4 h yea ). Using a ionale wi hin hese cou ses will p o ide an
oppo uni y o s udying how a ionale can assis in he s uden s’ p og ession om dualism
owa d he highe le els o de elopmen .
SUMMARY AND CONCLUSIONS
S uden s p og ess h ough se e al s ages as hey gain knowledge and expe ience. They un
in o di icul y when hey need o ope a e a highe cogni i e le els han hey a e accus omed
o. The abili y o make decisions when con on ed wi h unce ain y and ambigui y is
impo an since he p oblems hey will ackle become mo e ealis ic and beyond he poin
whe e, i hey pe o m he igh sequence o ac ions, hey can p oduce a single co ec
answe . The abili y o syn hesize and e alua e solu ions becomes c i ical o p oblem sol ing
and c ea i i y. Rela ed o his is he need o s uden s o mo e beyond duali y, whe e he e a e
always igh and w ong answe s, owa ds highe le els o hinking whe e hey can begin o
analyze he e idence and unde s and ha no all c i e ia a e equally alid in e e y con ex .
Ou expe imen wi h using wo a ionale ep esen a ions, he SEURAT RMS and Pugh’s
(1991) o al design me hodology (as pa o a w i ing assignmen ), indica ed ha s uden s a
he college sophomo e le el do indeed s a ou a a ai ly low le el. The expe imen sugges s
ha using he RMS does no inhibi c ea i i y when compa ed o esul s using he mo e
adi ional w i ing assignmen . Using a ionale, howe e , did no esul in a wide a ie y o
ideas. This could be because he s uden s we e old how many ideas we e equi ed and some
may ha e s opped sea ching once hey achie ed hei “quo a.”
While no gi ing de ini e answe s on a ionale’s impac on c ea i i y, he e we e insigh s
o be gained om he use o a ionale. The a ionale p o ided insigh in o s uden hinking o
Bu ge & B inkman
126
Table 6. Rela ing Bloom’s Taxonomy, Re lec ion and Ra ionale, and he Pe y Scale.
Bloom Re lec ion and Ra ionale Pe y
Knowledge Gi en a gene al decision p oblem, lis and de ine he
al e na i e solu ions, as desc ibed in class.
Example: Wha da a s uc u es can be used o s o e
lis s o i ems?
S uden s a all le els should be able
o do his since i could be di ec ly
ecalled om hei lec u e no es.
Comp ehension Gi en a gene al decision p oblem, and a se o
c i e ia o making a selec ion, explain why hese
c i e ia a e impo an .
Example: Why is i impo an o be able o e icien ly
emo e elemen s om a lis o i ems?
Gi en a se o al e na i es o a gene al decision
p oblem, di e en ia e be ween hem. (This may equi e
gi ing he s uden s he c i e ia).
Example: Wha is he di e ence be ween wo da a
s uc u es ha s o e lis s o i ems?
S uden s a all le els should be able
o explain he c i e ia, bu s uden s
s ill in he dualism s age may show
biases owa d ce ain al e na i es
( he “ igh ” one) when di e en ia ing
be ween op ions and may no
explo e hem in de ail.
Applica ion Gi en a speci ic decision p oblem, gi e a lis o
possible solu ions.
Example: Gi en a design ha equi es so ing and
sea ching a lis o i ems, which da a s uc u es could
be used o sol e i ?
S uden s in dualism may ha e
di icul y p o iding mo e han one
solu ion o mo e han one alid
solu ion.
Analysis Gi en a speci ic design p oblem, p o ide a lis o
possible solu ions and map hose o a se o design
c i e ia.
Example: Gi en a design p oblem ha equi es
so ing and sea ching a lis o i ems, lis he
app op ia e da a s uc u es o s o ing he i ems and
how hey ela e o c i e ia, such as ime equi ed o
sea ch, ime equi ed o add new i ems, e c.
S uden s in dualism may ha e
di icul y p o iding mo e han one
solu ion o mo e han one alid
solu ion. I mul iple solu ions a e
p oduced, hey may ha e ouble
p oposing a gumen s opposing
solu ions hey ha e al eady
deemed “co ec ” o iden i ying
a gumen s suppo ing solu ions
o he han he “co ec ” one.
Syn hesis Gi en a speci ic design p oblem, de ine he c i e ia
ha should be used in o de o make a decision.
Example: Gi en a design p oblem ha equi es
so ing and sea ching a lis o i ems, wha c i e ia
should be used o e alua e candida e da a
s uc u es? Which c i e ia a e mo e impo an o he
speci ic p oblem?
This is no clea . Will s uden s in
dualism only come up wi h c i e ia
ha apply o hei chosen
al e na i e, disca ding any c i e ia
ha do no suppo hei belie s?
Will s uden s in he mul iplici y
s age ha e issues iden i ying some
c i e ia as being mo e impo an
han o he s o will hey conside all
c i e ia equally alid?
E alua ion Gi en a speci ic design p oblem, de ine al e na i es
and he key c i e ia, and use his in o ma ion o
selec a solu ion.
Example: A design p oblem ha equi es so ing and
sea ching a lis o i ems, wha a e he candida e da a
s uc u es, wha c i e ia apply in e alua ing he
app op ia eness o each da a s uc u e o sol ing he
p oblem, and gi en hose c i e ia, which solu ion is
he bes choice?
S uden s in dualism a e likely o
ha e he same issues lis ed abo e,
and a e likely o ha e di icul y
ge ing o he e alua ion s age.
S uden s in mul iplici y may ha e
di icul y in making a selec ion, e en
a e iden i ying al e na i es and
c i e ia.
Ra ionale Assis ing S uden Cogni i e De elopmen
127
he ins uc o o use o assess bo h he s uden ’s unde s anding o he p oblem a hand and
whe e hey a e likely o be in hei de elopmen along he Pe y (1970) scale. Unde s anding
whe e he s uden s a e de elopmen ally, ela i e o whe e we wan hem o go, is impo an in
deciding how o help hem p og ess. Ra ionale can be a aluable ool in bo h aiding and
assessing ha p og ession.
Ou expe imen demons a ed ha a ionale p o ides a mechanism o s uden s o
exp ess he esul s o he analysis, syn hesis, and e alua ion equi ed o design solu ions o
p oblems and p o ides assis ance du ing he p ocess. Explici ly exp essing a ionale o hei
wo k encou aged bo h e lec ion on why hey made hei choices and he ac i e conside a ion
o mul iple al e na i es. This expe imen demons a ed ha s uden s using a ionale
conside ed all easonable al e na i es and we e able o selec c i e ia and e alua e al e na i es
in a way ha indica ed hey we e p og essing in hei in ellec ual de elopmen .
REFERENCES
Amabile, T. M. (1983). The social psychology o c ea i i y: A componen ial concep ualiza ion. Jou nal o
Pe sonali y and Social Psychology, 45, 357–377.
A man, C. J., Ca della, M. E., Tu ns, J., & Adams, R. S. (2005). Compa ing eshmen and senio enginee ing
design p ocesses: An in-dep h ollow-up s udy. Design S udies, 26, 325-357.
Bax e Magolda, M. B. (1992). Knowing and easoning in college: Gende - ela ed pa e ns in s uden s’
in ellec ual de elopmen . San F ancisco: Jossey-Bass Inc.
Bloom, B. S. (Ed.). (1956). Taxonomy o educa ional objec i es: The classi ica ion o educa ional goals. New
Yo k: Da id McKay Company, Inc.
B uegge, B., & Du oi , A. H. (2004). Objec -o ien ed so wa e enginee ing: Using UML, pa e ns, and Ja a
(2nd ed.). Uppe Saddle Ri e , NJ, USA: P en ice Hall.
Bu ge, J. E., & B own, D. C. (2004). An in eg a ed app oach o so wa e design checking using a ionale. In J.
Ge o (Ed.), Design Compu ing and Cogni ion '04 (pp. 557–576). Camb idge, MA, USA: Kluwe
Academic Publishe s.
Ca ey, T., McKe lie, D., & Wilson, J. (1996). HCI design a ionales as a lea ning esou ce. In T. Mo an & J. M.
Ca oll (Eds.), Design a ionale: Concep s, echniques, and use (pp. 373–392). Mahwah, NJ, USA:
Law ence E lbaum Associa es.
Ca oll, J. M. (2009, June). The essen ial ension o c ea i i y and a ionale in so wa e design. Mani es o om
he wo kshop on C ea i i y and Ra ionale in So wa e Design, S a e College, PA, USA.
Casaday, G. (1996). Ra ionale in p ac ice: Templa es o cap u ing and applying design expe ience. In T. Mo an
& J. M. Ca oll (Eds.), Design a ionale: Concep s, echniques, and use (pp. 351–372). Mahwah, NJ,
USA: Law ence E lbaum Associa es.
Daugh y, J., Bu ge, J., Ca oll, J. M., & Po s, C. (2009). C ea i i y and a ionale in so wa e design. SIGSOFT
So wa e Enginee ing No es, 34(1), 27–29.
Dewey, J. (1933). How we hink. New Yo k: D.C. Hea h and Company.
Dym, C. L., Agogino, A. M., E is, O., F ey, D. D., & Lie e , L. J. (2005). Enginee ing design hinking, eaching,
and lea ning. Jou nal o Enginee ing Educa ion, 94, 103–120.
Fische , G., Lemke, A. C., McCall, R., & Mo ch, A. I. (1991). Making a gumen a ion se e design. Human–
Compu e In e ac ion, 6, 393–419.
Fox, C. (2006). In oduc ion o so wa e enginee ing design. Bos on, MA, USA: Addison Wesley.
Bu ge & B inkman
128
Gamma, E., Helm, R., Johnson, R., & Vlissides, J. (1995). Design pa e ns: Elemen s o eusable objec -
o ien ed so wa e. Reading, MA, USA: Addison-Wesley.
Glass, R. S. (1995). So wa e c ea i i y. Englewood Cli s, NJ, USA: P en ice Hall.
Hamil on, H. R. (1974). Sc eening business de elopmen oppo uni ies. Business Ho izons, 7(4), 13–24.
King, P. M., & Ki chene , K. (1994). De eloping e lec i e judgmen : Unde s anding and p omo ing g ow h
and c i ical hinking. San F ancisco: Jossey-Bass Publishe s.
King, P. M., & Ki chene , K. (2002). The e lec i e judgmen model: Twen y yea s o esea ch on epis emic
cogni ion. In B. K. Ho e & P. R. Pin ich (Eds.), Pe sonal epis emology: The psychology o belie s abou
knowledge and knowing (pp. 37–61). Mahway, NJ, USA: Law ence E lbaum.
Kloss, R. J. (1994). A nudge is bes : Helping s uden s h ough he Pe y scheme o in ellec ual de elopmen .
College Teaching, 42, 151–158.
MacLean, A., Young, R. M., Bello i, V. M., & Mo an, T. P. (1991). Ques ions, op ions, and c i e ia: Elemen s
o design space analysis. Human-Compu e In e ac ion, 6, 201–250.
Mo an, T. P., & Ca oll, J. M. (Eds.). (1996). Design a ionale: Concep s, echniques, and use. Mahwah, NJ,
USA: Law ence E lbaum Associa es.
Pe y, W. G. (1970). Fo ms o in ellec ual and e hical de elopmen in he college yea s: A scheme. New Yo k:
Hol , Rineha , and Wins on.
Pugh, S. (1991). To al design. Wokingham, UK: Addison-Wesley Publishing Company.
Rice, J. A. (1995). Ma hema ical s a is ics and da a analysis (2nd ed.). Belmon , CA, USA: Duxbu y P ess.
Schön, D. A. (1983). The e lec i e p ac i ione . New Yo k: Basic Books, Inc.
Shah, J. J., Smi h, S. M., & Va gas-He nandez, N. (2003). Me ics o measu ing idea ion e ec i eness. Design
S udies, 24, 111–134.
an Aals , J. W., an de Mas , C. A. P. G., & Ca ey, T. T. (1995). An in e ac i e mul imedia u o ial o use
in e ace design. Compu e s and Educa ion, 25, 227–233.
VanGundy, A. B. (1981). Techniques o s uc u ed p oblem sol ing. New Yo k: Van Nos and Reinhold
Company.
Whi ing, C. S. (1958). C ea i e hinking. New Yo k: Reinhold.
Au ho s’ No e
This wo k was suppo ed by NSF CAREER Awa d CCF-0844638 (Bu ge). Any opinions, indings, and
conclusions o ecommenda ions exp essed in his ma e ial a e hose o he au ho (s) and do no necessa ily
e lec he iews o he Na ional Science Founda ion (NSF).
We would like o hank he s uden s aking he Da a S uc u es cou se o hei willingness o pa icipa e in
his expe imen , and ellow pa icipan s in he wo kshop on C ea i i y and Ra ionale in So wa e Design.
Co espondence should be add essed o:
Jane Bu ge
Miami Uni e si y
205V Ben on Hall
Ox o d, Ohio, USA
bu [email protected]
Human Technology: An In e disciplina y Jou nal on Humans in ICT En i onmen s
ISSN 1795-6889
www.human echnology.jyu. i