P ac ice Pape
Recommended ci a ion: Qi, H., Li, B., & Chen, C. (2025). Teach Enginee ing
S uden s Expe Thinking Th ough Thinking Aloud Du ing P oblem-Sol ing. In
Kangaslampi, R., Langie, G., Jä inen, H.-M., & Nagy, B. (Eds.), SEFI 53 d Annual
Con e ence. Eu opean Socie y o Enginee ing Educa ion (SEFI), Tampe e,
Finland. DOI: 10.5281/zenodo.17631216.
This Con e ence Pape is b ough o you o open access by he 53 d Annual Con e ence
o he Eu opean Socie y o Enginee ing Educa ion (SEFI) a Tampe e Uni e si y in
Tampe e, Finland. This wo k is licensed unde a C ea i e Commons
A ibu ion-NonComme cial-Sha e Alike 4.0 In e na ional License.
1
ENHANCE STUDENTS' PROBLEM SOLVING SKILLS THROUGH
THINK-ALOUD EXERCISE
H Qia, B Li a , C Chena
a Uni e si y o Cali o nia San Diego, La Jolla, Uni ed S a es
Con e ence Key A eas: Enginee ing skills, p o essional skills, and ans e sal skills
Keywo ds: Think aloud, P oblem-sol ing, me acogni ion
ABSTRACT
Think-aloud is a aluable s a egy o enginee ing s uden s o de elop
p oblem-sol ing skills. By a icula ing hei hough p ocess, s uden s become mo e
awa e o hei easoning, moni o hei unde s anding, and iden i y p oblem-sol ing
gaps and concep ual e o s. This pape p esen s a se ies o in-class in e en ions
implemen ed o os e hink-aloud habi s and enhance me acogni i e p oblem-sol ing
skills among enginee ing s uden s, including o al in-class s a egy discussions and
s uc u ed w i en homewo k. To e alua e he impac o he s uc u ed homewo k and
s uden s' pe cep ion, a mixed-me hods s udy was conduc ed. Quan i a i e analysis
e ealed posi i e co ela ions be ween he quali y o s uden s’ w i en
p oblem-sol ing p ocesses and hei exam pe o mance. In addi ion, su ey
esponses indica ed mixed s uden s’ pe cep ions: while some s uden s exp essed
conce ns abou he addi ional wo kload, many alued he s uc u ed w i en
homewo k app oach o i s cla i y and e lec i e bene i s.
1 INTRODUCTION
P oblem-sol ing is a c i ical skill in enginee ing, equi ing no only echnical
knowledge bu also s a egic hinking and adap i e easoning (B ans o d e al., 2004;
Voss 1983). Resea ch in he lea ning sciences highligh s a key dis inc ion be ween
expe s and no ices: how hey o ganize knowledge and app oach p oblem sol ing.
Expe s end o iden i y ele an p inciples, de elop a plan, and an icipa e
ou comes—o en e balizing hei hinking along he way (Chi e al., 1981). No ices,
by con as , o en di e in o calcula ions wi hou a clea s a egy, bypassing he
planning phase and a emp ing o unde s and p inciples only a e eaching a
solu ion (La kin, 1983). This di e ence e lec s no jus dispa i ies in con en
knowledge bu in me acogni i e abili y— he capaci y o plan, moni o , and e alua e
one’s hinking—which is essen ial o e ec i e and ans e able p oblem-sol ing.
Wi hou explici suppo in de eloping hese skills, s uden s o en eso o o e
memo iza ion: ecalling solu ions o amilia p oblems bu s uggling o apply he
same concep s in new con ex s.
Think-aloud is a cogni i e s a egy in which indi iduals e balize hei hough s in eal
ime while engaging in a ask, such as eading, analyzing, o p oblem sol ing
(Cowan, 2019). I makes in e nal easoning isible, allowing bo h lea ne s and
ins uc o s o obse e and e alua e cogni i e and me acogni i e p ocesses, helping
s uden s mo e beyond o e memo iza ion o a deepe engagemen wi h co e
p inciples. By a icula ing why a concep o me hod is used, s uden s de elop a
2
clea e unde s anding o p oblem s uc u e and easoning, which in u n p omo es
concep ual lea ning and he ans e o knowledge ac oss con ex s (Schmidmaie e
al., 2013 & B ans o d e al., 2004). P e ious s udies ha e documen ed some o he
impac s. McKeown and Gen ilucci (2007) ound ha hink-aloud s a egies suppo
English language lea ne s by enhancing hei language comp ehension and
academic engagemen . Ebne (2013) demons a ed ha hink-aloud echniques can
imp o e ocabula y acquisi ion among uni e si y s uden s. O lieb, No is, and Eh i
(2012) showed ha using hink-aloud me hods signi ican ly enhances science
comp ehension in young child en. Ma hew (2014) employed hink-aloud p o ocols o
assess unde g adua e s uden s’ unde s anding o enginee ing concep s, while
Dounas-F aze (2016) used hem o help s uden s model and oubleshoo elec ical
ci cui s in a physics cou se. Ga din (2010) and Van Some en e al. (1994)
in es iga ed he ole o hink-aloud ac i i ies in os e ing expe -like hinking,
emphasizing how e balizing hough p ocesses can mi o he easoning pa e ns o
expe ienced p oblem sol e s.
Thus, designing lea ning ac i i ies ha p omo e hink-aloud is essen ial o
de eloping s uden s’ me acogni i e and p oblem-sol ing skills. The e a e many
widely used pedagogical app oaches ha can p omp s uden s o hink aloud, such
as Think-Aloud Pai P oblem Sol ing (TAPPS) (Noh e al., 2004; Johnson, 1999;
Ro iqah e al., 2020), Soc a ic Ques ioning (Makhene, 2019; Sahamid, 2016),
P oblem-Based Lea ning (PBL) (Wijnia e al ,2024; T ullàs e al. 2022), s uc u ed
e lec ion (Ma heson e al., 2017; Dounas-F aze & Reinholz, 2015), and o al exams
(Qi e al, 2022; Iannone e al, 2020; Ga dne & Gio dano, 2023). Fo example,
Henjes (2007) used s uc u ed e lec ion o help s uden s sol e complex wo d
p oblems, demons a ing ha guided e baliza ion enhances p oblem-sol ing
e ec i eness. Qi (2022) implemen ed bo h o al exams and ideo-based assignmen s
in a S a ics and Dynamics cou se o engage s uden s in hink-aloud p ac ices and
imp o e hei p oblem-sol ing skills.
Despi e i s documen ed bene i s, li e a u e on he impac o he use o hink-aloud in
unde g adua e enginee ing cou ses emains ela i ely limi ed, pa ly because o he
dynamic and non-linea na u e o me acogni i e p ocesses, which can be
challenging o cap u e and assess. Fo example, du ing Think-Aloud Pai P oblem
Sol ing in mid- o-la ge class sizes, i can be di icul o ins uc o s o e alua e he
e ec i eness o s uden s’ me acogni i e p ocesses and p o ide imely eedback. On
he o he hand, while ideo assignmen s mo e na u ally p omo e he hink-aloud
p ocess h ough e bal a icula ion, hey emain challenging o assess and g ade
e icien ly. Al hough o al exams (Qi, 2022) ha e been shown o be e ec i e in
p omo ing deep hinking and e ealing s uden s’ easoning, hey can be di icul o
implemen a scale wi hou adequa e eaching assis an (TA) suppo . Thus, i is
impo an o design and implemen lea ning ac i i ies ha ac i ely engage s uden s in
hink-aloud p ocesses, while also cap u ing hei easoning in a o ma ha allows o
e icien g ading and ac ionable eedback.
This pape p esen s a se ies o in-class in e en ions implemen ed o os e
hink-aloud habi s and enhance me acogni i e p oblem-sol ing skills among
enginee ing s uden s, including o al in-class s a egy discussions and s uc u ed
w i en homewo k. The o al in-class s a egy discussion helps s uden s o m he
habi s o hinking aloud and ge ing pee and ins uc o alida ion du ing class ime,
3
bu i is challenging o cap u e each indi idual s uden ’s s a us. S uden s we e asked
o documen in w i ing hei p oblem-sol ing s a egies and key akeaways while
sol ing homewo k p oblems. The s udy aims o unde s and he impac o
documen ing he hink-aloud p ocess on s uden s’ lea ning pe o mance, and o
unde s and s uden s’ pe cep ions o such lea ning ac i i ies. The wo p ima y
esea ch ques ions a e:1) How does he quali y o w i en hink-aloud assignmen s
co ela e wi h s uden s’ assessmen pe o mance? 2) Wha a e s uden s’ pe cep ions
o enhancing hink-aloud h ough w i en assignmen s?
2 METHODOLOGY
2.1 Desc ip ion o he s udy - S udy se up and da a collec ion me hod
The s udy was conduc ed du ing he Win e 2025 e m a a public esea ch uni e si y
in he Uni ed S a es, in a 2nd yea enginee ing cou se: S a ics and In oduc ion o
Dynamics, wi h a o al en ollmen o 208 s uden s. The class me wice weekly o
80-minu e lec u es, which we e augh by he p o esso and in oduced new
concep s, along wi h a 50-minu e discussion session led by he TA ha ocused on
p oblem-sol ing. The hink-aloud lea ning ac i i ies in his cou se consis ed o h ee
main componen s:
Raise s uden s’ awa eness on di e ence be ween o e lea ning &
p oblem-sol ing As pa o a b oade ins uc ional in e en ion aimed a imp o ing
s uden s' lea ning skills, he ins uc o deli e ed an 80-minu e lec u e on
e idence-based lea ning s a egies. This session in oduced he concep o
elabo a ion—explaining ideas o onesel o o he s—and i s ole in enhancing
unde s anding and e en ion. S uden s we e also in oduced o Bloom’s Taxonomy
(Bloom, 1964) and emphasized ha he goal o lea ning ex ends beyond o e
memo iza ion o concep ual unde s anding, and p oblem-sol ing. The lec u e
highligh ed how s a egies like hink-aloud suppo deepe lea ning.
In-lec u e “Think-Pai -Sha e” o p omp s uden s o exe cise hink-aloud Du ing
in-class p oblem-sol ing, he ins uc o egula ly inco po a ed “ hink-pai -sha e”
ac i i ies (Mundelsee & Ju kowski, 2021). S uden s we e p omp ed o discuss
p oblem-sol ing s a egies wi h a pa ne , ocusing on key decisions and he
a ionale behind each s ep. These discussions ypically las ed 1–2 minu es and
occu ed in small g oups o 2–3 s uden s sea ed nea by. P omp s we e p o ided o
guide s uden s' hinking; an example is shown in Box 1. S uden s we e encou aged
o ac i ely pa icipa e in hese g oup discussions. Howe e , due o ime cons ain s,
he le el o engagemen and ou comes o hese discussions we e no o mally
eco ded.
S uc u ed W i en Documen a ion o Think-Aloud P ocesses in Homewo k
S uden s we e gi en oppo uni ies o engage in hink-aloud exe cises h ough hei
weekly homewo k. Each assignmen consis ed o 3 o 6 ques ions and included wo
pa s. Pa 1 was comple ed online and au oma ically g aded h ough a digi al
lea ning pla o m. Pa 2 equi ed s uden s o documen hei p oblem-sol ing
p ocess o selec ed ques ions. Speci ically, hey we e asked o include:
a) a p e-a emp ou line o hei p oblem-sol ing s a egy,
b) a de ailed eco d o hei ac ual wo k, including diag ams, equa ions, calcula ions,
4
o code, and
c) a pos -p oblem-sol ing e lec ion summa izing key lea ning akeaways.
S uden s we e p o ided wi h sample en ies o he w i en documen a ion o
Think-Aloud p ocesses and we e encou aged o adop a o ma ha sui ed hei
hinking s yle, as long as hey clea ly demons a ed hei easoning, p epa a ion, and
lea ning—no jus he inal answe s (see Box 2 o examples).
Box 1. Sample in-class hink-aloud exe cise p omp s
Box 2. Examples o w i en documen a ion o Think-Aloud p ocesses o homewo k
2.2 G ading o he W i en Documen a ion o Think-Aloud P ocesses in
Homewo k Assignmen s
The online po ion o he homewo k was au o-g aded h ough he cou se pla o m,
allowing s uden s o ecei e immedia e eedback on answe co ec ness. The w i en
documen a ion was submi ed ia G adescope and g aded ou o 50 poin s: 30 poin s
o he p oblem-sol ing wo k, and 20 poin s o he p e-a emp s a egy ou line and
pos -a emp e lec ion. Each homewo k assignmen con ained 3 o 6 ques ions. Due
o limi ed TA esou ces, 1–2 esponses pe s uden we e andomly selec ed o
de ailed g ading based on pe o mance, while he emaining we e e alua ed o
comple ion. The w i en documen a ion o Think-Aloud p ocesses we e each g aded
on a 3-poin scale: 0 = missing, 1 = insu icien ly de eloped s a egy, and 2 =
well- hough ou s a egy o e lec ion.
5
2.3 Analysis Me hod
This s udy employed a mixed-me hods app oach. Fi s , we examined he co ela ion
be ween s uden s’ p oblem-sol ing p ocesses and hei pe o mance ou comes. The
p oblem-sol ing p ocess was assessed h ough he p e-a emp s a egy ou lines
and pos -a emp e lec ions documen ed in homewo k. Pe o mance ou comes we e
measu ed using se e al me ics, including online homewo k sco es, mid e m exam
g ades, and inal exam g ades (see Table 1).
Due o limi ed TA esou ces, h ee inal exam ques ions and hei co esponding
homewo k ques ions we e andomly selec ed o analysis. We conduc ed mul iple
co ela ion analyses be ween he ollowing pai s: P1 o al & O1; P1 o al & O2; P1-1
& O3-1; P1-3 & O3-3; P1-5 & O3-5.
Table 1. Da a used o measu e p oblem- hink aloud p ocess and p oblem-sol ing ou come
(i=homewo k #, j= inal exam ques ion numbe )
P oblem-sol ing hink-aloud p ocess
P oblem-sol ing ou come
P1-j:(subse o P1) homewo k ques ion 1
hink-aloud g ade o co esponding inal
exam con en
P1-1: homewo k 6 Ques ion 1
P1-3: homewo k 5 Ques ion 3 and Ques ion 4
P1-5: homewo k 6 Ques ion 6
O1: Quiz 2 in class g ade
O2: Final exam o al g ade
O3-j: Final Exam ques ion-
by-ques ion g ade
Nex , we in es iga ed s uden s’ pe cep ions o he homewo k o ma , pa icula ly he
u ili y and p e e ence o he w i en documen a ion, including he p e-a emp ou line
and pos -a emp summa y. This da a was collec ed h ough an end-o -qua e
su ey adminis e ed a e he inal exam. A o al o 196 ou o 208 s uden s
comple ed he su ey and consen ed o he esea ch, which was dis ibu ed
alongside a inal lea ning e lec ion assignmen ha con ibu ed modes ly o hei
cou se g ade.
S uden s we e asked o a e hei ag eemen wi h he s a emen : “The homewo k
w i en documen a ion o he hink-aloud p ocess helps me imp o e my
p oblem-sol ing skills,” using a Like scale (s ongly disag ee o s ongly ag ee).
They we e also asked o selec hei p e e ed homewo k o ma . Desc ip i e
analysis was conduc ed on he quan i a i e esponses. Addi ionally, s uden s we e
in i ed o explain hei homewo k o ma p e e ence h ough open-ended commen s.
A hema ic analysis was pe o med o be e unde s and he easoning behind hei
choices.
3 RESULTS
3.1 How does he quali y o w i en hink-aloud assignmen s co ela e wi h
s uden s’ assessmen pe o mance?
A Pea son co ela ion analysis was conduc ed using s uden s’ homewo k
pe o mance index. The co ela ion esul s and co esponding R alues a e
p esen ed in Table 2. A low posi i e co ela ion was obse ed in mos compa isons.
The s onges co ela ion was ound be ween he o e all homewo k pe o mance
6
index and he inal exam sco e. While co ela ions be ween hink-aloud pe o mance
and speci ic exam ques ions we e weake , he da a sugges s ha s uden s e ain
c i ical in o ma ion mo e e ec i ely when hey explici ly a icula e hei
p oblem-sol ing app oach in weekly assignmen s. This delayed e ec implies ha
egula p ac ice o w i en documen a ion o Think-Aloud p ocesses assignmen may
help s uden s de elop mo e e ec i e enginee ing p oblem-sol ing s a egies o e
ime.
Table 2. Co ela ion s udy o hink-aloud p ocess and p oblem-sol ing ou come and hei Pea son
co ela ion R alues
Co ela ion S udy
R- alue
P1 o al & O1
R=0.238
P1 o al & O2
R=0.358
P1-1 & O3-1
R=0.154
P1-3 & O3-3
R=0.184
P1-5 & O3-5
R=0.052
3.2 Wha a e s uden s’ pe cep ions o enhancing hink-aloud h ough w i en
assignmen s?
S uden s’ pe cep ions o he hink-aloud ac i i ies we e analyzed using su ey da a.
App oxima ely 38.3% o s uden s ag eed o s ongly ag eed ha hese ac i i ies
enhanced hei p oblem-sol ing skill, 23.5% we e neu al, and 38.2% el he
ac i i ies did no con ibu e o imp o emen .
S uden s we e also asked o indica e hei p e e ed homewo k o ma om i e
op ions:1) online au o-g aded + w i en documen a ion wi h p e- and pos -a emp
summa ies (cu en o ma ); 2) online au o-g aded + w i en documen a ion wi hou
w i en documen a ion o Think-Aloud p ocesses 3) online au o-g aded only; 4)
w i en documen a ion only wi h w i en documen a ion o Think-Aloud p ocesses;5)
w i en documen a ion only wi hou w i en documen a ion o Think-Aloud p ocesses.
A majo i y o s uden s (56.6%) p e e ed he second op ion—online au o-g aded
homewo k combined wi h w i en documen a ion bu wi hou he w i en hink-aloud
documen a ion. This sugges s ha while s uden s alue ecei ing immedia e
eedback and eco ding hei calcula ions, many a e less inclined o engage in
e lec i e componen s o he assignmen .
S uden s alued he online po ion o he homewo k o i s immedia e eedback and
abili y o alida e hei answe s. The w i en componen was app ecia ed o wo
main easons: (1) i p o ided p ac ice in enginee ing documen a ion and se ed as a
e e ence o u u e e iew o discussion, and (2) i helped mi iga e he isk o losing
poin s when inal answe s in he online sys em we e inco ec . While 38.3% o
s uden s ecognized he alue o he p e- and pos -a emp summa ies, only 17.9%
p e e ed he cu en ull- o ma homewo k (online + w i en documen a ion wi h
w i en hink-aloud). In con as , 14.8% p e e ed online-only homewo k, and 10.8%
p e e ed w i en-only assignmen s. O e all, 74.5% o s uden s a o ed a hyb id
7
app oach ha combined au o-g aded online p oblems wi h w i en documen a ion,
while 25.5% p e e ed a single-mode o ma (online-only o w i en-only).
A hema ic analysis o open-ended su ey esponses p o ided u he insigh in o
s uden s’ p e e ences ega ding he homewo k o ma , pa icula ly he inclusion o
p e- and pos -a emp summa ies. On he posi i e side, many s uden s el ha his
o ma helped imp o e hei p oblem-sol ing skills and acili a ed deepe lea ning.
They sha ed ha a icula ing hei s a egies be o e and a e sol ing p oblems
ein o ced key concep s and suppo ed long- e m e en ion. They iewed he
summa ies as use ul ools o consolida ing p oblem-sol ing s a egies and
app ecia ed he immedia e eedback p o ided by he online po ion o he homewo k,
which allowed hem o alida e hei app oach in eal ime. O he s no ed ha he
e lec i e p ocess ac ed as a checkpoin ha helped cla i y hei unde s anding and
guided hei lea ning. This posi i e heme is e iden in sample s uden commen s:
“Helped me ully unde s and he p oblems.” “The summa ies a e e y use ul o
e en ion o p oblem-sol ing s a egies. I ind ha he online po ion o he homewo k
is use ul o ge ing ins an eedback on you app oach by being able o check he
co ec ness o an answe .”
Howe e , no all s uden s ound alue in he w i en hink-aloud. A numbe o
esponses exp essed us a ion wi h wha was pe cei ed as unnecessa y
complexi y. These s uden s el ha combining au o-g aded online wo k wi h w i en
documen a ion, pa icula ly he w i en hink-aloud, made he p ocess cumbe some
and o e ly ime-consuming. Some pe cei ed he e lec i e w i ing as edundan o
unp oduc i e, no ing ha i did no meaning ully enhance hei unde s anding. Some
sample commen s: “I’m an enginee w i ing essays shouldn’ ha e o be a
p e equisi e o u n in my enginee ing homewo k” “I o en el ha p e and pos
a emp s we e epe i i e and no p oduc i e o me pe sonally.” These mixed
esponses highligh he impo ance o balancing cogni i e bene i s wi h wo kload and
pe cei ed ele ance in designing e ec i e lea ning ac i i ies.
To be e unde s and which g oups o s uden s ound he hink-aloud exe cises
help ul, we compa ed s uden s’ pe cei ed help ulness a ings wi h hei sco es on a
beginning-o -qua e p e-quiz, which assessed ounda ional physics knowledge
equi ed o he cou se. P e-quiz sco es we e g ouped in o i e ca ego ies based on
s uden s' esponses o he pe cei ed help ulness o he hink-aloud ac i i ies. The
esul s o his analysis a e p esen ed in Fig. 1. This pa e n is aligned wi h cogni i e
psychology li e a u e, sugges ing ha he expe e e sal e ec is a play. S uden s
wi h mo e p io knowledge will ind ins uc ional in e en ions ha sca old lea ning o
be mo e edious and less help ul. Some high-pe o ming s uden s may e en be
nega i ely a ec ed by hese in e en ions.
This pa e n is u he suppo ed by bo h he median and a e age p e-quiz sco es
ac oss he g oups. S uden s who pe cei ed he hink-aloud p omp s as unhelp ul
sco ed signi ican ly highe on he p e-quiz han hose who ound hem help ul. This
sugges s ha s uden s who en e ed he cou se wi h s onge ounda ional knowledge
may ha e el less need o s uc u ed e lec ion, while hose wi h weake
backg ounds we e mo e likely o bene i om and app ecia e he hink-aloud
ac i i ies.
8
Fig. 1. S uden s’ esponse o “The homewo k w i en documen a ion w i en hink-aloud summa y
helps me o imp o e p oblem-sol ing skills” s hei p e-quiz sco e.
Table 3. Co ela ion s udy o hink-aloud p ocess and p oblem-sol ing ou come and hei
Pea son co ela ion R alues
Pe cei ed Use ulness o Homewo k
Think-aloud P omp
Median P e-quiz Sco e
A e age P e-quiz Sco e
S ongly ag ee
45.00
50.00
Ag ee
45.00
50.00
Neu al
60.00
57.77
Disag ee
61.00
58.59
S ongly disag ee
69.00
65.78
Pe son R
-0.952
-0.955
4 DISCUSSION AND CONCLUSIONS
In his s udy, we examined how hink-aloud s a egies can be augh and hei impac
on s uden lea ning. The indings open new possibili ies o u u e pedagogical
p ac ices and esea ch. Howe e , se e al challenges emain. Fi s , cap u ing
s uden s’ ac ual me acogni i e ac i i y is inhe en ly di icul . The w i en p e-a emp
ou lines and pos -a emp e lec ions may no ully ep esen he cogni i e and
me acogni i e p ocesses s uden s engage in du ing p oblem-sol ing. Second, w i ing
down he dynamic and non-linea hink-aloud p ocess can be edious, less engaging,
and e en dis up i e o s uden s’ na u al low o s a egy de elopmen . Mo e o ganic
app oaches—such as ideo- eco ded hink-alouds—could o e iche insigh in o
s uden hinking. Howe e , hese come wi h p ac ical challenges, pa icula ly in e ms
o g ading and scalabili y. Thi d, hese limi a ions highligh he b oade challenge o
de eloping alid, e icien , and scalable me ics o assessing hink-aloud p ocesses
while also p o iding s uden s wi h cons uc i e eedback o help hem e ine hei
me acogni i e s a egies, p oblem-sol ing skills, and expe hinking.
5 ACKNOWLEDGEMENTS
This ma e ial is based upon wo k suppo ed by he Uni ed S a es Na ional Science
Founda ion unde G an No.2315777.
