"Mathematical Knowledge For Teaching": Adapting U.S. Measures For Use in Lreland

D e l a n e yS, . , B a l l ,D . L . ,H i l l ,H . C . , S c h i l l i n gS,. G . ,& Z o p ,f D . , ( 2 0 0 8 )".
M a t h e m a t i c a l
knowledge for teaching": AdaptingU.S.measures for use in lreland.Journalof
Mathematics Teacher Education, 11(3), 171-197
J Math TeacherEduc (2008) tl:t7t-19'l

DOI 10.I 007/s10857-008-9072-I
"Mathematical knowledgefor teaching": adapting

U.S. measuresfor use in lreland
Sedn Delaney . Deborah Loewenberg Ball . Heather C. Hill .

Stephen G. Schilling . Deborah Zopf
Published online: l9 January 2008

@ Springer Science+BusinessMedia B.V. 2008
Abstract This article describesa study in which measuresof mathematicalknowledge

for teaching developed in the United States were adapted to measure mathematical
knowledgefor teachingin Ireland.When adaptingthe measuresit was not assumedthat
the
mathematicalknowledgeusedby Irish and U.S. teachersis the same.Insteadpsychometric
and interview-basedmethods were used to determine a correspondencebetween the
constructsbeing measured,and ensure the integrity of item performancein the
Irish
context. The study found overlap betweenthe knowledge that is used to teach in
both
Ireland and the united states, and that the items tapped into this knowledge.However,
specificfindings confirm the usefulnessof conductingextensivecheckson the validity of
items used in cross-nationalcontexts.The processof adaptationis describedto p.ouid"
guidancefor othersinterestedin using the iiems to measuremathematicalknowledge
for
teachingoutsidethe United States.The processalso enabledthe authorsto raisequestions
about the assumptionsthat lie behind the practice-basedconstruct of mathematical
knowledgefor teaching.
Keywords Mathematicalknowledgefor teaching . practice of teaching .

Test . Subjectmatter . Ireland . United States
Abbreviations
CCK Common contentknowledge
IRT Item ResponseTheory
KCS Knowledgeof content and students
KCT Knowledgeof content and teachins
An earlier draft of this article was presentedat the annual meeting of the American
Educational Research
Association, Montr6al, CN, April 15, 2005.
S. Delaney (8)
coliiste Mhuire, Marino Institute of Education, Griffith Avenue, Dublin 9. Ireland
e-mail: sdelaney@umich.edu
S. Delaney ' D. L. Ball . H. C. Hill . S. G. Schilling .D. Zopf

University of Michigan, Ann Arbor, MI, USA
Q Sp.i"g.t
,,\"^!,,2i{#,^":,y:^,,!.::1,t1c_s^!,.e_7cner
,.v vvq,,,st v, ,vtaLtrctuaalus leacner
tE^dy,ca,t!on
ispubtished
byspringer
is_availab
roucauon E rhe orisinatpubtication
le
at w.ww. sprinserlink.
com (http://www. K
Reproducedwith permissionof tfre;
i
permission.
t72
S. Delaney et al.
LMT Learning Mathematicsfor Teachins

MKT Mathematicalknowledgefor teaching
NCTM National Council of Teachersof Mathematics
PISA Programfor InternationalStudentAssessment
SCK Specializedcontentknowledge
TIMSS Trends in InternationalMathematicsand ScienceStudy
U.S. United Stdtes
Introduction
The last two decadeshave seengrowing interestin cross-nationalstudies

of teachingand
learning. The Trends in InternationalMathematicsand ScienceStudy (TIMSS)
and the
Program for International StudentAssessment(PISA) have demonstrated
strengthsand
weaknesses in studentachievementacrossdevelopingand industrializedcountries.
As
theseresultshave beendisseminated,therehasbeena lively effort by scholars
interestedin
determining which aspectsof high-achieving countries' mathematics
instruction matter
(Stigler and Hiebert 1999).The quantity and quality of teachers'
mathematicalknowledge
has been one leading areaof interest(An et al. 2004;Ma 1999).However,
while student_
level measuressuch as thoseincluded in TIMSS and PISA are carefully
investigatedfor
neutrality with regard to languageand culture, measuresof teacherknowledge
are not.
This article arguesfor and providesan initial exampleofthe cross-cultural
adaptationof
a set of measuresof teacherknowledge.It arguesthat beyondthe literal translation
needed
to convert items from one languageto another,test developersneedto
carefully account
for differencesin the culture of teachingmathematicsitself. Severalstudies
suggestthat
the work of teachingin the United Statesdiffers from the work of teaching
elsewhere(e.g.,
cogan and Schmidt1999;Santagata2004; Schmidtet al. 1996;Stigleraid Hiebert 1999).
Stigler and Hiebert (1999), for instance,claim that different beliefs about
the nature of
mathematics,the nature of leaming, the role of the teacher,the structure
of a lesson,and
how to respondto individual differencesamong studentslead to different
teachingprac-
tices in the United Statesand Japan.Others claim rhat the spreadof global
models of
schoolinginfluencenationaleducationsystems(Meyer et al. lggz). everiir
t11isis true the
global models interact with nationallaws, customsand expectations,
and local influences
could still affect how lessonsare planned,for example uuluing mathematical
reasoning
over memorization(LeTendreet al. 2001).If the work of teachingdiffers from
one counrry
to another,an instrumentto measureknowledgefor teachingn""dr to be
sensitiveto such
differences.
This is especially true for the measureswe describe here, which were
intended to
capture "mathematicalknowledgefor teaching" (MKT), or the mathematics
teachersuse
in classrooms,a construct that are explained in more detail below. In
addition to the
differencesin mathematicsteachingidentifiedin the aboveliterature,there
may be less
visible-but critically irnportant, from a measurementperspective-differences
in key
tasks of teachingmathematics.For instance,the use of mathematicalterms
and conven-
tions might differ; the presenceand prevalenceof particular contentstrands particular
at
gradesmight differ; the typical responsesof srudentsmight vary.
Given that the MKT
items were intended to measureteachers'use of classroo..,ûih"-utics
in the United
States,cross-nationalvariability in classroommathematicswould threaten
the interpreta-
tion of resultsfrom.studiesin othercountries.
Q Springer
TheJournalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

is availableat www.springerlink.com (http://www.sprinqerlink.com/conienUt460664281673730/).
Reproduced prohibited
with permissionof the copyrightowner.Furtherreproduction without
permission.
Adapting measuresof mathematical knowledge for teaching t73
This article emerged from an attempt by the lead authorr to use the U.S. items to
measurethe mathematicalknowledge held by primary teachersin Ireland. Before using
the U.S. items to investigateIrish teachersmathematicalknowledge, a pilot study was
deemed necessaryto investigate how multiple choice items would be selected and
adapted.Therefore the pilot study centers on the selection, adaptation,and evaluation
of items, some of which would be used in a subsequentstudy of Irish teachers'
mathematical knowledge for teaching. This article is based on the pilot study and
representsan attempt to identify and respond to issues that arise in adapting MKT
items for use in another country. Ireland was particularly appealing becausethe lan-
guage of instruction (English) was largely held constant; this allowed for greater
attention to the other issuesof adaptationthat might arise. It was also anticipatedthat
the multiple-choice items could subsequentlybe used in Ireland, in pre-testsand post-
tests, to measurehow mathematicalknowledge grows during pre-serviceteacher edu-
cation, during professionaldevelopmentinterventions,or with teaching experience(Hill
and Ball 2004). Eventually,following extensivecross-national validation,it is possible
that the items could be used to compare teachers'mathematical knowledgefor teaching
acrosscountries.
As the literatureon test adaptationrecommends,we did not assumethe items could be
used in Ireland without modification.Becausetheseitems were basedon a constructthat
emergedfrom studying the practice of mathematicsteaching in the United States,we
made item adaptationa study in itself prior to using the items as a measuringinstrument.
We asked:what methodologicalchallengeswere encounteredwhen attemptingto use the
items outsidethe United States?What choicesdid we make when adaptingthe items and
why did we make them? How did we evaluatethe successof the adaptations?What did
these initial explorationssuggestabout the suitability of the U.S. measuresfor studying
mathematicalknowledge of teachersin Ireland? Answering these questionswould help
us solve the immediateproblem of adaptingthe measuresfor Ireland. Moreover, it would
be of help to other researcherswho wished to adapt the measuresfor use in their
countries, and would lead to greater understandingof the construct of mathematical
knowledge for teaching.
In answering the questions,we begin with an overview of researchon teachers'
knowledgeof mathematicsand on other instancesof translatingtestsfrom one country to
another.We then describethe methodsusedin the current study.The methodsincluded a
moderate-scalepilot of an untimed written test administeredto teachers(n : 100) and
follow-up interviews with five teachersafter administeringthe pilot. In the resultssection,
we outline how psychometricanalysesallowed us to fine-tuneour understandingof how
the translated'items performedin Ireland.Note our interestin this article is not abouthow
teachersperform on the items but in how the adapteditems perform in a new setting. A
well-performing item is one that generallyhigh-scoringteachersanswercorrectly and one
that low-scoring teachersget wrong. We concludeby discussingthe findings and recom-
mendingstepsthat researchers might take when adaptingmeasuresof teacherknowledge
from one country to another.
'
The lead author spent I I years teaching in Irish primary schools, where he taught all class levels and
worked for almost 2 years as a resource teacher with specific responsibility for mathematics.
2 We
use the terms "translate" and "adapt" interchangeably throughout this article.
Q Springer
The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

is availableat www.springerlink.com (http://www.sprinqerlink.com/contenVt460664281673730/).
permission.
174 S. Delaney et al.
Background
Severalstudiesover the past two decadeshave investigatedhow teachersacquire math-

ematicalknowledgeand how they apply it when teachingmathematics(e.g.,Ball 1990;
Borko et al.1992; Eisenhartet al. 1993;Steinet al. 1990).Building on this work and on
Shulman's (1986).hypothesisabout a professionallyuseful form of subject matter
knowledgecalled "pedagogicalcontentknowledge,"scholarsat the Universityof Mich-
igan have been developinga theory about the mathematicsthat primary school teachers
need to know. This knowledge,called mathematicalknowledgefor teaching(MKT), not
only includes aspectsof pedagogicalcontent knowledge, but also incorporatessubject
matter knowledge,both common and specialized.The relationshipbetweenpedagogical
content knowledge and subject matter knowledge in the construct of mathematical
knowledgefor teachingcan be seenin Fig. I (from Hill et al. 2007). The figure illustrates
that MKT consistsof disciplinaryknowledge,including both mathematicalknowledgethar
is held by well-educatedadultsand mathematicalknowledgethat is specializedto the work
of teaching. In addition to disciplinary knowledge, MKT includes pedagogicalcontent
knowledge, that is knowledge of how to make mathematicalideas understandableto
studentsand knowledgeof what studentsfind difficult in mathematicsand students'typical
perceptionsand misconceptions (Shulman1986).
The conceptof mathematicalknowledgefor teachingemergedfrom studyingrecordsof
practice of mathematicsteaching, and identifying the recurrent tasks teachersdo that
require mathematicalknowledge,reasoning,and insight (Ball and Bass 2003a;Ball et al.
2005). For example,teachersanalyzethe mathematicalstepsthat producestudents'errors.
They also design explanations,assessstudents,pose and respondto mathematicalques-
tions, evaluatethe quality of teachingmaterials,deploy representations,
explain concepts,
and show why procedureswork. Such tasks demandmathematicalknowledge (Ball and
Bass2000,2003b).
In order to learn more about the mathematicalknowledgethat is neededfor teaching,
the researchersdevelopedassessmentitems that could measurethe MKT held by large
Knowledge
Common of Content
Content Specialhed and
Knowledge Content Students
(ccK) (Kcs)
Knowledge
Knowledgeof
(sc() Curriculum
Knowledge at Knowledge
the of Content
mathematical and Teaching
horizon (Kcr)
Fig. I Mathematical knowledge for teaching (Hill et al. 2007)
Q Springer

permission.
Adapting measuresof mathematical knowledge for teaching
175
numbersof teachers.Theseitems, developedby the Learning Mathematics

for Teaching
(LMT) project' requireteachersto identify the most appropriate
answerfor multiple-choice
questionsthat are basedon school- and classroom-relatedscenarios.
Preliminary psychometric analyses, including factor analyses
of u.S. teachers,
responsesto the items, supportthe existenceof the mathematicalknowledge
for teaching
constructand suggestthat this constructhas specific dimensions
or domains (Hill et al.
2004)' Four domainsof.mathematicalknowledgefor teachinghave
beenhypothesizedand
analysessuPportthe existenceof the first three.The four dimensions
are commoncontent
knowledge(ccK), specializedcontentknowledge(scK), knowledge
of contentand stu_
dents (KCS) and knowledgeof content and teaching (KCT) (Ball
e*ral. 2005). CCK and
SCK are subjectmatter knowledge and KCS and KCT, along with
knowledgeof curric_
ulum, are a form of pedagogicalcontentknowledge (Ball et al.
2006). common content
knowledgerefersto the_mathematical knowledgeand skill possessed by any well-educated
adult, e.g.,correctlysubtracting75-18. Specializedcontentknowlejge
is mathematical
knowledgeand skill usedby teachersin their work but not generally
neti uy well-educated
adults,e.g.,knowingalternativealgorithmsfor calculatingZS-f S. Knowledge
of content
and studentsincludesknowingaboutbothmathematics andstudents, e.g.,recognizingwhy
a studentmight give the answer63 to75-18. Knowledgeof content
andteachinginvolves
knowing about both mathematicsand teaching,e.g., knowing instructional
advantagesof
different representationsof 75-18. Although these domains have
been identified in the
U.S., the domainsof knowledgemay differ in other settings.
Nevertheless,mathematicseducatorsin other countriesliuu.
interestin using
thesemeasures'Somefurther believe that the items could be used "*pr"rr"d
to compareknowledge
of teachersin different countries.The comparativeinterestmay havebeenprompted
by the
useof high-profile testssuchas the Trendsin InternationalMathematics
and ScienceSiudy
(TIMSS)' which measuredstudents'mathematicalachievement
acrosscountries.Indeed
some studies have already comparedU.S. and chinese teachers'
knowledge of mathe_
matics (e.g., An et al. 2004; Ma 1999) and these studiesconcluded
that the chinese
teachers'knowledge of mathematicsfor teaching was generally superior
to that of u.S.
teachers.
Unfortunately, these studies typically provide little information
as to how the mea-
surementinstrumentswere adaptedfor use outside the united
states. For example,An
et al' (2004) statethat "questionnaireswere preparedfirst in English
and then translated
into Chinese" (p. 15l), but little information * giu.n about translation
issuesthat arosein
the research.This is critical, becausemisunderstandingsof terms
can alter whether and
how instrumentsdiscern teacherknowledge.These problems even
occur without trans-
lating items into a different language;Borko and her colleaguesdescribed
how multiple
possibleinterpretarions of whatis meantby "to providean explanationfor" (lgg2, p.2i2)
causedproblemsfor one studentteacher.If the meaningof a mathematical
practicecan be
misunderstood in onelanguage,thereis room for evenmoremisunderstanding
whenterms
are translatedinto a secondlanguage.
Ma does not explicitly describehow the mathematicalteachingtasks
she uses,which
were developedin the united States (Bail lggg), were adapteJ
for use with chinese
teachers'In severalplacesthroughouther book she placesin context,
explainsor refers to
the Chinese words for particular expressions,mathematicalconventions,
units of mea-
surement,mathematicsteachingpracticesand mathematicalterms.
In many casesthese
notes are placed in footnotes,giving the impressionthat they are
supplemenrary,even
thoughthey may be centralto the main data.Further,the teaching
tasksthemselvesmay or
may not have been appropriate as contexts for evaruating Chin"r"
teachers, use of
0 Springer

permission.
t76
S. Delaney et al.
mathematicalknowledge.Indeed,Ma (1999) changedthe context of ..subtraction

a with
regrouping" scenariofrom one that assessedU.S. teachers'ability
to evaluatetextbook
selectionsto one that assessed how Chineseteacherswould approac-h teachingthe problem
andwhat prior knowledgestudentsneeded.Changingthe coniext of an
item may alter how
teachersrespondto it, but Ma did not discusshow this change
might have alTectedher
comparisonof u.S. and chinese teachers'mathematicalknowledee.
Internationalcomparativestudiesin other fields over the last 210years
have led to the
developmentof guidelinesfor translatingor adapting tests from onl
.oun,.y to another
(Grisay2002; Hambletonet al. 2005; Maxwell 1996).This
literatureprovidesexrensive
descriptionsof hgw items were adaptedand testedfor equivalence
in different countries,
including thosewherethe samelanguageis spokenbut whereotherdifferences
exist. Many
of the recommendedprocedurescan be appliedto translatingmathematical
knowledgefor
teaching items, such as using translatorswho are native speakersof
the target language,
who know the target culture and who have expertise ln uoth subject
matter and test
developmentmethods.Furthermore,these studiesdiscusshorv psychometric
data can be
used to evaluate the successof adapting the measures.Yet the,process
of translating
mathematicalknowledge for teaching items is also different from
other kinds of item
translation.
One differenceis that the mathematicalknowledgefor teachingitems were
not initially
designedto be usedin other settings.In fact, they were deliberateiydesigned
to reflect the
work of teachingin one country, the United States.In contrast,when
deJetopingitems for
TIMSS in 2003, National ResearchCoordinatorsfrom each participating
counrry con-
tributed to the designand preparationof the test instrument(Chrostowski
and Malak 2004).
This lack of input from potential intemational usersof the items makes
translationmore
difficult becausethere is a greaterrisk that items may be biasedin
favor of U.S. teachers
and unlike the PISA studiesonly one sourcelanguagefor items exists (e.g.,
Grisay 2002).
A secondway in which translatingMKT items is different from transiating
TIMSS or
PISA items is that they are grounded not in the discipline of mathematics,
but in the
practice of teachingmathematics.This makesadaptingthe items
more costly in terms of
time and expertise.Like in TIMSS and PISA, one needsto recruit professional
translators.
and apply proceduressuch as double translationand reconciliation.In
addition, however,
one needsto recruit and train multi-disciptinaryteamsexpert in mathematics,
in measures
development,and in the practice of teathing in the country for which
items are being
adapted'Such teamscan adviseon how well the items relate to mathematics
teachingin
particular countriesand can identify areasof MKT not measured
by the irems.
A third differencealso relatesto the contentbeing measured.Tests
such as TIMSS and
PISA measuremathematicalknowledge, which is commonly assumed
to be universal.3
Notwithstanding this qommon belief, comparative studies of mathematics
achievement
generallyhave beencriticizedfor their assumptionof an "idealized
internationalcurric-
ulum, definedby a common setof performancetasks" (Keitel and Kilpatrick
lggg). Keitel
and Kilpatrick further say that "no allowanceis madefor different aims,
issues,history and
context acrossthe mathematicscurricula of the systemsbeing studied',(p.
za!. we argue
that theseand many more potentialdifferencesexist in the mathematicai
knowledgeused
for teachingin different countries.Actual and potential areasof difference
in mathematical
knowledge for teaching across countries relate to teachers,students,
mathematicsand
teachingmaterialsand include:
'
Some will argue that assuming mathematics to be universal is mistaken
and that it is culturally bound
(e.9., see Jaworski and Phillips 1999).
Q Springer

is availableat www.springerlink.com (http://www.sprinqerlink.com/contentit4606642B1673730/).
permission.
t77
what teachersdo during mathematicslessons.For example,the

amountof time teachers
devote to whole class instruction compared to working with groups
or individuals
variesacrosscountries(Stigleret al. l9g7);
teachers'conceptionsaboutmathematicsand aboutmathematics
teaching.For example
one study of teachersfrom England and Hungary linked different
conleptions about
mathematicsand mathematicsteachingto the presenceof mathematics
wall displaysin
English classrooms,a feature absentin Hungarian classrooms(Andrews
and Hatch
2000);
the classroomcontextsin which the knowledgeis used.Some
teachers,for example,
work in linguistically homogenousclassroomswhereasothers
work in murtilingual
classrooms.{n multilingual situationsteachersmust usetheir
mathematicalknowledge
to explain concepts and to understandstudents' thinking processes
without always
understandingsrudents'language(Gorgori6 and planas200i);
differencesin the types and sophisticationof the explanations
studentsmake (Silver
et al. 1995).Teachersmay needdifferent levels ofknowledge
to respondappropriately
to typical explanationsin their country;
respondingto student errors. In some countriesteachersmight
decide to mitigate a
student'serror whereasteacherselsewherehave beenfound tJexpress
disappointment
at enors through the use of harsh or ironic comments
lsantagata2004;;
the presenceand prevalenceof specific mathematicaltopics.-For
example,the grade
level at which topics appearhave been found to differ u-ong
.ount i", (e.g., Fuson
et at. 1988);
the mathematicallanguageused in the school. Spellings,units
of measurementand
punctuationhave been found to vary among Engliih speaking
countries(chrostowski
and Malak 2004);
o the contentof the textbooks.This can vary
from country to country making different
demandson, and providing different supportsto, teachers(e.g.,
Mayer et al. 1995).
All of thesepotential differencesin the practiceof teachingmathematics
renderguidelines
available from previo rs studies only partially helpful in this
translation endeavor.
Translation guidelines from TMSS and pISA offer an informative
starting point, but
additional challengescan be expectedif MKT items are translated
for use outside the
United States.
Nevertheless,despitethe many potentialdifferences,it seemsprobable
- that somedegree
of overlap exists betweenmathematicalknowledgefor teachingin
different countries.For
example,one common contentitem asksrespondentswhetherthe
statement,..0 is an even
number," is true' The answerdependson the definition ofeven
numbersand it cannotvary
amongcountriesunlessa fundamentalmathematicalassumption
is changed.Moreover, it
is likely that this is somethingthat should be known by ali
teachers,(fJr example when
teachingstudentsto classifynumbersas odd or even in early grades'or
when deciding
whetherthe sum or productof two numbersis odd or evenin
hig"hergruJ..;. If agreement
among countriesis likely on the importanceof such common
content knowledge,other
items, especially those that relate to knowledge of students
una to ,p".iulized content
knowledge,may reflect knowledgethat varies i.o-.ountry to
country.
In order to help specify how differencesmight occur,we describe
mathematicsteaching
in the two countriescentral to our analysi.. ih. ^"urures
originate in the united States
where there is no centralizedmathematicscurriculum and
at a=policylevel, mathematics
is influencedby the work of the National council of reaclers
1:i:1*c of Mathematics
(NCTM) and particularly the counciIs cuticurum
and Evaruationstand.ardsfor school
Q Springer
rhe Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

is availableat www.springerlink.com
thttp:ltryyw.springe
Reproducedwithpermissionofthecopyrightowner'F66u1 l
permission.
Mathematics(1989) and its Principles and Standardsfor School Mathematics(2000).4

Mathematicseducationin the Republic of Ireland has been less well documentedand a
brief overview is presentedhere to describe the context for which the measureswere
translatedflromthe U.S. version.
In Ireland disagreementsabout the teaching of mathematicshave never caught the
imaginationof media or the public in the way that characterizedthe "Math Wars" in the
United States(Wilson 2003).Neverthelessa study by Walsh (2006) of inspectors'reports,
curriculumdocuments,test results,notesfor teachers,and small scalestudiesbetween
1922 and 1990 reveals a persistent,silent war being waged betweenthe aspirationsof
policy makers and curriculum designerson one hand and the practice of teachersin
teachingmathi:maticson the other. Teacherswere continuouslyencouragedto emphasize
applicationsof mathematics,the useof concretematerialsand problemsbasedon students'
experiencesand interests,the promotionofthinking and the centrality oflanguageand oral
discussion.Although someof thesewere observedin practice,the predominantpicture in
schools for most of the 70 years was one of over-relianceon textbooks, emphasison
mechanicaland routine aspectsof mathematicsand difficult vocabularyin textbooks.In
general,studentswere consideredto perform well in arithmetic but less well in algebra,
geometryand problemsolving(Walsh2006).
A revised curriculum was introducedin 1999 which was structuredaround intercon-
nected strands (number, algebra, shape and space, measures,and data) spannedby
mathematicalskills. The curriculum advocatedprinciples such as building on previous
knowledge, teaching informed by assessment,teaching using constructivist and guided
discovery methods,using languageeffectively and accurately,manipulating materialsto
support concept acquisition, calculating mentally and estimating, using technology to
support mathematicalprocesses,and problem solving (Government of Ireland 1999).
Although the datescheduledfor full implementation,2002,wasrelatively recent,the initial
signsare that in many classroomslittle haschanged(Governmentof Ireland 2005; Murphy
2004; NationalCouncilfor Curriculumand Assessment 2005).
In summary,the United Stateshas been working at reforming mathematicsinstruction
longer than Ireland, with more resources,and with arguably more success.This may
changethe natureof MKT. However, at policy level at least,the contentstandardsof the
National Council of Teachersof Mathematicsare similar to the strandsof the Irish primary
mathematicscurriculum and the NCTM processstandards(National Council of Teachers
of Mathematics2000) are similar to the Irish curriculum skills (Governmentof Ireland
1999).
Method
Ireland is a good first testing-groundfor efforts to translatethe U.S. items. Ireland differs
from the United States in many features such as size, history, diversity and cultural
activities, such as sports. The countries also share similarities. Speaking in 2000, the
Deputy Prime Minister of lreland,-MaryHarney,claimed that "spiritually" the Irish are "a
lot closerto BostonthanBerlin".' Formalcollaborationbetweenthe two countriesoccurs
a
See http://www.standards.nctm.org/
s See http://www.entfemp.ielpresV2000/210700.htm (accessedon October 7th 2006).
Q Springer

is availableat www.sprinqerlink.com (http://www.springerlink.com/contenUt4606642B1673730/).
permission.
Adapting measuresof mathematical knowledge for teaching 179
on many issues,including education.6These differencesand similarities make keland a

suitable location to begin investigating how to adapt the measuresof mathematical
knowledgefor teachingfor use outsidethe united States;we might expect to find some,
but not radical, differencesbetweenthe two countries.For example,similar algorithmsare
found in both countries but minor differencesin layout may be noticed. Furthermore,
becauseEnglish is the principal languageof public schoolingin both countries,we could
study the effect of using the measuresof mathematicalknowledge in a different country
without the addedcomplication of translatingfrom one languageto another.
Both qualitative and quantitative methods were used to describe, and analyze the
successof, the adaptationsto items. Qualitatively,we documentedthe processof selecting
items, and then carefully recordedthe adaptationsmade for Ireland. when making the
adaptations,we conducteda focus-groupinterview with four practicing Irish teachersand
with one mathematician.In addition we conductedfollow-up interviews with five (dif-
ferent) teachersafter they had completed the test, to learn about their responsesto the
items. Quantitatively,we studiedhow the measuresperformedpsychometricallyin a pilot
study.
Approximately 110 item stemsTdeveloped by the LMT project were available for
selection.Theseitems had previously been piloted on various survey forms in the United
Stateson samplesof teachersranging in numberfrom 104 to 659.8The items were mainly
piloted in California's MathematicsProfessionalDevelopmentInstitutesbetween2001 and
2003 (Hill 2004a4).
The items were administeredto a conveniencesampleof 100 Irish primary teachersein
summer 2004. The teacherswere attending summer institutes in a range of curriculum
areasincluding visual arts, language,social, environmentaland scientific education,and
mathematics.Thesecourseslast for twenty hours over five days and teacherswho attend
them during their holidays may take three days of extra personal vacation during the
following school year. The lead author approachedteachersdoing these coursesin six
centersand invited them to participatein the study.Teacherswere told that it would take
about 90 minutesto complete the test, and participantswere offered a token of appreci-
ation. Teacherswere requiredto completethe surveyin the presenceofthe researcherand
no time limit was imposed.
After completing the questionnaire,follow-up interviews were conducted with five
teachers.A conveniencesample of teacherswho were willing to spend one extra hour
answeringquestionsaboutthe test was selected.Teacherswho completedthe questionnaire
towardsthe end of the testingperiod weremore likely to be invited to do the interview than
thosewho did the test earlier and only teacherswho completedthe questionnairein Dublin
were asked to participate in the interviews. The interviews were intended to yield data
about how well teachersbelievedthe items reflectedthe work of teachingmathematicsin
primary schoolsin Ireland. About sevengeneralquestionswere askedto elicit teachers'
" For example see http://www.ed.gov/news/pressreleaseV2003/ll/l I l92003.html (accessedon October 7th

2006).
7
Some item stems have three or more items attached and others have just one.
E These
numbers refer to teacherswho took the pre-test only and not to teachersfor whom pre- and post-test
data were available.
e
At least 93 of the teachers have spent I year teaching mathematics (and all other subjects) at primary
school level and most have spent longer. Seventy-nine teachers received their teacher certification from a
pre-service programme in the Republic of Ireland, 13 from Great Britain or Northern lreland, one from the
United States.One is certified to teach secondaryschool mathematics,one is qualified to teach other subjects
in secondary school and five did not supply information about their qualifications or experience.
Q Springer
fhe Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

is availableat www.springerlink.com (http://www.sprinqerlink.com/contenVt4606642B1673730/).
Reproduced with permissionof the copyrightowner.Furtherreproduction
prohibited
without
permission.
S. Delaney et al.
180
work of primary
general opinion on the items, the extent to which they related to the
characters in the questions, the cladty of the language,and
Ieaching,ihe authenticityof the
are likely to find difficult. The teachers were also askedif
the questionsthat respondents
items, where they draw
they'could think of teachingsituations,not reflectedin the chosen
the test were selectedand
on their mathematicalknowledge.In addition, 16 questionsfrom
they chosethat answer
intervieweeswere askedto statewhat answerthey gave and why
and not anotheranswer.Theseinterviews were recorded and transcribed.
analyzedquanti-
Teachers'responsesto the multiple-choiceitems were compiled and
item difficulties and slopes (how well
tatively using psychometricmethods.Previously,
with similar levels of knowledge) were calculated in the
items discriminateamongteachers
to independently calculate item difficulties based on Irish
U.S. contextbut it was iriportant
we wanted to know how the U.S. items performed in
teachers,responses.Specifically,
would shed light on whether they were measuring similar constructs
Ireland, becausethis
into BILOG-MG
in the two countribs. We entered the responsesof the Irish teachers
we inspected the point
(Zimowski et al. 1996)and conductedthreepsychometricanalyses:
them to the U.S. point biserial
biserial correlation estimatesfrom Ireland and compared
each domain of the test; and we compared
estimates;we calculatedcronbach's alpha for
Theory (RT) model in
the relative item difficulties from a one-parameterItem Response
(Rasch) model because the samplesize was too
each setting.We chosea one-parameter
model. The point biserial correlation indicates how an item
small to use a two-parameter
The higher an item's point biserial correlation, the stronger
correlateswith all other.items.
between the item and the construct being measured' Another way of
the relationship
the better the item can
thinking about this is that the higher the point biserial correlation,
on the underlying construct.An
discriminatebetweenindividuals who are closer together
poorly, in that it is mostly
item which has a low point biserial correlation discriminates
of items vis-d-vis one another' Bilog-MG
measuringnoise.Difficulty refersto the difficulty
in standard deviations, on a scale where 0
(Zimows[i et al. 1996)reportsitem difficulties
abitity. Difficulties lower than 0 describe easier items and
representsthe averaget"*h".
more difficult items. Scale reliability measures the consistency of
hiiher difficulties reflect
testing occasions' A
test-takers,scores achieved over multiple items or over multiple
alpha and it is
widely usedmeasureof reliability from classicaltest theory is Cronbach's
responses are highly corre-
reportedon a 0-l scale.Higher figures indicate that the item
laied'and that they ur" measuringthe samething. Another way of thinking
"oniirt"ntiy differences"' that is
about this is that it measuresthe amount of "observed individual
(Cronbach and Shavelson 2004, p. a01) in teachers'math-
attributableto true variance"
If the point biserial correlations, the difficulty levels and
ematicalknowledgefor teaching.
differed among countries, this would suggest that the items were
the overall reliabilities
performing differently acrosscultures'
Results
outside the United

This article is primarily about how the items performed in a setting
performed on the items. Therefore the results begin with a
Statesrather than how ieachers
were selectedfor inclusion in the test' Changes to the items are
descriptionof how items
teachersrespondto the
then carefully documentedbecausethesechangesmight affect how
after they had completed
items.Next we presentthe resultsof interviewswith five teachers
of item performance
the test.Finally, we presentthe findings from psychometricanalyses
in Ireland and the United States'
Q Springer
fhe Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

is availableat www.springerlink.com (http://www.sprinqerlink.com/contenUt4606642B1673730/).
permission.
Adapting measuresof mathematical knowledge for teaching l8l
Table I Number of items in each category chosen for pilot test of items in Ireland
Item category Numbers and Pattems, functions Geometry

operahons and algebra
Content knowledge 28 25 3l 84
Knowledge of students 24 aA
and content
Total 25 JI 108u
" The 108 items were based on a total of 49 item stems
Item selection
The lead author selectediiems for use in Ireland basedon his knowledgeof the work of
teachingmathematicsin Irish primary schoolsand on resultsfrom psychometricanalyses
of item pilots in the U.S. From the bank of approveditems,39 item stemswere selectedfor
use.Nine of the 110 U.S. item stemsavailableat the time were excludedfrom consid-
eration becausethey relatedto topics that are not generallytaught at primary schoollevel
in Ireland, but instead appear mainly in the secondaryschool curriculum. The topics
includeddivision of fractionsby fractions,abundantnumbers,division of a whole number
by a non-unit fraction, rotational symmetry, coordinate geometry, equations with two
variables,linear functions, and exponentialgrowth.
The flnal set of items selectedfor the Irish study containeditems from each of four
scalesthat had beenidentified in the U.S. studies:contentknowledge(including CCK and
SCK) of number and operations,content knowledge of patterns,functions and algebra,
contentknowledgeof geometry,and a further set of number and operationsitems which
requiredknowledgeof contentand students(KCS). Ten additional item stemswere used:
three "items-in-progress"belongingto the LMT Project and sevenwritten specificallyfor
use in lreland. Theseten item stemsdo not form part of the analysisin this article because
they have not been subjectedto the rigorous item vetting process(Bassand Lewis 2005).
The total numberof items selectedin eachmathematicalstrandand in eachdomain of the
constructis summarizedin Table 1.
The most important criterion for selectingitems was alignment with the knowledgeof
primary mathematicsteachingand its mathematicaldemandsin Ireland.As in any situation
whereitems developedexclusivelyin a country are selectedfor usein the second,we could
not chooseitems to reflect MKT usedin the secondcountry (Ireland in this case)that does
not arise in the first.
Item adaptation
After item selection and initial adaptationsby the lead author, items were reviewed in
advance of test administration by a focus group of four experienced,practicing Irish
teacherswho helped determinewhether the items reflectedsituationsthat would arise in
Irish classrooms.Further modificationsof items followed, basedon the focus group dis-
cussion.Finally, one Irish mathematiciancritiqued the changesto determineif they were
mathematicallysound The instructionsto the focus group teachersand to the mathema-
tician, noneof whom had prior knowledgeof the constructof mathematicalknowledgefor
teaching,were to acceptiterns without change,or to proposeadaptationsto items so that
they soundedrealistic to Irish teachersand so that they preservedmathematicalintegrity.
@ Springer I

is availableat www.springerlink.com (http://www.sprinqerlink.com/contenUt460664281673730/).
permission.
i'
t82 S. Delaney et al.
We compileda table listing all the changesthat were madeto the items in adaptingthem
for use in Ireland. Of the 42 item stemsthat were originally written in the United States
only 7 were unchanged.Having considereddifferent ways of categorizingthe changesthat
were madeto the items, we decidedthat the following four categoriesbest summarizethe
changes:
l. Changesrelatedto the generalculturalcontext;
2. Changesrelatedto the schoolculturalcontext;
3. Changesrelated to mathematicalsubstance;
4. Otherchanges.
Translationsthat relatedto the generalcultural contextincluded:changingpeople'snames
to make them soundmore familiar to Irish teachers;alteringspellingsto reflect differences
betweenU.S. and British English; adaptingnon-mathematical languageand culturally
speciflcactivities to give the items a local flavor. This categoryof changesis summarized
in Table 2. These changesare important in making the items appear authentic to Irish
teachersand similar changeswere made to TIMSS items (Maxwell 1996).
The secondcategoryof changesrelated to the cultural context of the school or of the
educationsystemin general.This includedlanguageusedin schooland structuralfeatures
of the wider educationsystem.Examplesare given in Table 3. Adapting languagerelated
to the generalcontextor to the schoolcontextof the items is important so that teachersin a
second location are not distractedby unfamiliar terms or contexts. Knowledge of the
educationsystemin the "target" country is requiredto make suchchanges.Thesechanges
do not affect the mathematicalsubstanceof the items and thereforeare unlikely to com-
promise the items' ability to measuremathematicalknowledgefor teaching.
Table 2 Exemplars of general contextual changes made to items and frequency of changes
Type of change Example from original Example from adapted Number of items changed
U.S. form Irish form (Max. 98)
People's names Mr. Ives Mr. Fennelly 69

Mrs. Kwon Mrs. Kehoe
Nathaniel Darren
Spellings Factorization Factorisation
Realize Realise
Non-mathematical Cookie Biscuit
language Candy Sweets
Movie theater Cinema
Activities Baking cookies Baking scones
Table 3 Exemplars of school contextual changes made to items and frequency of changes
Type of change Example from Example from Number of items

original U.S. form adapted Irish form changed (Max. 98)
School language Students'papers Pupils'work

Text Textbook
Structure of education The schooldistrict's The schoolcurriculum
system curriculum
University Collegeof Education
A Springer

is availableat www.springerlink.com (http://wvvw.sprinqerlink.iom/contenvt+ooo5+zg16zgzgo4.
Reproduced with permissionof the copyrightownerfurther reproOuctionEotriOiteO
witnout
permission.
Adapting measuresof mathematical knowledge for teachins 183
In contrastto the previoustwo categories,the third categoryrelatedto the mathematical

substanceof the items. Units of measurementwere changed;suchchangeswere similar to
ones made for TIMSS and PISA, although fewer were necessaryin those assessments
becausemost participating countries used a common system of measurement(metric)
(Maxwell 1996; wilson and Xie 2004). of the 98 items that originated in the LMT
database,12 items (lLVo) requhedadaptationof the measurement units. In some cases,
these translations were straightforwardand the senseof the item was preserved.For
example,$20 might become€20 or l0 yards might become10 metres.In mosrcasessuch
adaptationsare similar to context changessuch as changesof namesand spellings.
Not all changesof measurementunits, however,were so straightforward.In four items,
the numerical quantity also had to be changedin order to preservethe senseof the item.
For example,sweetsare often sold by the ouncein the United States,but sweetsare rarely
sold by the analogousunit in lreland, the gram. This promptedfurther changesin the item:
insteadof a graph in intervalsof single ounces,the item now requireda graph in intervals
of ten grams.This type of changewas more problematicthan the generalcultural changes
mentionedearlier becauseit risks making the mathematicsmore difficult for one popu-
lation taking the test.
Other mathematical changes referred more specifically to teachers' mathematical
knowledgeand are less likely than the one just describedto have occurredin TIMSS or
PISA (seeTable 4). They included changesto the mathematicallanguagethat is usedin
schools,changesto representationsthat are commonly used in schools and changesto
Table 4 Exemplars of mathematical changes to items and frequency of changes
Change Example from Example from adapted Number of items

original U.S. form Irish form changed(Max.98)
Units Pennies One-cent coins t6

$ €
Inch Centimetre
I ounce l0 grams
School Mathematical Ones Units
Language
Polygon base Polygon face
Polygon with only two 2-D shape with only
sides two straight sides
Representation
1/\.8 0 180
9x20
-</ \ ,/,/\
(9) 20 ,/
9 x 4xs
V /.,, \
9 x 2x2 \ 5
4 (5)
,4. \-./
t,)€J
Anticipated student solution (7 x 8) "Seven times five (7 x 6) "Seven fives are
methods is thirty five, forty+wo, thirty five, thirty-six,
forty-nine, fifty-six" thirty-seven, thirty-
eight, thirty-nine, forty,
forty-one, forty-two"
6 Springer

is availableat www.sprinqerlink.com (http://www.sprinqerlink.com/contenVt4606642B1673730i).
Permission.
S. Delaney et al.
184
anticipatedstudentresponses.The mathematicallanguageusedin Irish schoolsdiffers in

somewaysfrom languagethat is usedin the U.S., resultingin changesto l'|Vo of the items.
value,
In most cases,precisetranslationsfor the termswere possible.When teachingplace
for instance, in Ireland the term rznir is used in the national curriculum whereas one is
typically used in the United States.A rightangled ffiangle in Ireland is more commonly
Jt"O u right triangle in the United States.Studentsin the United Statesencountetinput'
output maihines and Irish students:use function machines.In the United Statesa flat shape
is calledaplanefigure or a2-Dfigure whereasin Irelandit is calleda2-D shape.Ineachof
these casesan exact translation from the U.S. to the kish term is possible. The term
.,polygon," however,is frequently used in school texts in the United Statesand rarely in
lrist-t"^ts, where the term 2-D shapeis preferred.The problem is that polygon and 2-D
shapeare not synonymsbecausethe set of 2-D shapesincludes circles, semicirclesand
ovaisand the setofpolygonsdoesnot. Therefore,qualifyingstatements (e.9.,a 2-D shape
with straight sides)may need to accompanythe terms in a particular country.
Repres-entations of mathematicalideascan also vary from country to country' Take, for
an item that included a representationofa factor tree.Although the changes(see
"*urnpl",
Table 4) may seemsuperficialto a mathematician,all four Irish teacherswho saw the U.S.
representationprior to the pilot test agreedthat it neededto be modified in order to appear
familiar to Irish teachers.With ttre item changedas recommendedby the teachers,74Voof
the 100 lrish teachersrespondedcorrectly to this item. In the subsequentlarge scalestudy,
where the U.S. diagram was included unchanged,only 587oof a different sampleof 503
Irish teachersrespondedcorectly. It is possiblethat the unfamiliar diagramdistractedIrish
teachersin the secondstudy, negatively impacting their scores'
The four Irish focus-groupteachersalso believed that the method usedby a studentto
- 3 and then skip counting
multiply two single-digitnumbers,a x b, (multiplying a by b
6 x 9 solvedby calculating 6 x 6 : 36, 42, 48' 54)' in a U'S'
tt..e *ottipl", of o;
".g.,in Ireland and they agreed that multiplying two numbers and
item would be unusual
countingon in oneswould be more likely. Howevercountingin onesfor 3 x a would also
be unrealisticso the context was changedto have the studentin the Irish item count on a
ones.As a result the mathematicaldemandof the item was slightly changedfrom caleu-
Iating a x b to calculating a x (b - 2) (see Table 4). This change again highlights a
chaGnge faced in adaptation.Is the mathematicalknowledgein this item about knowing
how studentsarelikely to respondto a computationaltask?If so,the adaptationseemsto be
appropriate.If, however,the mathematicalknowledgebeing measuredis a teacher'sability
- (5 x 7) + (3 x 7), then the changeis not
,oûpptythe distributive property, (8 x 7)
upp.opiiut" becauseit simply makes the same task easier,(6 x 7) - (5 x 7) t (1 x 7)'
In-order to assist the accuracyof translation in the 2000 PISA study, test developers
provideda "short descriptionof the QuestionIntent" (Grisay2002,p.60) for eachitem;
ihis featurecould be developedfor the mathematicalknowledgefor teachingitems in the
future.
The final categoryrefers to changesnot necessitatedby cultural requirements.These
included deletion of an answerchoice, alterationsto visual appearance(e.g., font size or
typeface),furtherclarificationof an item or alterationto the meaningof a distractor.Because
thesechangeswere not made in responseto possibledifferencesin mathematicalknowl-
US
edge,and the items had previouslybeenthrough a rigorousdevelopmentprocessin the
(seleBassand Lewis 2005),the changeswere unnecessary. It is conceivable,however,that
in some cultures (e.g., where multiple-choiceformats are unfamiliar to potential respon-
dents) such changesmay be necessary.Such changesmay be problematic becausethey
could make an item easieror more difficult; they may make it more or less discriminating;
fl Springer

is availableat www.springerlink.com (http://www.sprinqerlink.com/content/1460664281673730/).
permission.
Adapting measures of mathematical knowledge for teaching l6)
and they may also changehow effectively the items measurethe underlying construct.In
this instance,however,they do not shedlight on the cross-culturalnatureof the constructof
marhematicalknowledgefor teachingand thereforethey will not be discussedfurther here.
Figures2 and 3 proyide an example of how some of the principles, identified above,
were applied and they illustrate how an item was adapted.This specificitem was not used
in the Irish test but a similar one was included.lo
o There are no people's namesin the item, no changeneededthere.
r The $ sign would havebeenchangedto a € (Euro) but the valuesstill make sense.
o No changeneededbecauseof different school mathematicallanguage.
o No changeneededbecauseof differencesin the educationsystem.
. "Taffy" could be changedto toffee, but making toffee is not an activity that is familiar
to most Irish children so that part would be changedto a scenarioinvolving making
Rice KrispierMbuns.
. A final type ofchange adjustedthe item to accountfor a differencein the Irish primary
curriculum which requiresstudentsto divide whole numbersby unit fractions (and not
to divide fractionsby fractions)so 1% was changedto 3.tr
The original item is reproducedin Fig. 2 and the changesthat were made are reflectedin
the rewritten item in Fig. 3.
Follow-up interviews
After the items were adaptedand administered,we interviewedfive of the Irish teacher-
respondentsabout the items in the test. We wanted to investigatehow successfulthe
translationof the items from the U.S. context had been by asking if the items sounded
authenticto Irish teachersand whetherthe items relatedto the work of teachingmathe-
matics in Ireland.
Four of the five teachersinterviewed believed that the teachersand the students
describedin the items seemedauthenticand the fifth teacherremarkedonly that severalof
the charactershad Irish names.Most of the teachersalso commentedthat the languagein
the items wds clear, althoughtwo commentedthat some of the mathematicalterms were
unfamiliar. One teacher was unsure of the meanings of congruent and tessellateand
anotherteacheralso included congruent,tetrahedron,and polygon among the terms that
were difflcult. Theseunfamiliar terms,which were both specificand exceptionalin the test
as a whole, point to terms that might need to be translatedor explained in a future
administrationof the questionnaire.12One of the teacherssuggestedthat "if you're doing
shapeswith a class, you're going to look at [the relevant vocabulary] beforehandand
you're going to kind of study it and know exactly what [the students]ask you" (Pilot
lo Many of the items used in the test are not released and therefore cannot be reproduced in this article.
Released items can be accessed at the website: http://www.sitemaker.umich.edr-r/lmt/files/
LMT_sample-items.pdf.
rr
Although this change would probably reduce the challenge of the word problem for primary school
students (Greer 1987), we do not know if it would affect the difficulty of the item for teachers.
12 It is interesting to note that "congruent" is one of the terms which TIMSS was requested to change.
"Same shape and size" was considered an acceptabletranslation but "equal" was considered too imprecise
(Mullis et al. 1996, pp. l-6).
0 Springer

is availableat www.springerlink.com (http://www.sprinqerlink.com/content/14606642816737300.
Permission.
S. Delaney et al.
to illustrote
WhlA"f th" f.ll"*ing story problemscouldbe used
rldividedbylt(MorkyEs,NO,orI'MNOTSUREforeachpossibility')
I'm not
Yes No sure
o) Youwont to split rf piesevenlYbetween

two families.Howmuchshouldeochfomily
get?
b) Youhove$1.25ond moysoondoubleyour
money.Howmuchmoneywouldyou end uP
with?
c) Youore mokingsomehomemode toffy ond

the recipe collsfor t| cupsof butter'
=
Howmonysticks of butter (eqchstick
need?
t cup)will You
Fig. 2 How an item looked before adaptation
illustrate
Whi.h th" f"llowing-tory problemscouldbe usedto
"f or eachPossibility')
3 dividedby iZ lfnotXyES, NO, or rM NOT SUREf
f'm not
Yu No sure
o) Youwontto sPlit 3 piesevenlybetweentwo
fomilies. Howmuchshouldeochfomilyget?
b) Youhove€3 andmaysoondoubleyour
money.Howmuchmoneywouldyouendup
with?
c) Youore makingR'iceKrispieobunsondthe
bqrs of chocolote(eochbor =
Fig.3 How an item looked after adaptation (changes underlined)
for respondingto a questionor

transcript2358).This doesnot, however,preparea teacher
statementfrom a studentthat occursout of context'
-a -i;;".;i;r
not com-
parricipantsalso identified as difficult, items featuring materials
referred to tetrominoes and another to
monly available in lrish schools.One such item
they did not understand the item that used the
uig.u.u tiles. Two of rhe teacherssaid
understandthat at all' I couldn't
algebratiles: "I never saw that sort of illustration.I didn't
Q Springer

is availableat www.springerlink.com (http://www.sprinqerlink.com/contenVt4606642B1673730/).
permission.
Adapting measuresof mathematical knowledge for teachins 187
work that out" (Pilot transcript l29l). A, third respondentfound it "a bit taxing" because
"you're transferringalgebraicthinking into graphics" (pilot transcnpt 2346).
When the Irish teacherswere askedif they thought that the mathematicsin the items
was the kind that teachersencounterin Irish classrooms,they all agreedthat, in general,it
was. Three teachersqualified their answersto this question.One believed that most Irish
studentswould not producethe type of mathematicsproducedby the hypotheticalstudents
in some of the u.S. questions.Anotherteacherspeculatedas to why this might be. He
commentedthat someof the mathematicsincluded in the items would not be observedin
an Irjsh classroom.He noted that a relatively high proportion-25 out of 108-of the items
in the Irish test relatedto pattems,functionsand algebraand Irish studentsdo not currently
spend much time on the algebra strand of the curriculum. Consequently,if the algebra
items, for example, included in the Irish questionnaireare based around mathematical
conjectures,questionsor other utterancesthat Irish studentsare unlikely to produce in
class, it is possible that these items do not reflect mathematicalknowledge that Irish
teachersusewhen teachingmathematics.Since 1999,however,algebrahas beena specific
strandof the Irish primary mathematicscurriculum.
Notwithstanding these commentsfrom the interviewees,all the Irish teacherscom-
mentedthat the items seemedauthenticbasedon their teachingexperience.Items relating
to content or contextsthat were unclear,difficult or unlikely to occur in Irish classrooms
were exceptional.
Overall the follow-up interviews we conducteddemonstratedtheir potential to yield
important data in relation to translatingitems. We collecteddata that helpedus to explain
unexpectedresponsesand reasonsfor errors that may be country-specificand that sug_
gestedalternative,credible responseoptions that are not currently included in the items.
The interview data also helped us to check if the charactersin the questionsseemed
authentic to teachersand if the situationsarose when teachingmathematicsin the new
setting.In addition,we usedthe interviewsto identify languagethat causeddifficulty for
respondentsin the new setting. We found the interview data to be less effective for
identifying contentthat arisesin the work of teaching,but that is not coveredby the items
and which might form the basis of future items.
Because we interviewed only five teachers,we do not claim to have definitively
resolvedall the issuesraisedin the previousparagraph.We have,however,identifiedareas
that are worthy of further investigationand gatheredevidenceto supportthe potential of
more systematicuse of follow-up interviews in future translationof MKT items.
Psychometricanalyses
Datafrom the sampleof 100 Irish teacherswere analyzedand comparedto 6ataobtained

from teachersin the United States.Although neither group of teachersconstituteda ran-
dom sample,IRT modets are consideredto be robust to this difficulty, becauseIRT is
focusedon comparingitems, ratherthan individuals.We beganby inspectingpoint biserial
correlationsfor number and operationscontent knowledge (CCK and SCK) items. The
point biserial correlationsof content knowledge items in both settings were not on an
interval scaleso we first performeda FisherZ transformationon them. Figure 4 correlates
the U.S. and Irish FisherZ transformedbiserialcorrelationsof contentknowledgeitems. In
generalthereis a moderatbcorrelationbetweenrhepoint biserials(r:0.43) but item ld,
which is more than two standarderrors from a best-fit line between the two sets of
estimates,is a clearoutlier.
Q Springer

permission.
S. Delaney et al.
188
o
,F u.o
-6
E
o..
9 o.o
o.e
.q
o
E
o.4
^j
o
@
0.2 o ltem Ic
o
.t2 Item 1d
u-
s 0.(
0.0
.9
o.4 0.5 0.6 0.7 0'8 09

o.g
U.S. FisherTransformedBiserialCorrelations
Fig.4AregressionlinefittedtoascatterplotoftheU.s.andlrishFisherZtransformedbiserial
correlations
a representationwas an accurate
Item ld required the respondentto considerwhether
Like items la, lb and 1c, the only adaptations
representationof one-fourth(seeFig. 5a).
a change in school language in the item stem' Although many
were a changeof name and
reproducedin Fig' 5(a)
respondentsin both countries believed that the representation
that someIrish teachersbelieved
representedone-fourth,the follow-up interviewsindicate
interviewedexpresseduncertainty
that it representedone-eighth.Three of the five teachers
only on the upper part of
about whether the dark lin" *u, deliberatelyor unintentionally 'that's
struck me, one-eighth' the
the horizontal line. one teacher commented: "That
(Pilot transcript2346). The combinationof
minureI saw it. It didn't strike me as a quarter"
the point biserial correlations suggest that this item
evidencefrom the interviews and
how U.S. and Irish teachers interpret linear representations
surfacesa differencebetween
it as an area,rather than a linear
of fractions.SeveralIrish teachersappearedto consider
one fourth' In the united states'
representation,with just 62Voof them identifying it as
g5voofrespondentscorectlyidentif,edthediagramaSrepresentingone-fourth.This
conjectureissupporteduyoefinaingofthesubsequentstudyoflrishteacherswherethe
was shadedaboveand below the line
."pi"sentution *u, ctangla so that the numberline
(seeFig. 5b). Selectionind adaptationof items for the subsequentnational study of Irish
teaching was informed by findings from the pilot
teachers,mathematicalknowledgefor
percentagethat correctlyidentified
study.of 503 sampledteachersin the secondstudy,the
in the revised diagram rose to 92Vo,much closer to the u's'
the fraction as one-fourth
continuedto usean areaconception
finding.It is possibie,however,that the Irish teachers
this question so an alternative
of fractions rather than a linear one when answering
Fig. 5 (a) Linear rePresentauon (a)'rltl

of a fraction fiom a content
knowledge item in the actual lllrl
study. (b) Revised linear
representationof a fraction from
a content knowledge item in the
(b)
follow-up studY
Q Springer

is availableat www.sprinqerlink.com (http://www.sprinqerlink.com/contenUt460664281673730/).
Reproduced with permission
of the prohibited
copyrightowner.Furtherreproduction without
oermission.
Adapting measuresof mathematical knowledge for teaching 189
representationmay be neededif the goal were to elicit Irish teachers'understandingof

linear representationsof fractions.This result showsthat teachers'performanceon items
across countries can be sensitive to what may appear to be minor differences in
representation
In the number concepts and operationsknowledge of content and students(KCS)
domain,one item had a negativepoint-biserialcorrelationscoreof -0.051, indicating that
respondentswho scoredwell on other items in this subtestwere slightly more likely to get
this wrong than right. The item askedteachersto identify similar errorsthat studentsmade
when adding three two-digit numbers,such as failing to heed place value, and failing to
note a carry of 2 tensratherthan I ten. Our initial hypothesisfor this differencecenteredon
the fact that the algorithm is typically recordeddifferently in lreland. In the secondstudy,
where the algorithm was recordedas it usually is in lreland, the point biserial correlation
rose to 0.306. Again what appearedto be a minor variation in the layout of an algorithm
may have affectedteacherperformanceon the item (Table 5).
After examining point biserial correlations,we inspectedrelative item difficulties to
determinewhether the sameitems, in general,were easy and difficult in both countries.
Overall, Fig. 6 and Table 6 show quite a strong correlation(0.74) betweenthe items that
Table 5 Point biserial correla-

Item number U.S. form on which Irish point U.S. point
tion estimates for number and
on Irish form item first used n biserial biserial
operations content knowledge
estimates estimates
items ordered by estimates on
Irish test form
2a c0r 0.679 o.702
6d 402 o.574 0.622
2b c0l 0.547 0.487
3b A0l 0.513 0.495
2c c01 0.513 0.47|
4b c0l 0.487 0.542
3a B0l 0.470 4.527
3c A01 0.450 0.562
5d c0l 0.412 0.449
6b A01 0.403 0.518
5b c0l 0.368 0.528
4a c01 0.366 0.553
5a c01 0.359 0.380
JO B0l 0.347 0.565
lb B0l 0.326 o.534
5e c01 0.309 0.588
la B0l 0.309 0.550
6a A01 0.304 0.418
5c c01 0.288 0.413
7 A0l 0.276 o.479
6c A01 0.249 0.296
4c c01 0.218 0.529
3e A0l 0.r99 0.429
o B01
Form number refers to the form lc 0.156 0.548
as described in technical reports 1d B0l 0.020 0.538
by Hill (2@4a-d)
Q Springer

permission.
S. Delaney et al
t90
L000
0.000 o
llem ld o
E o
o
i5 OO
.c -1.000
o
"
o*"o
;o
d
-?.000 o
a
ul
-3.000
ItBmI c
ftem1e
-4.000
-,1.g00 -3.000 '2.000 -1'000 0'000 1'000
USDifrs
Fig.6AregressionlinefittedtoaScatterplotoftherelativedifficultiesofitemsinthelrishandU.S.
section of the test
n"irion, of tie number concepts and operations
Table 6 DifficultY estimates for U.S. difficulties (SE) Equated Irish

number and oPerations content difficulties (SE)
knowledge items with equated
Irish difficultY estimates -2.r'18 (0.232) -3.64387(0.650)
la
-2.319 (0.246) -2.r0639 (0.365)
lb
-1.993(0.215) -3.40017(0.576)
-2.178 (0.232) -0.70s74 (0.256)
1d
-0.103(0.201) 0.091738(0.300)
2a
-0.726 (0.211) -0.4669 (0.289)
2b
-r.478 (0.244) -1.39959(0.319)
2c
- 1 . 5 1 4( 0 . 1 8 2 ) -0.85471(0.292).
3a
-2.222 (0.t65) -1.28219(0.313)
3b
-0.00137(0.279)
3c 0.323(0.121)
-0.197(0.143) 0.045589(0.273)
3d
-1.093(0.130) -0.18677(0.261)
3e
-t.o75 (0.223) -0.90572(0.285)
4a
-1.678(0.258) -1.s8742(0.331)
4b
-0.1'74(0.212) -1.52265(0.310)
4c
-1.128(0.22s) -2.02219 (0.360)
5a
-0.971(0.219) -2.2817s (0.392)
5b
-0.00137(0.268)
5c 0.286(0.203)
0.329(0.203) 0.83902(0.30o)
5d
-1.826 (0.269) -2.022t9 (0.357)
- 1 . 1 5 3( 0 . 1 3 1 ) -r.46031(0.310)
6a
-0.823 (0.125) -1.39959(0.312)
6b
-0.044 (0.120) -0.6094(0.269)
6c
-1.136(0.100) -2.28'775(0.405\
6d
'7 0.s96(0.123) 0.233422(0.268)
@ Springer
TheJournalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

is availableat www.sprinserlink.com (http://www.sprinqerlink.com/content/t460664281673730/).
permission.
Adapting measuresof mathematical knowledge for teaching l9l
Table 7 Coefficient alpha results for each domain of the test piloted in Ireland
Domain of mathematical knowledge for teaching Number of items Cronbach's alpha measure
Numbers and operations, content knowledge 28 0.745

Numbers and operations, knowledge of students 1A
0.563
and content
Patterns, functions and algebra content knowledge 25 0.811
Geometry JI 0.804
Irish and U.S. teachersfound difficult. We did, however,identify three outliers. Item ld,
which featured the linear fraction representation,was relatively more difficult for Irish
teachersand two other items (la and lc) relating to the same stem on the topic of
representationsof % were relatively more difficult for U.S. than Irish teachers.Items la
and lc require respondentsto state whether or not different representations-one area
model and one discretemodel-show %. Without further investigationno obviousreason
emergesto explain the why theseitems were relatively more difficult for the U.S. teachers.
We alsocalculatedCronbach'salphafor eachdomainof the test (Table7). We cannot
specifically compare these estimatesto those obtained in the United Statesbecausethe
items on the forms were different. However, the estimatesare broadly similar to someof
the early resultsobtainedin the United States(seeHill et al. 2004).The greatestdifference
can be seenin the domainof knowledgeof studentsand content(numbersand operations).
The knowledgeof studentsand contentreliabilities are noticeablylower than thosefound
in the other threedomainsin both settings,but in Ireland,the measureis lower than any of
the U.S. measures.This may indicate that studentshave different misconceptionsor dif-
ficulties or that they approachproblemsdifferently in the two countries,but further studv
would be neededto establishwhv.
Discussion
Adapting items developedin one country to measureteachers'mathematicalknowledgein

anotherenafles researchersto build on expertisethat has been developedby international
researchgroups.We have shown,however,that translatingsuchmeasuresfrom one setting
to anotherfairly similar setting is a non-trivial process.Despite the relative similarities
betweenthe United Statesand Ireland,we found differencesin mathematicallanguage,in
measurementunits, in representationsof mathematicalconcepts,and in knowledge of
studentsand content. Although such differences may be dismissed as mathematically
trivial, some evidence pointed to teachers' performanceson items being sensitive to
apparentlyminor variationsin items.With non-Englishspeakingcountries,or in countries
with markedly different ways of teaching mathematics,many more changes may be
required.Therefore,explicit guidelinesfor adaptingteachingknowledgeitems needto be
developed.Basedon the follow-up interviewsand on the psychometricanalyseswe believe
that our adaptationswere generallyappropriateand that the results sectionof this article
provides working guidelineson which those adaptingitems in future can build. If future
translationsbasedon theseguidelinesare documented,researcherswill be betterplacedto
interpret test results.becausethe changesmade, or the absenceof changes,may help to
explain differencesin teachers'scoreswhich are not attributable to differencesin their
knowledge.Further, apparentdifferencesin knowledge may be accountedfor by differ-
encesin the mathematicalknowledgeused by the teachersin the respectivecountries.
Q Springer

is availableat www.sprinoerlink.com (http://www.sprinqerlink.com/contenUt460664281673730/).
prohibited
without
permission.
S. Delaney et al.
t92
Thosewhowishtoadaptmeasuresofteachers'mathematicalknowledgewillalsohave
the United States-Irelandproject' A key question
to grapplewith other issu"s us we did in
is:whatconstitutesthemathematicalknowledgeforteachinginaparticularcountryand
whohastheexpertisetodecidethis?Muchcanbeagreedwithlittleargument.Teachersin
of algorithmsusedfor arithmeticoperations'to
a country needto traueo.ep understanding in the
be able to manipulate used iaterials and to know terms that appear
"#-onry is an important factor in determiningmath-
curriculum. Although ttre school'curriculum
e m a t i c a l k n o w l e d g e , o t h e r f a c t o r s a l s o i n f l u e n c e t h e m a t h e m a t i cand
a l kassessment
nowledgethata
interaction,materialsavailable
teacherneeds:typical forms of classroom
proceduresused,fbr"û*pt".Allofthese.factoJsneedtobeconsideredwhendetermining
in a particular country--
iir" .utt"*utical knowledgeusedin teaching
is tiat some mathematics topi"r upp". in different gradelevels in
one particularissue
e a c h c o u n t r y , l e a d i n g , i n t h e c u r r e n t s t u d y , t o e x c l u s i o n o f U ' s . i t e mthey
s b aare
sed ontopicsthat
based
to removeitems because
do not appearin the Irisi curriculum.The decision
primary curriculum illustratesone challengefaced
on topics that do not appearin the Irish
inusingthetestoutsidetheUnitedStates.Knowingcuniculumcontentisonecomponent
ofteachers,mathematicalknowledge,butfew*ould*gu.thatteachersneedtoknoworrly
teachersbe ableto divide fractionsby fractions'
the mathematicsof the curriculum.-should
f o r e x a m p l e , a n d b e a b l e t o r e s p o n d t o s t u d e n t s ' q u e s t i o n s a b o u t d i v i sDiscussion
ionoffractionsby
in the primary schoolcurriculum?
fractionseven if suchcontentis not included
with expertisein the theory of
among teachers,teacher educatorsand mathematicians
in local knowledge of teaching offers one
mathematicaltcnowteogeioi t"u"t ing and/or
way to resolve this issue'
teaching'Many item adaptations
Another issuecenterson the lexicon for mathematics
language and consideringhow termsmight
requiredpaying .lose utt"ntion to mathematical
Referring to an earlier example'it
be moie or less familiar to teachersfrom each"oontty'
couldbearguedthatbecausethetermspolygonand2.Dshapedescribeparticularmathe.
(or shouldbe included in the mathematical
matical concepts,tne term-polygon equiuatent;
(or equivalent)shouldbe part of the mathematical
discourseof Irish reachersand2-D shape
discourseofU.S.t"u"i'".,,evenifcurriculaineachsettingdonotusesuchterminology.A'
question.thereforeiswtratsnoutdbeincludedinthelexiconandhowshoulditchange
a m o n g e l e m e n t a r y , m i d d l e a n d h i g h s c h o o l t e a c h e r s i n t h e U n i t e dby Statesandbetween
questioncan be addressed studyingthe
primary and secondaryt"u.h.r, in ireland?The
working at particulargradelevels' in particular
languageteachersuse and encounterwhen
accuracyand adequacyof.the terminologyused'
countriesanaexaminingthe mathematical
teacherscurenrly hold, but we do not claim
This study identifleiknowledge that Irish
needto know' Teachersmay know what
that this is all the mathematicsttrat Irish teachers
reasonsand this knowledgemay influencehow
they know for historical, social or cultural
teachersteach.Teaching*"tt'oo'change.Teacherpreparationevolves.Curriculaare
reformed.Teachersmayneedtoacquirenewknowledgeandameasureofmathematical
knowledgemustbeuul"tou..o*modatesuchknowledge.ThinkaboutGulliveramongthe of height to
Lilliputians, ro, ii ,t. t-ittiputians had limited their measurement
"*u-p]".
themselvestheywouldhavehadnowaytomeasuretheheightofGulliverwhenhearrived
scalewas built with only people the size of
in their land. Similarly, if their measurement
would not measurethe locals precisely
Gulliver in mind, it is possiulethat the intervals
When we pilot our measureswe are both
becausethe intervals would be too far apart.
measuresthat can be used to measure
measuring current tnowledge and deveioping
mathematics'. evenif that knowledgeis not
knowledgethat we believe is usefulin teaching
causesa tension when adapting measures
currently widely helJ among teachers.This
@ Springer
The Journatof MathematicsTeacherEducationis publishedby Springer.The originalpublication

permission.
193
becauseif, for example,Irish teachersscorepoorly on a particular

item comparedto u.S.
teachers,this may be becausethe item is tappinginto knowledge
that Irish teachersdo not
use in teaching.Whetherit is knowledgethat Irish teachersshiutduse
in their teachingor
whetherthat knowledgeis peripheralto the work of teaching
in Ireland, requiresa judg_
ment call on the part of the test translator.
The writers of teacherassessments can supportthe adaptationprocessin at least two
ways' If test authors define the "Question Intent" for each
itemj adapterscould try to
maintain the integrity of the intent in their translation.Test
developerscould also inves-
tigate the possibility of training multi-disciplinary,multilingual
tea;s to vet items prior ro
their being usedin new countries.When items have been
selectedand adaptedbasedon
decisionsand criteria outlined above,changesto items might
be submittedtbr approvalto
a team of educatorsand mathematicianswith expert knowledge
of the mathematical
knowledge for teaching construct, prior to administeringthe
sirvey. This would help
ensurethat the adaptationsdid not compromisethe ability of
the ite;s to measureMKT.
When comparingteachers'knowledgeacrosscountries,it is possible
that adaptationof
items from one country alone may not be sufficient. One needs
to be cognizantof iden-
tifying potential new material for items that reflect distinctive
aspectsof mathematical
knowledge for teaching as observedin the new setting; aspects
thut u." not currently
reflected in the item bank developedbased on studying
the practice of mathematics
teachingin the United States.For examplein Ireland a national
mathematicscurriculum
exists for mathematicsteachingand a third grade teachercan
be expectedto know the
mathematicsthat a pupil entering the class from secondgrade
can be expectedto have
encounteredpreviously. This knowledge will not be tapp;d
in the existing u.s. items.
Another specificexampleis that the Irish curriculu. ,.quir",
studentsto draw, construct
and deconstructnets of 3-D shapes(Governmentof Iieland
1999) but no U.S. items
featuring nets have been developedto date.
This researchhas benefits for practicing elementaryschool
teachers,particularly for
teachersin the country for which the items are adapted.pre_service
and in_service
mathematicseducationcan be improved becausethe measures
would allow professional
developersto measureteacherlearningrather thanjust teachers,
level of satisfactionwith
professionalworkshops(wilson and Berne 1999).The process
of adaptingthe measures
also has the p.qtentialto inform teachersabout differences
in tvtKi 1rom country to
country. Such awarenesscould help teachen to preparefor
teachingin new settingsand
could enableteachermobility acrosscountries.
As with any study,the current study has somelimitations. First,
_ the sampresize of 100
Irish teacherswas small, restrictingthe analyseswe could conduct.
of theseteachers,only
five wpre interviewed about their responsesto items. Further,
the timing of theseinter-
views,:priorto analyzingdata,meantthatthe itemsdiscussed
in detailweri not necessarily
the items that performeddifferently than in the u.s. Third,
the forms that were usedin the
Irish test consistedof different items than on any single
u.S. form and rhis made psy_
chometriccomparisonsproblematic.Fourth,we did not have
analogousinterview datawith
U'S' teachersto comparetheir views of the items.Finally, althougliwe
speakof the United
Statesas an entity, the data were drawn from a handful of states-and
it is iossibte that the
constructmay differ among other U.S. statesto a similar greater
or extent than it differs
betweenIreland and the stateswhere U.S. data were collected.
Despitetheselimitations,
novel and multiple sourcesof data were usedto explorehow
mathematicalknowledgefor
teachingmay differ acrosstwo countries.
we lnow return to summarizeour answersto the four research
questions.First, we
encounieredfour methodologicalchallengeswhen using
u.S. items to measureMKT in
I e Springer

is availableat rnnrvw.sprinqerlink.com
(http://www.sprinqerlink.com/content/t+6066+28i6237300
of the copyrightowner.Furtherreproduction
pronlUiteO
witfrout
permission.
Fig. 7 Mathematical knowledge

for teaching in the United States The UnitedStates lreland
(ElementarySchool) (PrimarySchool)
and lreland
Ireland. One significant challenge is deciding what MKT is core for primary school
teachingacrosscountriesand what is specificto a country, determinedby factors such as
its curriculum, pupil interactionpatterns,assessment procedures,mathematicallanguage
and the teachingmaterialsavailable.In other words, if MKT in the United Statesand in
Ireland were representedon a Venn Diagram,what would be in eachof the sectionsA, B,
C? (SeeFig. 7). A secondchallengearosebecausethe items originatedexclusivelyin the
U.S. Therefore, familiarity with education and culture in both settings was neededto
translatethe itemsfor Ireland.A third challengearosefrom the fact that we found teachers'
performanceson someitems to be sensitiveto what might appearto be minor variationsin
representations. Fourth, in one item deciding on an appropriatetranslationwas hampered
by not havinga "questionintent" to accompanythe question.
With regardto our secondresearchquestion,whenadaptingthe itemsmostchangeswere
madeso that the U.S. originatingitems would appearrealisticto teachersin Ireland.For this
reasonnames,spellings,cultural activities, featuresof schools,measurementunits, repre-
sentations of concepts, mental arithmetic approachesused by students and school
mathematicallanguagewere changed.We envisagedthat other changesto the format of
questionsmay be neededin a settingwheremultiple-choicequestionsmay not be familiar to
potential respondents.We evaluatedthe successof the adaptationsby using follow-up
interviewsin which we askedthe Irish teachersif the itemsappearedauthenticto them andif
the mathematicswas of the kind that Irish teachersencounter.We also evaluatedthe item
adaptationsby looking at how the point biserial correlationestimatescorrelatedbetween
Ireland and the United States.Finally, the initial explorationssuggestthat with the transla-
tions describedabovethe vast majority of elementaryschoolitems developedin the United
Statesw.€resuitable for use with primary teachersin Ireland. The processalso revealed
instanceswheretranslatingitems, from one primarily English-speakingsettingto another,
was problematicand it raisedquestionsaboutadaptingMKT items that needfurther inves-
tigation.Thesequestionsinclude:to which adaptationswereteachers'performanceson items
more sensitive?Why are someitems more difficult for teachersin one settingthananother?
and why do KCS items havelower reliability measuresthanother items in both settings?
This article has describedmethodologicalchallengesthat arise when using U.S. mea-
sures of mathematicalknowledgefor teachingto measuresuch knowledgein other
countries.It suggeststhat intemationalcomparisonsof teachers'mathematicalknowledge
need to be consideredin the light of differencesthat may exist in the knowledge that
teachersuse in each country. Much work remains to be done to reach a greaterunder-
standing of the relationship and interaction between mathematics teaching and
mathematicalknowledgefor teachingin the United Statesand elsewhere.
Acknowledgements The researchreported in this article was supported in part by grants from the U.S.
Department of Education to the Consortium for Policy Research in Education (CPRE) at the University of
Pennsylvania (Grant #OERI-R308A60003) and the Center for the Study of Teaching and Policy at the
University of Washington (Grant #OERI-R308B70003); the National Science Foundation's Interagency
Q Springer

is availableat www.sprinqerlink.com (http://www.springerlink.com/contenUt4606642B1673730i)
' prohibited
without
permission.
195
Educational Research Initiative (IERI) to the University of Michigan (Grant

#s REC-gg7gg63 & REC-
0129421)' the william and Flora Hewlett Foundation, and the
Athlntic itiiuntrrropi"r, the Deparrmenr of
Education and Science (Ireland) grant R/D7l03 and by coldiste
tvttruireMarino, Dublin. opinions expressed
in this article are those.ofthe authors, and do not reflect the views
or trre u.s. nepartment ofEducation, the
National Science Foundation, the william and Flora Hewlett
Foundati.n, irre Atlantic philantkopies, the
Department of Education and Science (Ireland) or Coldiste Mtruire
MJno, ouutin. rn" authors would like
to thank Hyman Bass, Merrie Blunk, Carolyn Dean, Imani coffney,
rennir". l"rnir, Laurie Sleep, M;k
Hoover Thames for their help in developing aspects of this article.
The authors would also like""d to thank
Jeremy Kilpatrick and Larry Ludlow who ."ia ana offered feedback
on earlier drafts of this articte. Thanks
also to Dina Tirosh and four anonymous reviewers for helpful comments
which improved this article. Errors
are the responsibility of the authors.
References
An' S', Kulm, G., & wu,.Z' (2_004).The pedagogical content

knowledge of middle school, mathematics
teachers in china ang th: u.s Journar of Mathematics Teacher
Eiucation, 7, 145-172.
Andrews, P', & Hatch, G. (2000). A comparison.of Hungarian
ana Enllish teachers, conceptions of
mathemarics and its teaching. Educationar studies in lriathenatics,
ai g, l_e+.
Ball' D' L' (1988). Knowledge and reasoning in mathematical
pedagogy: i,xamining what prospective
teachersbring to teachereducation. unpublished doctoral dissertati'on]'uictrigan
stute uniu".rity, easi
Lansing.
Ball' D' L' (1990)' The mathematical understandingsthat prospective
teachers bring to teacher education.
The Elementary School Jouma!, 90(4),449466.
Ball' D' L', & Bass,]I (20o0)' Interweaving content and pedagogy
in teaching and learning to teach. In J.
Boaler (Ed.), Multipte perspectives on the teachiig olrd-ieamirg
oiman"matics (pp. g3-10a).
Westport, CT: Ablex.
Balt' D' L', & Bass, H. (2003a), Making mathematicsreasonable
in school. In J. Kilpatrick, W. G. Martin, &
D' schifter (Eds),! rlsea-rc! companion to pinciples and standards
for schiot mathematijcs6;.;7-
44). Reston, VA: National Council of Teachers of Mathematics.
Ball' D' L', & Bass' H' (2003b). Toward a practice'based theory
of nnthematical knowledgefor teaching.
Paper Presented at the Proceedings of the 2002 Annual M""iing
or tire Canaolan Mathematics Edu-
cation Study Group, Edmonton, AB.
Balt, D. L., Bass, H., Hill, H., Sleep, L., phelps, Thames, M. H. (2006). Knowing and using
mathematics in teaching. Paper Presented at the.G.,_&
Leaming Network Conference, T"*,i.;
a;il:
Quantity, and Diversity, washington DC, January 30-3i. Retrieved on January gth
200g from
http://www-personal.umich.edu/-dbalvpresentatio;/013100_Nse_rr,rsp.par
_
Ball, D. L.,,Thames,_M. H., & phelps, G. (2005). Knowkdge
oy iatneiatics jor tuaching: what makes it
speciar? Paper presented at the AERA Annual confeience, Montrdal,
_ Cunaau.
Bass, H', & Lewis, J. (2005, April 15). what's in collaborative
work? Mathematicians and educators
developing measuresof mathematical knowledge.for'teaching. paper pa"r"nt"o
at the Annual meeting
of the American Educational ResearchAssociation, Montr6al,
Canada.
Borko, H., Eisenhart, M., Brown, c. A., underhill, R.-G.,
Jones, o., a egara, p. c. (1gg2). Learning to
teach hard mathematics: Do novice teachersând^theirinsiruciors
giu" ,rp ,oo easlly? Journal
Research in Mathematics Educarion, 23(3), 194_222. for
Chrostowski,S. J.' & Malak, B. (2004).Translationand cultural
adaptationof the TIMSS 2003 instrumenrs.
In M o' Martin, I V. S. Mullis & S. J. Chrostowski (Eds.),
TIMSS 2003 technical report. chestnut
Hitl, MA: TIMSS & pIRLS Intemational Study Center, Bosion
Coilege.-
Cogan, L' s', & schmidt, w' H. (1999). An examinationor'inrt-.iion.ot
pLtices in six countries.In G.
Kaiser, E' Luna & I. Huntley (Eds.), International comparisons
trmoihe-ati"s
London: Falmer press. "ducation(pp.6g-g5).
cronbach, L' J', & shavelson, R. J. (2004). My cunent thoughts
on coefficient alpha and successorpro-
cedures. Education and psychological Measurement, 64A),
39 l4lg.
Eisenhart,M., Borko, H., Underhilr, R., Brown, C.,.Jones,o.,
['egaro, p. (1993). conceptual knowredge
falls through the cracks: complexities of leaming to teach mati'ematics'ior
unoerst anding.Joumalfor
Research in Mathematics Education, 24(l), g4O.
Fuson, K' c', stigler, J:.Y' & Bartsch, K: (1988). crade placement
of addition and subrraction topics in
Japan, Mainland china, the soviety Union, Taiwan and the
united iiite". torôt yor Research in
Mathematics Education, I 9(5), 449456.
Q Springer

is availableat www.springerlink.corn(http://www.sprinqerlink.com/contenVt460664281673730/;
prohibited
without
permission.
r96 S. Delaney et al
Gorgori6, N., & Planas,N. (2001). Teaching mathematicsin multilingual classrooms.Educational Studies in
M athematics,47(l), 7-33.
Covernment of lreland. (1999). Primary school cuniculum: Mathematics. Dublin, Ireland: The Stationery
Office.
Government of lreland. (2005). An evaluation of curriculum implementation in primary schools: English,
mathematics and visual arrs. Dublin, Ireland: The Stationery Office.
Greer, B. (1987). Nonconservation of multiplication and division involving decimals. Joumal for Research
in Mathematics Education, 18(l-), 3745.
Grisay, A. (2002). Chapter 5. Translation and cultural appropriatenessof the test and survey. In R. Adams &
M. Wu (Eds.), PISA 2000 Technical report. PNis: Organisation for Economic Co-operation and
Development (OECD).
Hambleton, R. K., Merenda, P. F., & Spielberger, C. D. (Eds.). (2005). Adapting educational and psy-
chological testsfor cross-cultural assessment.Mahwah, New Jersey: Lawrence Erlbaum Associates
Inc.
Hill, H. C. (2N4a). Content knowledge for teaching mathematics measures (CKTM measures) Technical
report on number and operations content knowledge items 2001-2@3. Ann Arbor, MI: University of
Michigan.
Hill, H. C. (2004b). Content knowledgefor teaching Mthematics measures (CKTM measures): Technical
report on geometry items-2002. Ann Arbor, MI: University of Michigan.
Hill, H. C. (2004c). Technical report on number and operations knowledge of studentsand content itents-
2001-2003. Ann Arbor, MI: University of Michigan.
Hill, H. C. (2004d). Technical report on patterns, functions and algebra items-2001. Ann Arbor, MI:
University of Michigan.
Hill, H. C., & Ball, D. L. (2004). Learning mathematics for teaching: Results from California's mathematics
professional development institutes. Journalfor Research in Mathematics Education, SS(5),330-351.
Hill, H. C., Ball, D. L., & Schilling, S. G. (2007). Unpacking "Pedagogicalcontent knowledge": Con-
ceptualizing and measuring teachers' topic-specific knowledge of students. Joumal for Research in
Mathematics Education (in review).
Hill, H. C., Schilling, S. G., & Ball, D. L. (2004). Deveioping measuresof teachers' knowledge for teaching.
The ElernentarySchool Journal,105(1), 1l-30.
Jaworski, 8., & Phillips, D. (1999). Looking abroad: Intemational comparisons and the teaching of math-
ematics in Britain. In B. Jaworski & D. Phillips (Eds.), Compaing standards internationolly: Research
and practice in mathematics and beyond (pp.7- 22). Oxford: Symposium Books.
Keitel, C., & Kilpatrick, J. (1999). The rationality and irrationality ofinternational comparativestudies.In
G. Kaiser, E. Luna, & I. Huntley (Eds.), Intemational comparisons in mathematics education (pp.24l-
256). London: Falmer Press.
LeTendre, G. K., Baker, D. P., Akiba, M., Goesling, B., & Wiseman, A. (2001). Teachers' work: Institu-
tional isomorphism and cultural variation in the U.S., Germany, and Japan. Educational Researcher,
J0(6),3-15.
Ma,"L. (1999). Knowing and teaching elementary mathenatics. Mahwah, New Jersey: Lawrence Erlbaum
Associates Inc.
Maxwell, B. (1996). Translation and cultural adaptation of the suwey instruments. In M. O. Martin & D. L.
Keliy (Eds.), Third intemational mathematics and science study (TIMSS) technical report (Yol. 1, pp.
8-1-8-10). ChestnutHill, MA: Boston College.
Mayer, R. 8., Sims, V., & Tajika, H. (1995). A comparison of how textbooks teach mathematical problem
solving in Japan and the United States.American Educational Research Journal, 32(2),443460.
Meyer, J. W., Ramirez, F. O., & Soysal, Y. N. (1992). World expansion of mass education. Sociology of
Education, 65(2), 128-149.
Mullis, I. V. S., Kelly, D. L., & Haley, K. (1996). Translationverificationprocedures.In M. O. Martin & I.
V. S. Mullis (Eds.), Third international mathematics and science study: Quality assurance in data
collection (pp. l-l-1-14). ChestnutHill, MA: Boston College.
Murphy, B. (2004)- Practice in Irish infant classrooms in the context of the Irish Primary School Curriculum
(1999): Insights from a study of curriculum implementation. Intemational Joumal of Early Years
Education, I 2(3), 245-257.
National Council for Curriculum and Assessment. (2005). Primary curriculum review, phase I: Final
report. Dublin: NCCA.
National Council of Teachers of Mathematics. (20OO).Principles and standards for school mathenatics.
Reston, VA: National Council of Teachers of Mathematics.
Santagata,R. (2004). "Are you joking or are you sleeping?" Cultural beliefs and practices in Italian and
U.S. tachers' mistake-handling strategies.Linguistics and Education, 15, 141-164.
'zl )Prmger
The Journatof Mathematics TeacherEducationis publishedby Springer.The originalpublication

is availableat www.sprinqerlink.com (http://www.sprinqerlink.com/contenVt+O066aZgi6Z3Z30D
prohibitedwithout
permission.
t97
Schmidt, W. H., Jorde, D., Cogan, L. S., Barrier, 8.,

Gonzalo, I., Moser, U., et al- (1996). Characterizing
pedagogical flow: An investigation of mathematics
and. science teaching i rir'rounrrirr.Dordrecht,
NL: Kluwer Academic publishers.
Shulman' L' s' (1986)' Thoservho understand:Knowledge
growth in teaching. Educational Researcher,4-
Silver, E. A., Leung, S. S., & cai, J. (1995). Generaring

multiple solutionsfor a problem: A comparisonof
the responsesof u.S. and Japanesestudents. Ediationat
studies in Mathematics, 2s(r),35_54.
Stein,M' K', Baxter,J. A., &.Leinirardt,G. (1990).subject-matter
knowledgeand elementaryinstrucrion:A
case from functions and graphing. American Edricationat
Research lzumat, i/t[, u.,a,o.
stigler, J. w., & Hieberr, J. (r999J. The-teaching gap:
Bestideasfrom tn, *orti,i ior:hersfor improving
education in the classroom. New york: Thi lree press.
Stigler, J. w., Lee, s.-y., & stevenson,H. w. (19g7).
Marhematicsclassroomsin Japan,Taiwan an<lthe
Unired Stares. Child Development, 5g(5), l2j2_12g5.
Walsh' T' (2006)' A critical analysi.s-ofcurricular policy
in lrbh primary schools, IggT-lggT.Unpublished
doctoral dissertation, National University of Ireland Maynooth,
_- Kildu.".
wilson, M., & Xie, y' (2004). The imperial vs. metric study'(MS).
n"t.i"u"a May 6th 2006,2006.
wilson,.S' M. (2003). Catifomia dreaming: Reforming
mathematics ed.ucation. i.qe*-Haven, CT: yale
University Press.
wilson' S' M', & Berne, J. (1999). Teacher leaming
and the acquisition of professional knowledge: An
examination of researchon contemporary profesiional
develoiment. nrrii-- oi n"iorch in Ed,ucation,
24, 173-209.
Zimowski, M, F., Muraki, E., Mislevy, R. J., & Bock,
R. D. (1996). Bitog-MG J. Chicago: Scientific
Software Intemational.
4\^
. ? sPrmger

is availableat www.sprinqerlink.com (http://www.sprinoerlink.com/content/t+o066+zBi6z373o/)
prohibited
without
permission.

"Mathematical Knowledge For Teaching": Adapting U.S. Measures For Use in Lreland

Uploaded by

Document Informationclick to expand document informationMatemática

Document Informationclick to expand document information

Copyright:

Available Formats

"Mathematical Knowledge For Teaching": Adapting U.S. Measures For Use in Lreland

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

"Mathematical Knowledge For Teaching": Adapting U.S. Measures For Use in Lreland

Uploaded by

Copyright:

Available Formats

D e l a n e yS, . , B a l l ,D . L . ,H i l l ,H . C . , S c h i l l i n gS,. G . ,& Z o p ,f D . , ( 2 0 0 8 )".

J Math TeacherEduc (2008) tl:t7t-19'l

"Mathematical knowledgefor teaching": adapting

Sedn Delaney . Deborah Loewenberg Ball . Heather C. Hill .

Published online: l9 January 2008

Abstract This article describesa study in which measuresof mathematicalknowledge

Keywords Mathematicalknowledgefor teaching . practice of teaching .

S. Delaney ' D. L. Ball . H. C. Hill . S. G. Schilling .D. Zopf

LMT Learning Mathematicsfor Teachins

The last two decadeshave seengrowing interestin cross-nationalstudies

TheJournalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

Severalstudiesover the past two decadeshave investigatedhow teachersacquire math-

Fig. I Mathematical knowledge for teaching (Hill et al. 2007)

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

numbersof teachers.Theseitems, developedby the Learning Mathematics

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

mathematicalknowledge.Indeed,Ma (1999) changedthe context of ..subtraction

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

what teachersdo during mathematicslessons.For example,the

rhe Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

Mathematics(1989) and its Principles and Standardsfor School Mathematics(2000).4

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

on many issues,including education.6These differencesand similarities make keland a

" For example see http://www.ed.gov/news/pressreleaseV2003/ll/l I l92003.html (accessedon October 7th

fhe Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

outside the United

fhe Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

Item category Numbers and Pattems, functions Geometry

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

People's names Mr. Ives Mr. Fennelly 69

Type of change Example from Example from Number of items

School language Students'papers Pupils'work

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

In contrastto the previoustwo categories,the third categoryrelatedto the mathematical

Table 4 Exemplars of mathematical changes to items and frequency of changes

Change Example from Example from adapted Number of items

Units Pennies One-cent coins t6

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

anticipatedstudentresponses.The mathematicallanguageusedin Irish schoolsdiffers in

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

o) Youwont to split rf piesevenlYbetween

c) Youore mokingsomehomemode toffy ond

Fig. 2 How an item looked before adaptation

Fig.3 How an item looked after adaptation (changes underlined)

for respondingto a questionor

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

Datafrom the sampleof 100 Irish teacherswere analyzedand comparedto 6ataobtained

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

o.4 0.5 0.6 0.7 0'8 09

Fig. 5 (a) Linear rePresentauon (a)'rltl

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

representationmay be neededif the goal were to elicit Irish teachers'understandingof

Table 5 Point biserial correla-

The Journalof MathematicsTeacherEducationis publishedby Springer.The originalpublication

-,1.g00 -3.000 '2.000 -1'000 0'000 1'000