A Scientific Perspective
1 9 on Structured, Task-Based
Interviews in Mathematics
Education Research
Gerald A. Goldin
Rutgers, the State University of New Jersey
‘This chapter consider ove methodological aspect of uallivereseurch—ihe we of structed,
‘askbusd inurviews ln observing nd lterpeing mathematical behavior. Sever siete
‘sues and their implictions ae dicused briefly, including: (a) the reproductliy,
‘compare, an laity of observation: () the peeriabty of research Badin: 6)
the importace of mathemati conte an structures; (@) the role of cogatve theory In
esigaing and interpreting interviews; aod, (c) the interplay among task and contextual
vasibles In evaluating tk-bse interview methods sclenifaly, I argue aglast some
lis tat have been adarced inthe name of epistemological schools of thought «aglag
from radical positivism and bebavioram to aa contrvcviem, socal constucivisn, sd
‘posmodemism. Finally, some broadly aplcble principles sad tchiques ae proposed for
Improving the quay of t-bused interview research
‘The penpective ole here is tha ofa physical scien as well as « mathematics
‘ducaor who as bee involved in empirical an theoretical seach oa mathematical problem
solving for 25 yea. maaan that sound plciples of ecient lavestigaton, a developed
tc applied in moder seins, shouldbe appli to this eadeaver oo. This shald ever te
one dogmatically oF automatically. However, it should be done rigorously, paying cal
senton 10 the reasoning behind the application of the meibndologial Wess of scence.
‘Although Ihave leaned mach rom the reerch of thers wh bold iffret views, 1 rein
tity wncoovince by be armen that ee anced ocasonally, claiming that scenic
methods of inquiry ae logue for, oc eleva to, the study of buman peychowocia
setives such as waching and leaming muzbemusics snd mathematical problem solving.
si518 GOLDIN
Becawe one purpose ofthis book is to conser quality sundads for quliatve mca
‘methods, ts eseatial to conse the fundamental scientific ime,
My expsisooes with t-te interview methodology eign and evolved through a
series of sues of invdoal muhematcal problem solving by elemeatry school, high
Schol, and college sudots ao aul, coo i collborto with my stadents ater
& Goldin, 1991s, 1991b; DeBels& Goldin, 191; Goldin, 198; Goa & Lands, 1985,
1986; Goin & Luger, 1975; Goin & Wate, 1982; Lager, 180; Wate, 1980). Most
‘cently, members ofa group of nvetigato tht I led at Rutgers Univesy have been
soalyzing and interpreting the resus ofa sees of five wi-bted interviews in eementary
‘school mathematics, We cad these interviews as par of longtudaal sty of inva
stildea’s matemstical development (Golda, 1993, Golda, DeBelis, DeWind-King,
Pamanino, & Zang, 1992). Between 1992 nd 1994, smocd interviews were coded
Witham Inia groop of 2 thi nt fourth-grade chien, 19 of whom complet! the full
‘series, Partial ress have been paned (DeBeli, 196; DeBells & Goldin, 1953, 1997;
Goldin Pustantino, 1996 Zang, 1994, 1955). The development interview srg fo this
‘series was guided by the views denribed in this hap, nt, my views wer infeed by
Insights goed during the sy.
1s not my itenton to dese the specifics of these sides here, bt to focus on
‘methodological suggestions and conclusions dave in part fom them The caper is
‘rpaize follows. The next section summarizes the meaning, inportace, and itons
of ask-base interview esearch in matematc edactoa. Her It to explain th notion of
Structured interviews that ar designed to lave bypotees sng qaliatve analyses of
au, and offer some ref examples. The eas presented cary forwaraod expend consderly
‘on eae dlberions about the mearuemeat of mathematical poten solving outcomes
(Cobb, 1986; Goldin, 1982, 1986; L. Han, 1986; Loca, etal, 1980) and the relation
‘between cognitive theory and asesiment (Goldin 19922). This followed by a dscesson of
‘key siemific sues in connection withthe methodology, and theca fo explicit ejection of
eran damaging conclusions deve trom amisive epistemoloyicel belief systems. The
Sol secon offers a prelimiony set of broad, siding principles and wcbnigucs for
‘nabishing and enbancing the quality of tdk-bsed Ioteview reearch in the domain of
‘mathematics.18, STRUCTUKD, TASKED-BASED INTERVIEWS $19
TASK-BASED INTERVIEWS
Structured tscbuod iervlews forthe suy of matheadcal bebuvioe involve minimally &
subject the probem solver) and an inmviewer (he clinician, iterating in relation to one or
‘more tks (quetons, problem, or activites) loosed tothe subject by the clinician in a
‘planed way. The later component jifes the term tas-ated, 20 tht the subjects!
Imerecons we not merely with te Interviewers, but with the tsk environments, Goup
‘mtcvews with wo or more subjects fil also within the purview ofthis discussion, lang
‘othe need to expand our intereons of some ofthe des,
‘Nonmally, provision is made for observing and recoring for later anlyis what takes
lice ding the interview, though mnio- andor videotuping, observers’ notes, sa the
subject's work Explicit provision is made too for contingencies that may occur at the
loterview procs, pomibly by mean of branching sequences of besrnte questions, hints,
‘elated problems ia sequence, rerospective quests or ater icterventons bythe scan. 1
{is this explicit provision for contigeoces, gee with the anton to the eqeene at
Structures ofthe tusks, tay distngusbes the "srocurs” iterviews disused bee fom
“weructund” intviews, which may be limited to “tne” problem solving (where no
‘subsoil assistance that woul alate & solution i given by the elinican to the subject)
(Ft the banding of contingences on an improvisational basi. By analyzing verbal mat
‘oavertl behavior or ineracton, the searcher bopes to make inferences sbout the
‘matbemaical thinking, laming, aor problem solving of the subject. From these
lnferenes, we hope to deepen our undesandiag of vou expects of mathematics education.
‘Wemay sim to test one or more explicit hypotheses, using qualiaive analyses of the dain,
‘eg may eck mec to cba deseipive epri atOu he subject earog er problem
‘solving: or we may hope to achive an itemnediae gl, sich as reiting o elaborating a
conjectie
‘Ofcourse, he design of srucured k-based interviews needs to tke into acount te
‘search purposes, These may inne (or example) exploratory investigation; refinement of
observation, description, ffereace, or analysis wchlgues: development of constructs al
conjectures: investigation or testing of advance hypotheses; anor ingiy fata be
spplicabiity of «model of teaching eaming, or problem soving In aio the design is
fected by the comple of he phenomena inthe syste being iavesigaed