Date post: | 21-Feb-2018 |
Category: |
Documents |
Upload: | sirojul-umam-nuruddin |
View: | 219 times |
Download: | 0 times |
of 27
7/24/2019 Qualitative Data Analysis 1
1/27
Focusing and Bounding
the Collection of DataTH E SUB STANTIVE START
Contrary to what you might have heard, qualitative research
designs do exist. Some are more deliberate than others. At the
proposal stage and in the early planning and start-up stages,
many design decisions are being made some explicitly and
precisely, some implicitly, some unknowingly, and still others
by deault. !he qualitative researcher is beginning to ocus on
the study"s issues, the cases to be studied, the data to be
collected, and how these data will be managed and analy#ed.
!his book is about analysis.$hy are we talking about
design% As &igure '.( suggests, study design decisions can, in
a real sense, be seen as analytica sort o anticipatory data
reductionbecause they constrain later analysis by ruling out
certain variables and relationships and attending to others.
)esign decisions also permit and support later analysis* they
preigure your analytic moves.
Some design decisions are mainly conceptual* the con-
ceptual ramework and research questions, sampling, case
deinition, instrumentadon, and the nature o the data to be
collected. +thers discussed in Chapter (, though they
appear in the guise o management/ issues, are equally
ocusing and bounding* how data will be stored, managed,
and processed* what computer sotware may be used
tosupport the work* and which agreements are made with the
people being studied.
$e cannot deal thoroughly here with qualitative research
design* see the detailed, helpul suggestions made by
0arshall and 1ossman '232. 4n this chapter we discuss the
analytic issues that arise as a study is bounded, ocused, and
organi#ed. $e provide speciic examples, but want to
emphasi#e that these issues must be dealt with uniquely in
any particular study. !hey may be approached loosely or
tightly* in either case, initial design decisions nearly always
lead to redesign. 5ualitative research designs are not
copyable patterns or panaceas that eliminate the need or
building, revising, and choreographing/ your analytic work
6reissle, '22'.
7/24/2019 Qualitative Data Analysis 1
2/27
Tight Versus Loose So!e Trade"offs
6rior to ieldwork, how much shape should a qualitative
research design have% Should there be a preexistent con-
ceptual ramework% A set o research questions% Some
predesigned devices or collecting data% )oes such prior
bounding o the study blind the researcher to important
eatures in the case, or cause misreading o local inorm
7/24/2019 Qualitative Data Analysis 1
3/27
&ocusing and 7ounding the Collection o )atax(
ants" perceptions% )oes lack o bounding and ocusing
lead to indiscriminate data"collection and data overload%
!hese are recurrent questions in qualitative analysis, and they
have started up lively debate. 8et"s try to order the terms o
the debate and to explain our own position.
Any researcher, no matter how unstructured or inductive,
comes to ieldwork withsomeorienting ideas. A sociologist
may ocus on amilies or organi#ations rather than, say, on
rock ormations or anthills and, within that ocus, will look
or data marked by conceptual tags roles, relationships,
routines, norms. 4 that researcher looks at closets or
lunchrooms, it is not with the eyes o an architect or a cook,
but with an interest in what the room and its contents have to
say about the patterns shared by people using it. A
psychologist would orient dierently toward the same
phenomena, 99seeing: motivation, anxiety, communication,
and cognition.
!he conventional image o ield research is one that keeps
prestructured designs to a minimum. 0any social
anthropologists and social phenomenologists consider social
processes to be too complex, too relative, too elusive, or too
exotic to be approached with explicit conceptual rames or
standard instruments. !hey preer a more loosely structured,
emergent, inductively grounded: approach to gathering data;
!he conceptual ramework should emerge rom the ield in
the course o the study* the important research questions will
come clear only gradually* meaningul settings and actors
cannot be selected prior to ieldwork* instruments, i any,
should be derived rom the properties o the setting and itsactors" views o them.
$e go along with this visionup to a point.
7/24/2019 Qualitative Data Analysis 1
4/27
generali#able indings, but they are less case-sensitive and
may entail bending data out o contextual shape to answer a
cross-case analytic question. !he solution may well lie in
avoiding the extremes.
$ith this backdrop, let"s look more closely at the aspects
o a study design involving decisions about ocusing and
bounding the collection o qualitative data in the ield. 4n this
chapter we ocus on conceptual aspects, including developing
a conceptual ramework, ormulating research questions,
deining the case, sampling, and instrumentation. $e turn to
management issues in Chapter (.
A. 7uilding a Conceptual &ramework
Rationale
!heory building relies on a ew general constructs that
subsume a mountain o particulars. Categories such as social
climate,/ cultural scene,/ and role conlict/ are the labels
we put on intellectual bins/ containing many discrete events
and behaviors. Any researcher, no matter how inductive in
approach, knows which bins are likely to be in play in the
study and what is likely to be in them. 7ins come rom theory
and experience and oten rom the general obBectives o the
study envisioned. Setting out bins, naming them, and getting
clearer about their interrelationships lead you to a conceptual
ramework.
)oing that exercise also orces you to be selectivetodecide which variables are most important, which relation-
ships are likely to be most meaningul, and, as a consequence,
what inormation should be collected and analy#edat least
at the outset. 4 multiple researchers are involved, the
ramework helps them study the same phenomenon in ways
that will permit an eventual cross-case analysis.
Brief Descri#tion
A conceptual ramework explains, either graphically or in
narrative orm, the main things to be studiedthe key
actors, constructs or variablesand the presumed rela-
tionships among them. &rameworks can be rudimentary or
elaborate, theory-driven or commonsensical, descriptive or
causal.
Illustrations
8et"s look at a ew examples. &irst, &igure =.' presents a
rudimentary, mostly descriptive ramework rom a large-
scale contract research study !he Detwork, 4nc., '2?2.
!he study"s general obBectives were to examine several
programs aimed at school improvement/ through the dis-
semination o exemplary innovations, toE understand the
reasons or implementation success, and to make policy
recommendations.
7/24/2019 Qualitative Data Analysis 1
5/27
A, 7uilding a Conceptual &ramework n (2
$e see here the ocusing and bounding unction o a
conceptual ramework. Some, not all, actors are going to bestudied, along with some, not all, aspects o their activity.
+nly some relationships will be explored, certain kinds o
outcomes measured, and certain analyses madeat least at
the outset.
Dow or a slightly more complex, more inerential con-
ceptual rame using some o the same variables &igure =.=.4t comes rom the same study. 4t is a reinement o the irst
illustration, with heavier bets being made on the in-
terrelationships. &or example, policymakers/ are hy-
pothesi#ed to inluence linkers/ through the provision o
technical assistance and through interventions in the linkers"
network.
!here are ew two-way arrows in this cut. !he researcher
isFdeciding to collect inormationselectively,at least on the
irst go, to test some hypotheses. Similarly, it looks as i the
Figure 2,2
Second Conceptual &ramework or a Study o the )issemination o @ducational 4nnovations
!he Detwork, 4nc., '2?2
6olicymaker
6olicymaker
7/24/2019 Qualitative Data Analysis 1
6/27
A, 7uilding a Conceptual &ramework n (2
study will ocus more heavily on linker behavior,: adopter
behavior,/ and implementation eectiveness/that is, on
variables coming later in the
Figure 23
Conceptual &ramework or a 0ulticase School 4mprovement/ &ield Study, 4nitial Gersion
7/24/2019 Qualitative Data Analysis 1
7/27
A, 7uilding a Conceptual &ramework n (2
0iles, '23(b, '23I.=
+nce again, we have the bins, labeled as eventse.g.,
prior history with innovations/,settingse.gN community,
district oice... adopting school/,processese.g.,
assistance,/ changes in user perceptions and practices/,
and theoretical constructse.g., organi#ational rules/.
Some o the outcomes are hypothesi#ed e.g., degree o
institutionali#ation/, but most are open-ended perceived
gains and losses/. !he directional arrows ollow time low,
but some bets still- are being made e.g., that most assistance
comes early and that reciprocal changes will occur among
the innovation, its users, and the organi#ation.
7ut the contents o each bin are less predetermined than
in &igure =.'. @ach researcher in the study will have to ind
out what the characteristics/ o the innovations are at the
ield site and how these actors will aect implementation
eectiveness./ !his is still a very general brie.
4t is also a brie that can change en route, as this concep-
tual ramework did. As qualitative researchers collect data,
they revise their rameworksmake them more precise,
replace empirically eeble bins with more meaningul ones,
and reconstrue relationships. Conceptual rameworks are
simply the current version o the researcherFs map o the
territory being investigated. As the explorerFs knowledge o
the terrain improves, the map becomes correspondingly more
dierentiated and integrated, and researchers in a multiple-
case study can coordinate their data collection even more
closely.
Variations
7/24/2019 Qualitative Data Analysis 1
8/27
A, 7uilding a Conceptual &ramework n (2
cate the relationships to be examined initially* some show
one-way inluence, others two-way. !here are purely, de-
scriptive bins curriculum, teachingOleaming styles and more
conceptual labels opportunity structure, survival/
knowledge and strategies. @nsuing, inevitable changes can
be mapped onto this rame, or they can call or a recasting.
Conceptual rameworks can also evol ve and develop out
o ieldwork itsel, 8, 0. Smith and Peith '2?' were
among the irst to do this graphically. 7ox =.= shows anexample drawn rom their study o the creation o a new
school. !he researchers noticed something they called
cloaking o organi#ational activities/keeping internaQ
unctioning protected rom external view. $hy did this
happen, and what were the consequences% Smith and Peith
believed that the school"s ormali#ed doctrine/its phi-
losophy and setup, along with procedural diiculties, sta
conlict, and poor program it with community biasesled to
the cloaking. !he cloaking, in turn, led to inaccurate public
perceptions and parental rustration* Pensington School
could not build long-term support or itsel.
Although the cloaking idea came rom prior organi#a-
tionalresearch, the way it patterned at Pensington plusFthe
other associated variables essentially were derived induc-
tively. Smith and Peith used many such emergent conceptualrameworks to explicate their understanding.
Ad'ice &
7/24/2019 Qualitative Data Analysis 1
9/27
A, 7uilding a Conceptual &ramework n (2
with all o the inormation at once. !ry your hand at it,
especially i you are a beginning researcher.
=. @xpect to do several iterations, right rom the outset. !here
are probably as many ways o representing the main
variables as there are variables to represent, but some
typically later cutsare more elegant and parsimonious than
others.
E (. 4 your study has more than one researcher, have each
ield researcher do a cut at a ramework early on and then
compare the several versions. !his procedure will show,
literally, where everyone"s head is. 4t usually leads to an
explication o contentious or oggy areas that otherwise
would have suraced later on, with ar more loss o time and
data.
I. Avoid the no-risk rameworkthat is, one that deines
variables at a very global level and has two-directional
arrows everywhere. !his avoidance amounts essentially to
making no ocusing and bounding decisions, and is little
better than the strategy o going indiscriminately into the
ield to see what the site has to tell./
7/24/2019 Qualitative Data Analysis 1
10/27
7, &ormulating 1esearch 5uestions m 23
research questions, conceptual rameworks, and sampling
matricesthough their choices are more implicit and the
links between ramework and procedures less linear. Dev-
ertheless these choices will serve to bound and ocus their
study-.
!ake, or example, the problem o understanding police
worka domain that remains somewhat arcane and obscure
despite or perhaps because p !G shows. 4t deserves an
inductive approach. 7ut which acets o police work will be
studied% ou can"t look at them all. And where will you study
them% ou can"t look everywhere. And when% 4 you delay/
that decision or a ew months until you"ve spent some time,
say, at the precinct station, that simply relects two tacit
sampling decisions start at *the precinct house, then recheck
ater a while.Suppose the implicit research question was,
7/24/2019 Qualitative Data Analysis 1
11/27
Figure $%)KeneraQ and Speciic 1esearch 5uestions 1elating to the Adoption )ecision School 4mprovement Study
7/24/2019 Qualitative Data Analysis 1
12/27
requisite conditions,/ and the idea o requisite conditions/
may not be retained throughout the study.
Ad'ice
'. @ven i you are in a highly inductive mode, it is a good idea
to Start with some general research questions.
!hey allow you to get clear about what, in the general
domain, is o most interest. !hey make the implicit explicit
without necessarily ree#ing or limiting your vision.
=. 4 you are oggy about your priorities or about the ways they
can be ramed, begin with a oggy research question and
then try to deog it. 0ost research questions do not come out
right on the irst cut, no matter how experienced the
researcher or how clear the domain o study.
(. &ormulating more than a do#en or so general questions is
looking or trouble. ou can easily lose the orest or the
trees and ragment the collection o data. shows this graphically; !here is a ocus, or
heart,/ o the study, and a somewhat indeterminate bound-
ary deines the edge o the case; what will not be studied.
Figure $%*!he Case as the Jnit o Analysis
DF 7+JD)A1 UU setting,
U concepts,4 sampling, etc.
tt
O$
O&+CJS
7/24/2019 Qualitative Data Analysis 1
13/27
you will know it. !he beauty o qualitative ield research is that there
is nearly always a second chance.Illustrations
$hat are some examples o cases% Sometimes the phe-
nomenon/ may be an individualin a deined context;
A patient undergoing cardiovascular bypass surgery, beore,
during, and T months ater surgery, in the context o his or
her amily and the hospital setting !aylor, 0ac8ean,
6allister,H $hite, '233
Dote that the heart/ here is the patient. !he boundary
deines amily and hospital as the context. !he researchers
will not, or example, interview the patient"s colleagues at
work or visit the restaurants where he or she dines. !he
bounding is also by time; Do inormation will be gathered
later than T months ater hospitali#ation.
$e can also expect that the boundary will be deined
urther bysamplingoperations, to which we"ll come in a
minute. &or example, these researchers will not be inter-
viewing the patient"s children, only the spouse. 7ut they will
be sampling diet, exercise, and blood count data, as well as
interview data rom the patient"s lived experience./
+ther examples o individualsas the case are;
A young person"s work experience in his or her irst real/ Bob
7orman, '22'; or example, 0iriam"s work as a
bookkeeper in 1iver City 7ank
An uncommonly talented mechanic, in his shop amid the context
o his riends, neighbors, and customers study o what happened when a ather gave a
treasured chess set to his young son. A case also may be
deined as an event,such as a school sta meeting* or as a
period of time,as in the classic study)idwest and *ts
+hildren7arker H $right, '2?', which includes +ne
7oy"s )ay,: a record o what 1aymond did rom the time he
got out o bed until he reentered it* or as a sustainedprocess
the adoption, implementation, and institutionali#ation o an
innovative program by a school in its district, as in the study
we already described
7/24/2019 Qualitative Data Analysis 1
14/27
A comment on notation. $e sometimes preerand use
here and there in this boohthe wordsitebecause it reminds
us that a case/ always occurs in a speciied social and
physicalsetting-we cannot study individual cases devoid o
their context in the way that a quantitative researcher oten
does.
Ad'ice
'. Start intuitively. !hink o the ocus, or :heart/ and build
outward. !hink o what you will.no* be studying as a way to
irm up the boundary. Admit that the boundary is never quite
as solid as a rationalist might hope.
=. )eine the case as early as you can during a study. Kiven a
starting conceptual ramework and research questions, it paysto get a bit stern about what you are deining as a case* that
will help clariy urther both the ramework and the questions.
(. 1emember that sampling operations will deine the cases
urther.
I. Attend to several dimensions o the case; its conceptual
nature, itssocial si&e,itsphysicallocation, and its temporal
extent.
Ti!e Re(uired
4 a starting conceptual ramework and research questions
are reasonably clear, a irst cut at case deinition usually takes
no more than a ew minutes* discussion among members o a
research team or with interested colleagues may occupy an
hour or two as clarity emerges during successive iterations othe deinition o the case./
). Sampling* 7ounding the Collection o )ata
Rationale
Sampling is crucial or later analysis. As much as you
might want to, you cannot study everyone everywhere doing
everything. our choiceswhom to look at or talk with,
where, when, about what, and whyall place limits on the
conclusions you can draw, and on how conident you and
others eel about them.
Sampling may look easy. 0uch qualitative research ex-
amines a single case,/ some phenomenon embedded in a
single social setting. 7ut settings have subsettings schoolshave classrooms, classrooms have cliques, cliques have
individuals, so deciding where to look is not easy. $ithin any
case, social phenomena prolierate science lessons, teacher
questioning techniques, student unruliness, use o
innovations* they, too, must be sampled. And the questions o
multiple-case sampling add another layer o complexity.
7/24/2019 Qualitative Data Analysis 1
15/27
Figure $%+
!ypology o Sampling Strategies in 5ualitative 4nquiry
Pu#ei, '22=* 6atton, '22
0urpose
)ocuments diverse variations and identiies important common patterns
&ocuses, reduces, simpliies, acilitates group interviewing
6ermits logical generali#ation and maximum application o inormation to other
cases &inding examples o a theoretical construct and thereby elaborate and
examine it @laborating initial analysis, seeking exceptions, looking or variation
4dentiies cases o interest rom people who know people who know what cases are
inormation-rich 8earning rom highly unusual maniestations o the phenomenon o interest
7/24/2019 Qualitative Data Analysis 1
16/27
teristics over time- Xa replication strategy. 0ost o these
strategies will increase conidence in analytic indings on the
grounds o representativeness.
See also Kuba and 8incoln '232, who advocate maxi/
mum variationsampling, a deliberate hunt or negative
instances or variations. !his process may take the orm o
questions to inormants, such as, $hom do you know who
sees things dierently%/ or $here can 4 ind patients who
dontkeep appointments%/
Wohnson '22, in his comprehensive treatment o se-
lecting inormants, also suggests dimensionalsampling; !he
researcher lays out the dimensions on which variability is
sought, then takes representative, well-inormed/ inormants
or each contrasting dimension. !he aim is to ind people who
are more knowledgeable, reliable, and accurate in reportingevents that are usual, requent, or patterned. !his strategy has
risks; Such inormants may assume greater uniormity than
actually exists 6elto H 6elto, '2?>* 6oggie, '2?=.
!he sampling strategies we"ve been discussing can be
applied both within or across cases. 8et"s turn to some o the
core issues in each o these domains.
Within/case sampling.5uantitative researchers usually think
o cases as individual persons, draw a sample/ o persons,
and then collect comparable data points/ rom each. 7y
contrast, a qualitative case/ may range widely in deinition
rom individuals to roles, groups, organi#ations, programs,
and cultures. 7ut even when the case is an individual, the
qualitative researcher has many within- case sampling
decisions; $hich activities, processes, events, times,
locations,and role partnerswill 4 sample%4n our cardiovascular bypass patient example, we might
want to sample diet and exercise activities* the processes o
understanding, taking in, and acting on medical advice*
events such as admission and discharge interviews* time
periods including prehospitali#ation, hospitali#ation, and
posthospitali#ation once every = weeks* locations including
recovery room, ward, and the patient"s home* and role
partners including the patient"s physician, ward nurses,
dietitian, and spouse.
$ithin-case sampling is almost always nestedor ex-
ample, studying children within classrooms within schools
within neighborhoods, with regular movement up and down
that ladder.
A second maBor point is that such sampling must be
theoreticallydrivenwhether the theory is prespeciied oremerges as you go, as in Klaser and Strauss"s '2T? theo-
retical sampling./ Choices o inormants, episodes, and
interactions are being driven by a conceptual question, not by
a concern or representativeness./ !o get to the construct, we
need to see dierent instances o it, at dierent moments, in
dierent places, with dierent people. !he prime concern is
with the conditionsunder which the construct or theory
operates, not with the generali#ation o the indings to other
settings.
!he third point is that within-case sampling has an it/
erativeor rolling/ quality, working in progressive waves/
as the study progresses. Sampling is investigative* we are
cerebral detectives, erreting out answers to our research
questions. $e observe, talk to people, and pick up artiacts
and documents. !hat leads us to new samples o inormants
and observations, new documents. At each step along the
evidential trail, we are making sampling decisions to clariy
the main patterns, see contrasts, identiy exceptions or
discrepant instances, and uncover negative instanceswhere
the pattern does not hold. +ur analytic conclusions depend
deeply on the within-case sampling choices we made.
So within-case sampling helps us see a local conigurationin some depth. $hat can adding cases do or us, and how do
we create a sample o cases%
)ultiple/case sampling.0ultiple-case sampling adds
confidenceto indings. 7y looking at a range o similar and
contrasting cases, we can understand a single-case inding,
grounding it by speciying howand whereand, i possible,
whyit carries on as it does. $ecan strengthen the precision,
the validity, and the stability o the indings. $e are ollowing
a replicationstrategy in, '22'. 4 a inding holds in one
setting and, given its proile, also holds in a comparable
setting but does not in a contrasting case, the inding is more
robust. !he multiple comparison groups/ used in grounded
theory work play a similar role.
$ith multiple-case studies, does the issue ogener/
ali&abilitychange% @ssentially, no. $e are generali#ing rom
one case to the next on the basis o a match to the underlying
theory, not to a larger universe. !he choice o cases usually is
made on conceptualgrounds, not on representative grounds.
!he cases oten are arrayed on a continuum e.g., highly
gited to underachieving pupils, with ew exemplars o each,
or they are contrasted e.g., assertive and passive
adolescents. +ther, unique properties may be added e.g.,
some assertive adolescents are rom cities, some rom rural
areas. 4 you look closely at the cells o such a sampling
rame, each is essentially unique. 7ecause case study
researchers examine intact settings in such loving detail, they
know all too well that each setting has a ew properties it
shares with manyothers, some properties it shares withsome
others, and some properties it shares with noothers.
Devertheless, the multiple-case sampling gives us conidencethat our emerging theory is generic, because we have seen it
work outand not work outin predictable ways.
0ultiple-case sampling, although it may have iterative
aspects, normally has to be thought through careully. An
explicitsampling frameis needed. 4t will be guided by the
research questions and conceptual rameworkeither pre-
speciied or emergent. +nce again, random sampling will not
help.
7/24/2019 Qualitative Data Analysis 1
17/27
). Sampling; 7ounding the Collection o )ata n '>
$e have to deal with the issue conceptually;
7/24/2019 Qualitative Data Analysis 1
18/27
events, and an abundance o logistical problems. Gery seldom
does a start-up sampling rame survive the lovely
imperection and intractability o the ield. 4t must be shited
and reramed.
7eing selective calls or some restraint in the classes, o
data you go ater.
7/24/2019 Qualitative Data Analysis 1
19/27
to
i
Y @ Z externally developed innovation x ' program names are pseudonyms, to avoid identiying speciic sites.
able 2.4
Characteristics o &ield Study SampleS4!@ C+D!@L! AS6 @ C ! S +& !
7/24/2019 Qualitative Data Analysis 1
20/27
to
i
8 as locally-developed innovation E 6rogram is used in this site with a comprehensive sample o learners, rather than with low-
achieving or marginal populations.!hble =.=
&inal Sampling &rame or &ield Study
are conceptually meaningul. !his eect is much more powerul than a series oindividual case studies over several years.
So we have done pretty well with a '=-case sample; careul illustration,
strong construct validity, legitimate claims to external validity with respect to the
core indings, and spadework or replication studies with well-identiied subsets
o settings and actorsall without control groups, random sampling, systematic
treatments, or other procedures required in experimental and correlationalresearch.
Dote that this sampling matrix does not resolve some o the importantwithin/
caseocusing and bounding decisions to be made in multiple-case studies. &or
example, i one case researcher observes only administrators and another only
teachers, the comparability o the two cases is minimal. &or these decisions, we
DA!4+DA8)4&&JS4+DD@!$+1P61+W@C!S !4!8@!G-C61+W@C!S(NDN) (IV-C)
'2?2 '2?3 @arlier '2?3 '2?? @arlier
@L6AD)4DK
SC+1@-+D7anestown1ural6ull-out, in-
@6S&Astoria!ownOsubur
b Add-on,inschool
46ACarson1uralAdd-on, in-
school
+DK+4DK
468@7urtonSuburb)rop-in, in-.
@C140asepa1ural)rop-in, in-
scaool
@7C@6erry-
6arkdaleSuburb
Subsystem,
!indale1eading!indale JrbansprawlSubsystem, in-
0attesonI-)Calston0etrourban
7entley Center6lummetJrbanSubsystem, in-
school
4DAC!4G@,)$4D)84DK
1A1@8ido1uralAdd-on, ield
@skimoCuir, )un
7/24/2019 Qualitative Data Analysis 1
21/27
to
i
usually need the conceptual ramework or the research questions. Jsing them,
we can agree on sampling parameters and comparable choices or initial
ieldworksettings, actors, events, processes.
Cross-case comparison is impossible i researchers operate in radically
dierent settings, use no coherent sampling rame, or, worst o all, i they ocus
on dierent processes. !hese processes in cases low out o the interplay o
actors, events, and settings. !hey are usually at the core o the conceptual
ramework, and serve as the glue holding the research questions together. Pey
processes can be identiied at the outset or graduallyoten via pattern codes,
relective remarks, memos, and interim summaries, as we"ll see in Chapter I,
7eing explicit about processes and collecting comparable data on them will not
only avoid costly distractions but also oster c omparability and give you easier
access to the core underlying constructs as you get deeper into data collection.
Ad'ice
'. 4 you"re new to qualitative research, rest assured that there is never enough:
time to do any study. So taking the tack, 4"ll start somewhere and take it rom
there,/ is asking or trouble. 4t is probably agood idea to start with a allback
sample o inormants and subsettings; the things you have to cover in light o
what you know at that point. !hat sam- pie will change later, but less than you
may think.
=. Wust thinking in sampling-rame terms is good or your study"s health. 4 you are
talking with one kind o inormant, you need to consider why this kind o
inormant is important and, rom there, who else should be interviewed or
observed. !his is also a good exercise or controlling bias.
(. 4n complex cases, remember that you are sampling people to gist atcharacteristics o settings, events, and processes. Conceptually, the people
themselves are secondary. !his means watching out or an overreliance on talk,
or on observation o inormants, while you may neglect sampling or key events,
interactions in dierent settings, and episodes embodying die emerging patterns
in the study. 1emember also to line up these sampling parameters with the
research questions as you go; Are my choices doing a representative, time-
eicient Bob o answering them% &inally, the sampling choices at the start o the
study may not be the most pertinent or data-rich ones, A systematic review can
sharpen early and late choices,
4n qualitative research, as well as in survey research, there is a danger o
sampling too narrowly. 4n act, points = and ( above tug in that direction; Ko to
the meatiest, most study-relevant sources. 7ut it is also important to work a bit at
the peripheriesto talk with people who are not .central to the phenomenon but
are neighbors to it, to people no longer actively involved, to dissidents and
renegades and eccentrics. Spending a day in the adBoining village, school,
neighborhood, or clinics is also worth the time, even i you don"t see the sense at
that point.
!here are rewards or peripheral sampling. &irst, you may learn a lot.
Second, you will obtain contrasting and comparative inormation that may help
you understand the phenomenon at hand by de-centering/ you rom a particular
way o viewing your other cases. As we all know, traveling abroad gives us
insights into our own culture,
Spend some time on whether your sampling rame is easible. 7e sure the time is
there, the resources are there, the requisite access to people and places is ensured,
and
F the conditions areright or doing a careul Bob. 6lan to study a bit less, rather
than more, and bank/ the extra time. 4 you are done, the time is yours or a
wider or deeper pass at the ield. 4 not, you will need this time t o complete your
more modest inquiry under good conditions.
!hree kinds o instances have great payo. !he irst is the apparently typical/
or representative/ instance. 4 you can ind it, try to ind another one. !he
second is the negative/ or disconirming/ instance* it gives you both the limitso your conclusions and the point o greatest variation. !he third is the
exceptional/ or discrepant/ instance. !his instance will allow you to qualiy
your indings and to speciy the variations or contingencies in the main patterns
observed. Koing deliberately ater negative and atypical instances is also healthy
in itsel* it may orce you to clariy your concepts, and it may tell you that you
7/24/2019 Qualitative Data Analysis 1
22/27
to
i
indeed have sampled too narrowly. 0ore on this in Chapter ', section 7.
?. Apply some criteria to your irst and later sampling plans.
7/24/2019 Qualitative Data Analysis 1
23/27
to
i
structured and ocused, So let"s not delude ourselves about total control/ and
precision/ in our instrumentationwhile remembering that attention to design
can make a real dierence in data quality and the analyses you can carry out.I
Asingle/casestudy calls or less ront-end preparation than dbes amultiple/casestudy. !he latter is looking orward to cross-case comparison, which requires
some standardi#ation o instruments so that indings can be laid side by side in
the course o analysis. Similarly, afreestandingstudy has ewer constraints than a
multimethodFigure $%, 6rior 4nstrumentation; Pey )ecision &actors
7/24/2019 Qualitative Data Analysis 1
24/27
to
i
study e.g., a ield study tied to a survey, an idea we discuss urther in Chapter
(. A 8basic(study oten needs less advance organi#ing than anapplied,
evaluation,orpolicy study. 4n the latter cases, the ocus is tighter and the instru-
mentation more closely keyed to the variables o interest.
(. 0uch depends on thecase definitionand levels o analysis expected. A
researcher studying classroom climate in an elementary school might choose to
look intensively in ( o the building"s (> classrooms, and so probably would be
right to start with a looser, orienting set o instruments. 4, however, an attempt is
made to say something about how classroom climate issues are embedded in the
working culture o the building as a whole, a more standardi#ed, validated
instrumenta questionnaire or a group interview schedulemay also be
required.
4n &igure =.? is a summary o some main issues in deciding on the
appropriate amount o ront-end instrumentation.
$e think there is wisdom in all three stances toward ront-end
instrumentation and its degree o structure. !he irst stance little prior
instrumentation puts the emphasis on certain types o validity;constructAre
the concepts well grounded%,descriptive$contextuai4s the account complete
and thorough%, interpretive)oes the account connect with the lived
experience/ o people in the case%, andnatural4s the setting mostly
undisturbed by my presence%.T
!he second stance a lot o preinstrumentation emphasi#esinternal validity
Am 4 getting a comparably measured response rom dierent people% and
generali&ability 4s this case a good instance o many others%, along with sheer
manageabilityo data collection.
!he third stance is both contingent and ecumenical, considering it unhelpul
to reach or absolute answers in relative instances. &igure out irst what kind o
study you are doing and what kind o instruments you are likely to need at
dierent moments within that study, and then go to work on the ones needed atthe outset. 7ut in all cases, as we have argued, the amount and type o
instrumentation should be a unction o your conceptual ocus, research
questions, and sampling criteria. 4 not, the tail is likely to be wagging the dog,
and later analysis will suer.
6ittle 0rior*nstrumentation 8*t 9epends '
7 6ot of 0rior
*nstrumentation
1ich context description needed
Context less crucial
Concepts inductively grounded in
local meanings
Concepts deined
ahead by researcher
@xploratory, inductive Conirmatory,theory-driven
)escriptive intent @xplanatory intent
7asic: research emphasis
Applied, evaluation
or policy emphasisSingle case 0ultiple cases
Comparability not too important Comparabilityimportant
Simple, manageable, single-level
case
Complex,
multilevel,
Kenerali#ing not a concern
Kenerali#abi li
tyOrepre-
Deed to avoid researcher impact
1esearcher impact
o less concern
5ualitative only, ree-standing study
0ultimethod study,
quantitative
7/24/2019 Qualitative Data Analysis 1
25/27
to
i
Brief Descri#tion
4nstrumentation comprises speciic methods or collecting data; !hey may be
ocused on qualitative or quantitatively organi#ed inormation, and maybe
loosely to tightly structured.
Illustration
7/24/2019 Qualitative Data Analysis 1
26/27
Figure $%-
@xcerpts &rom 4nterview Kuide,
School 4mprovement Study
((. 6robably you have a certain idea o howMMMMMMMMMMMMMMlooks to you
now, but keep thinking back to how it irst looked to you then, Bust
beore students came. . $hat parts or aspects seemed not worked out, not ready or use%
(T. Could you describe what you actually did during that week or so
beore you started usingMMMMMMMMMMwith students%
0robes%
-1eading
6reparing materials 6lanning
!alking with whom, about what
!raining
I. )id you make any changes in the standard ormat or the program
beore you started using it with students% $hat kind o changes with
things you thought might not work, things you didn"t like, things you
couldn"t do in this school%
0robes%
!hings dropped !hings added, created !hings revised
o the context, the actors, and how the school improvement
process seemed to be working locally. &rom that knowledge,we went or deeper and broader understanding.
Dow to dip into the guide near the point where the re-
search question will be explored &igure =.3.
!he interviewer begins by taking the inormant back to
the time Bust beore he or she was to use the innovation with
students, asking or detailed contextwhat was happening,
who colleagues were, and what eelings they had.
5uestions (( through (T move orward through time,
asking how the innovation looked, its ready or unready parts,
and what the teacher was doing to prepare or its use.
5uestion I comes directly to the research question, as-
sessing pre-use changes made in the innovation. !he probes
can be handled in various ways; as aids to help the
interviewer lesh out the question, as prompts or items the
inormant may have overlooked, or as subquestions derived
rom previous research.8ater in the interview, the same question recurs as the
interviewer evokes the teacher"s retrospective views o early
and later use, then moves into the present :$hat changes are
you making now%/ and the uture :$hat changes are you
considering%/.
So ail ield researchers are addressing the same general
question and are addressing it in similar ways chronologi-
cally, as a process o progressive revision, although the
question wording and sequencing will vary rom one re-
searcher to the next. 4 the response opens other doors, the
interviewer will probably go through them, coming back later
to the :transormation/ question. 4 the response is uncertain
or looks equivocal when the researcher reviews the ield
notes, the question will have to be asked again perhaps
dierentlyduring the next site visit.
pineal research and cognitive and social-psychological theory
both led us to the idea that people will adapt or reinvent
practices while using them.
!iresamplingdecisions are straightorward. !he question
addresses teachers in particular, and to get the answer, we
will have to observe or interview them or, ideally, do both.
$e should sample events such as the teacher"s irst encounter
with the innovation, and processes such as assessing its
strong and weak points and making changes in it to it one"s
practice.
8et"s look at the interview component o the instrumen/
tation.$e developed a semistructured interview guide. @ach
ield researcher was closely amiliar with the guide, but had
latitude to use a personally congenial way o asking and
sequencing the questions, and to segment them appropriately
or dierent respondents.
!he guide was designed ater ieldwork had begun. An
initial wave o site visits had been conducted to get a senseAdvice
'. $e have concentrated here on general principles o designing
appropriate instrumentation, not on detailed technical help.
&or the latter we recommend treatments such as Spradley
'2?2, $eller and 1omney '233, and 0ishler '23T on
interviewing. Wudd, Smith, and Pidder '22', 1. 7. Smith
and 0anning '23=, $erner and Schoeple '23?a, '23?b,
Koet# and 8eCompte '23I, and 7randt '23' are helpul
on other methods as well, including questionnaires,observation, and document analysis. Kood overviews are
provided by 0arshall and 1ossman '232, &etterman
'232, 7ogdan and !aylor '2?>, and i$in '2?3.
=. Simply thin:ingin instrument design terms rom the outset
strengthens data collection as you go. 4 you regularly ask,
Kiven that research question, how can 4 get an answer% it will
sharpensamplingdecisions 4 have to observeOinterview this
class o people, these events, those processes, help clariy
concepts,and help setprioritiesor actual data collection.
ou also will learn the skills o redesigning instrumentation
as new questions, new subsamples, and new lines o inquiry
develop.
Dot thinking in instrument design terms can, in act,
lead to sel-delusion; ou eel sensitive/ to the site but
actually may be stuck in reactive, seat-o-the-pants inter-
viewing. !hat tactic usually yields laccid data.
6eople and settings in ield studies can be observed more than
once. Dot everything is riding on the single interview or
observation. 4n qualitative research there is nearly always a
second chance. So ront-end instrumentation can be revised
in act, should be revised. ou learn how to ask a question in
the siteFs terms and to look with new eyes at something that
began to emerge during the irst visit. 4nstrumentation can be
modiied steadily to explore new leads, address a revised
research question, or interview a new class o inormant,
4n qualitative research, issues o instrument validity and
reliability ride largely on the skills o the researcher.
@ssentially apersonmore or less alliblyis observing,
interviewing, and recording, while modiying the observa-
tion, interviewing, and recording devices rom one ield trip
to the next. !hus you need to ask, about yoursel and your
colleagues,
7/24/2019 Qualitative Data Analysis 1
27/27
@. 4nstrumentation \ I
disciplinary grounding are considered assets. 7ut although
unamiliarity with the phenomenon or setting allows or a
ertile decentering,/ it also can lead to relatively naive,
easily misled, easily distracted ieldwork, along with the
collection o ar too much data.
!he problem is how to get beyond the supericial or the
merely salient, becoming empirically literate./ ou can
understand little more than your own evolving mental map
allows. A naive, undierentiated map will translate into
global, supericial data and interpretations-and usually
intoFsel-induced or inormant-induced bias as well. ou have
to be knowledgeable to collect good inormation 0arkus,
'2??. As Kiorgi '23T puts it, educated looking/ is
necessary. *.-.
4nexperience and single-discipline ocus can lead to asecond danger; plastering a ready-made explanation on
phenomena that could be construed in more interesting ways.
!hus presumably grounded/ theori#ing can turn out to be
conceptual heavy-handedness, without the researcher"s even
being aware o it. Kinsberg, '22, even suggests that the
researcher"s counter-transerence/ like the
psychoanalystFs unacknowledged eelings toward the patient
is at work during data collection and must be suraced and
tamed/ through discussions with peers, careul retrospective
analysis, and audits./ !ranscending personal biases and
limitations is not easy.
+n balance, we believe that a knowledgeable practitioner
with conceptual interests and more than one disciplinary
perspective is oten a better research instrument/ in a
qualitative study; more reined, more bias resistant, moreeconomical, quicker to home in on the core processes that
hold the case together, and more ecumenical in the search or
conceptual meaning.?
Ti!e Re(uired
4t is not possible to speciy time requirements or instru-
mentation. So much depends on the modes o data collection
involved, on how prestructured you choose to be, on the
nature o the research questions, and on the complexity o the
sample. 0ore structured instrumentation, done with diverse
samples, with a conirmatory emphasis, will take
substantially more time to develop.
Su!!ar. Co!!ents
$e"ve looked at substantive moves that serve to ocus
and bound the collection o datareducing it in advance, in
eect. !hese moves include systematic conceptual
framewor:sorgani#ing variables and their relationships,
research questionsthat urther deine the obBects o inquiry,
deining the heart/ and boundaries o a study through case
definition,planning or within-case and multiple-case
sampling,and creating instrumentation. A< o these moves
serve both to constrain and support analysis. All can be done
inductively and developmentally* all can be done in advance
o data collection. )esigns may be tight or loose. Such
choices depend on not only your preerred research style but
also the study"s topic and goals, available theory, and the
researcher"s amiliarity with the settings being studied.
4n the next chapter, we pursue other, more technical issueso ocusing and bounding a study during its design.
Dotes
'. !here is a paradox here 1. @.