What's Wrong With Ethnography

116 Pages·1992·32.393 MB·English

by  Martyn Hammersley

WHATS WRONG WITHH
ETHNOGRAPHY?
Methodological explorations
Martyn Hammersle
R
London and New York
CONTENTS
First published in 1992
by Rouledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Simultaneously published in the USA and Canada
by Rouledge List of figures
a division of Routledge,Taylor & Francis Acknowledgemenis X
270 Madison Ave, New York NY 10016 Introduction
Transferred to Digital Printing 2006
Part I Ehnography, theory and reality
O 1992 Martyn Hammersley
WHATS WRONC WITH ETHNOGRAPHY? THHE
Typeset in Baskerville by LascrScTipt Limited, Mitcham, Surrey
MYTH OF THEORETICAL DESCRIPTION 11
2 SOME QUESTIONS ABOUT THEORY IN
All rights reserved. No part of this book may be reprinted or
reproduced or utilized in any form or by any electronic, mechanical, or ETHNOGRAPHYA ND HISTORY 32
other means, now known or hereafter invented, including 3 ETHNOGRAPHYAND REALISM 43
photocopying and recording, or in any information storage or reuieval 4 BY WILAT CRITERIA SHOULD ETHNOGRAPHIC
system, without permission in wriing from the publishers. RESEARCH BEJ UDGED? 57
British Library Calaloguingin PublicatonDala
Hammersley, Martyn Part II Ehnography, relevance and practice
What's wrong with ethnography?
5 THE CENERALISABILITYO F ETHNOGRAPHY 85
1. Ethnography
6 CRITICAL THEORYAS A MODEL FOR
I. Tide
306.072 ETHNOGRAPIIY 96
7 PARTS THAT EVEN ETHNOGRAPHY CANNOT
Library of Congress Calaloging in Aublication Data
REACH: SOME REFLECTIONS ON THE
Hammersley, Martyn.
What's wrong with ethnography?/Martyn Hammersley. RELATIONSHIP BETWEEN RESEARCH AND POLICY 123
P cm 8 ON PRACTITIONER ETHNOGRAPHY 135
Includes bibliographical referencesand index.
1. Ethnology-Methodology. 2. Ethnology-Rescarch. Part II Qualitative versus quantitative method
3. Ethnology-Philosophy. I. Tide.
9 DECONSTRUCTING THE
GN345.H36 1992
305.8-dc20 91-10212 QUALITATIVE-QUANTITATIVE DIVIDE 159
10 THE OGIC OF THEORYTESTING IN CASE STUDY
CIP
RESEARCH 174
ISBN 0-415-05476-1
0-415-05477-X (pbk) 11 SO, W HAT ARE CASE STUDIES? 183
cONTENTs
Postscript 201
Bibliography 204
223
Name index
Subject index 228 FIGURES
11.1 Case study as one case selection strategy 184
11.2 Relationship between survey and case study 186
11.3 Relationship between experiment and case study 193
vii
ACKNOWILEDGEMENTS
Thanks to Barry Cooper, Peter Foster. Roger Comm, Colin Lacey
Donald Mackinnon, John Scarth, Thomas Schwandt and John K
Smith for comments on or discusons about various of the essays
included in this book. Thanks also to Joan, Rachel and Paul, ast
for being there.
Chapter 1 was previousiy pubiished in sigh tly different
version in Soccology, 24, 3, pp. 597-617.
Chapter 2 is a substantially modified veruon of Making Vice
of our Virtues', which appeared in L.F. Goodson and S.j. Ball (eda)
Defining the Cumcuhum, Lewes, Falmer, 1984
The author is much obliged to the Britiah Socological Aasoc
Don and to Falmer Press respectively for permissicon to reprint
hese articles.
INTRODUCTION
The tile of this collecion of essays is ambiguous. Read in one way
it might be taken to imply a defiant defence of ethnography
(what's wTong with it?). Alternatively, the tide may be interpreted
as promising the discussion of some serious problems facing this
approach to social research.' This ambiguity is intentional. It
captures what I take to be ethnography's ambivalent status at the
present time. On the one hand, it has come to be widely accepted
as a legitimate approach to social research. The proportion of
social researchers who use qualitative method has increased con
siderably in recent years. Criücisms that it is unscienific, or at best
can only serve as a preliminary to the 'real' (that is, quantitative)
work of social science, have declined sharply in many quarters.
Even many quantütaive researchers now accept that qualitaüve
rescarch has its own logic and criteria of validity. This reflects a
movement from a situation of methodological conflict to one of
delente, or at least of peaceful coexistence, between different
approaches to social research (Rist 1977; Smith and Heshusius
1986; Eisner and Peshkin 1990). At the same ime, criticism of
ethnographic practice has arisen from new direcuons; indeed,
from among ethnographers themselves. One reason for this is that
the widespread acceptance of ethnography has led to considerable
internal diversification of approach (Jacob 1987; Atkinson d al.
1988), which has generated internal criücism of various kinds (see,
for example, Wolcot 1980; Lutu 1981; McDermott 1982; Atkinson
and Delamont 1985). Political and philosophical trends outside of
ethnography, notably a growth in questioning of the value of
scientific ways of thinking and increased demand for knowledge
that is practically relevant, have also had their effects. These
developnents have led to criticism centring on rather different
INTRODUCTION
INTRODUCTION
issues froqmu atnhtoitsaet ivthea ta nwde rqeu aatli ttahteiv ef oruesse aorfc htheres ;c atrhloieurign hd etk h.e re aprrgauctmiceen thsa sth oef tdeinst ibnecetinv ee mcapphacaistiys eodf, etthhne ocgrirtaipcihsym t ob ecionngt rtihbautt eth tios
cbrpgtehreorimaetstTwipiratcehhieiiv eniesiicmrnsss e m a st hac,oi srcemr e oa teiutths hwnsaceutotosre n e mtlstoheipugfnae icnurcth eiiimmta pyleraal yesttkehs iisnlennyacngol etic ga eal artna isaitoupm panfrhr .toet yiocaTn ishiore neoa nwepsfl ynrcher tuares atrate drarndeiigtcenn uagaetml nrdc eb reeiirentn eiltcdca eaieksmnp mw eeer.ine tntFhd.h it ire.n en" ot (ggepotrrohaartIppnemt hh nosyitogchi rar osaealhu pc holthcdahuois aclu dbn nnme otb ostecen tdlb heeaoe o)a e fdrnno' orot relthmmoa argl aeityoxt py.tifr hm' eIaeafsips smeipeten lodwitten weridgnero esa o ptla tirrtoore ye etv bhavbiseeoeu rnuctas sor,ra eanwac tctnhnolieoudorrkndt/ o .ec prdoree nsrttshehipuaaahtrttuc eeethretta,h hlont nhnotiooe-s
vscoaocluuicaelh oerfed ae ltiihnty n?to egIrnmr astph hoey f p iinatss c t,go rmmeaputacerhri soaobfn it lwhiteiyt hat orqg ucuaamnptetitunartte isv utehp emp oenrtathitonudgre wt hoaesf wfinirocsult uldsde ihenmagv se t htpoer olufesoaeud n odtfo fiim'cetpxilopince arit(misoeene ns Ktfaortriie oegnteh'r n wo1ig9th8r a4rp hahenitcdo rpvircaaancl t iMdceeav.a iTcnehesne,
social phenomena. However, in recent years many have come to 1988). The second is often taken to imply the abolition of specialist
question this realist conception of validity, arguing that the data inquiry in favour of research carried out by pracütioners them-
which ethnographers use is a product of their participation in the selves.
field rather than a mere reflection of the phenomenon studied. The ambiguity of my tite, then, indicates what seems to me to
and/or is constructed in and through the process of analysis and be the currently ambivalent state of ethnography. On the one
the writing of ethnographic accounts. The most obvious example hand, in many fields it has achieved recognition so that the
of this ani-rcalist trend is the growing interest in the rhetorical quesuon of what is wrong with it may be interpreted rhetorically
strategies used by ethnographers, these often being treated as con- as implying its acceptance as one legitimate approach to social
stituting rather than merely represen tüng what they describe (see, for research amongst others. On the other hand, the emergence of
example, Tyler 1985; Clifford and Marcus 1986; Clifford 1988). some fundamental criticisms of ethnographic method suggests
Another area where traditional ideas about ethnography are that the quesion ought to be interpreted in a more literal way. In
being challenged concerns the relationship between research and this book I shall give emphasis to the second of these interpre
practice. In a varicty of fields conventional ethnography (along tations. I accept the value of ethnographic research and of the
with other forms of social research) has come to be criticised for methodological ideas associated with it. At the same time, I think
failing to contribute to practice; whether to political activity. it faces scrious problems. I believe that the challenge of anti-
narrowly defined, or to various relevant forms of ccupational realism and of those who advocate a more direct relationship
practice (such as that of policy-makers, administraters of various bet ween ethnography and pracice must be addressed. While I
kinds, social workers and schoolteachers). Examples of such criti regard many of the arguments of these critücs as mistaken, I think
cism are to be found in the writings of advocates of critical and that they point to important and neglected methodological
action ethnography (see, for example, Carr and Kemmis 1986). In questions. On top of this, I believe that the issues involved in the
arguing that ethnography should have a more direct relationship earlier dispute between quantitative and qualitative researchers
with practice, perhaps even being integrated with it, such critics have not yet been satisfactorily resolved. The debate simply
raise queries about what the purpose of ethnography is and should subsided, and, ironically, the present climate of ditente seems to
be, and how well current ethnographic research serves that discourage rational discussion of these issues almost as much as
purpose. Also arising here are questions about the relationship did the earlier period of cold war. Today, disagreements seem
between facts and values, and between researchers and practi often to be put down to the incommensurability of paradigms or
tioners. As with anti-realism, this type of criticism has often been are dealt with by means of pragmatic compromise. In my view
presented, not as a rejecion of ethnography in favour of some these issues sill require sustained attention, and the conclusions
other approach, but rather as capitalising on its strengths or even I draw about them have some radical implications for ethno
purging it of extraneous positivist elements. In the purse of such graphy.
2
INTRODUCTION
INTRODUCTION
So, the chapters in this book constitute my answer to the some
what ambiguous question posed in its ude. They range across the capture social reality more accurately than other approaches. OOn
threceo ahreeraesn ot fa dnedb aetfef eIc thiavvee r iedsepnotnifsieesd ;t oa nthde t hkeeyy oisfsfueer sw ihnavto Il vheodp. eI sthtued ioetdh earr eh avniedw, edit aiss cinotnrisntrsuicc titnog edthisntoingcrta pshoyci alt hwato rtlhdes. pAenodp leif
are emphasise, though, that they are separate essays. Even taken that idea is applied to ethnographers themselves it may seem that,
amths uea s btw ehsot lree stpheoyn sdeo I ncoatn cgoivnes taittu tthee a msyosmteemnatt itco twrehaattm seeenmt Ttoh emy ea rtoe craotnhsetrr uctht avne rsrieopnrse soefn rtüeanlgit yr. eIa laitryg,u ee tthhnaot gwrea pchanic reaccocgonuinsets thsiem fpalcyt
be the most pressing concerns in ethnographic methodology. that accounts are selective constructions without abandoning he
The first section of the book is concerned with the purposes of idea that they may represent phenomena independent of them-
ethnographic research, the nature of the theory and descriptions selves, and of the researcher, more or less accurately. And I sketch
it produces and draws on, and the question of how the validity of some of the implications of this 'subue realism' for ethnographic
ethnographic accounts is to be assessed, including whether they practice.
can reasonably claim to depict an independent social reality. In The issue of the assessment of ethnographic research, and
especially of the criteria on which such assessment should be
the article from which the book takes its title, I examine the
based, is the area where many of the issues dealt with in the first
concept of theoretical description, which is often treated as the section of this book become of most obvious significance. This is
intended product of ethnographic research. I consider the differ-
the topic of Chapter 4. Here I look at the two main surategies for
ent meanings of the term 'theory' that seem to be used by
specifying assessment criteria appropriate to ethnographic
ethnographers, and look at what is implied by the concept of accounts that are to be found in the literature: those derived from
ethnographic description. I show that none of the current quanitative methodology: and those presented as distinctive to
interpretations of the concept of theory is without problems, and ethnography. I argue that neither of these approaches is sais
that the common tendency to treat ethnographic description as factory, and present my own assessment framework centred on the
involving reproduction of the phenomena described is mis
concepts of validity and relevance. In discussing this framework I
leading. I argue that such description is always selective and that seek to show how validity can be conceptualised in such a way as to
for this reason the relevances and values that strueiure any void naive realism, and also to establish that emphasising the
ethnographic description ought to be made explicit iportance of relevance does not imply that all research must be
Chapter 2 takes a little further the issue of the nature of theory directy related to practice.
in ethnographic research. It seeks to show the parallels between The second section of the book concerns he relevance of
ethnography and historical research, particularly in the sorts of ethnographic research; that is, its relationship to other forms of
explanation they employ. While there has been rather littde practice. Chapter 5 deals with the problem of how an ethno
attention given to the form of explanation used by ethnographers, graphic study of a single, small-scale setting (or of a small number
there is a considerable literature dealing with historical explan- of such settüngs), at a particular point in ime can have relevance
ation. I draw on that literature to illuminate what seem to me to be for a wide audience of the kind that ethnographers typically
the two central problems of ethnographic explanation: the neces address in their publicaions. In other words, how ethnographic
sarily pragmatic and value-based character of the selection of findings can have general relevance. Occasionally, the settings
explanatory factors; and the question of the nature of the general studied by ethnographers have intrinsic interest to a large
assurmptions employed in such explanations. audience. In general, though, they do not There are two ways in
Chapter 3 discusses the realist assumptions that seem to be built which ethnographers typicaly seek to deal with this problem,
into ethnography, and the mounting challenge to them. I uy to though they do not always distinguish between these. One is to
treat the case (s) studied as representative of a larger population of
show that there is an ambivalence towards realism within ethno-
cases that is of wider relevance; the other is to use the case to
graphic thinking. On the one hand, justifications offered for
cthnography often involve the argument that it enables us to exemplify and/or establish a theory. The two strategies face
5
INTRODUCTION
INTRODUCTION
and require different sorts of evidencee., and
different problems in this chapter. My conccll usion is that the first practice, that it is invalid because the product of outsiders, and
these are cxamined is rr ather more straightforw that it is cxploitative - are considered. I conclude that they are not
sthtraante gthye ( esmecpoinridc a(lt hgeeonbryee retaitlchiasnla otiignorfnae)pr ehnecres)a, t bpuret stehnatt. neither is ni r- cleognivtiinmcainteg , faonrmd tohaf t iwnqhuilier yc,o nitvs enfutinocntaiol nr esceaanrncho ti sb ne oste t rhvee do nblyy
sued very successfullyin fluential bases for the argument that there practitioner ethnography.
spethhrtioahOscun tnloaidecgp erpob airefspo cthaah rcyiemth' i.cm oaItrolone st thda Ceidorhverayocp,tc taaretnereld a 6wta iohnI nautsom huiisbtp le isnrbo eeomt fw etahe tuieetmn h eoreasrt tshi cohnanaolvalgeelrde ad prf'ahcowryrin tai tocnhandist vwaInrai ertThi' Deo httfyehetc ee oofn irfn ne tisaalstlaksr tusueieeonccstn titsoinoohng nhi p wotoh blhfedei t qct chwhuoe ane qbnteurontaaio sqntktautu it aintisavng tcet iviov-tiqameet wuipavason.el sdi Itaen andq ieud voaa eqfcl i duhctaah itcilüaiadvtpsaeetete 'i IvrrI esea si rtdmegheauaenrtetct hid hftoeyheda raasl.t
approach and the criticisms that its advoca tes make of conven. we are not faced simply with two options, but rather with a range
tional ethnography. I then assess its viability. My conclusion is that of diffcrent possibilities. Furthermore, adopting one of these views
criücal theory is not a convincing basis for an alternative approach on one issue does not automatically commit us to particular views
to social research since it suffers from a number of serious on the others. In short, there is scope for far more han just two
problems of its own, as well as inheriing most of those that face positions overall. Moreover, I suggest that addressing method-
conventional researchers. ological issues in terms of the precepts of contrasting paradigms
Advocates of critical ethnography and others who argue for a provides an inadequate basis for deciding among the various
closer relationship between ethnographic research and practice methodological options that social researchers face.
often make strong claims for the potential practical contributüon Chapter 10 addresses the idea that quantitative and qualitative
of ethnography as compared with quantitative method. In Chapter research involve different kinds of inference. In an influential
7l argue that we must take care not to exaggerate either the likely
article Clyde Mitchell has claimed that ethnography relies on
validity of current ethnographic research or its possible contri- theoretical inference, where quantitative research depends on
buion to practice. I suggest that there is no general reason to boh this and statistical inference (Mitchell 1983). In my view,
believe that the findings of ethnography are any more valid than however, both types of research involve the same forms of infer-
those of quantitative research, and I suggest that the cortribution ence and therefore face the same basic methodological problems;
of any form of research to practice is necessarily rather srnall. Also, though they often adopt different strategies to deal with them.
Iquestion the argument that ethnographic research should be Furthermore, I suggest that there are misconceptions on the part
multi-purpose, designed simultaneously to serve the needs of of both qualitative and quantitative researchers about the nature
practitioners directly and to contribute to the cumulative of these types of inference. On the one hand there is the tendency
knowledge of the rescarch community. I argue for different forms to regard empirical generalisation from sample to populaüon as
of research designed to serve different functions and audiences, necessarily statüstical. On the other there is the idea that valid
rather than the adopion of a single all-purpose model; and that theoreical inferences can be based on intensive study ofa single
the relevance of ethnography to practice is most likely to be case in which causal or funcuonal relationships are effectively
general and indirect, rather than providing solutions to immediate visible. I argue that both these assumptions are mistaken.
practical problems. The final chapter, entitled 'So, what are case studies?', sketches
This theme of the nature of the contribuion that ethnography some of the implications of my rejecion of the quantitaive
can make to practice is taken up again in Chapter 8. I examine the qualitative divide. I take one aspect of social research method, the
case for practitioner ethnography, the proposal that ethnographic selecion of cases for study, and compare different strategies for
work should no longer be the domain of specialised researchers dealing with it: case study, survey and experiment. I argue that the
but should be open to all. The critücisms of conventional ethnogra- choice of one or other of these strategies involves trade-offs whose
phy advanced by advocates of this approach -that it is irrelevant to significance must be judged in terms of the purposes and circum-
6
INTRODUCTION
studies. And while the weaknesse. of. case
of particular they should
stances in various ways, not be
sisgtouncdoiayrl e cdra.e nsO enbaecrc eah amegreasli info,a rctaeht eead pr eo inmto isr 1ebt eoc sochhomoonwwpc l eetphxta utta hhlaisene dpi sr oinsb oletmemrenst ümes Part I
usefully of
and cannot
suggested,
competing paradigms.
There isc au rbrreinefly P ofsalcscinrigp l ewthhincohg sruamphmica rimseest hrondy odlioaggyn,o sIits monfa tk es ETHNOGRAPHY,
pcHthlroeeowasberel etvphmerarsot, bwhleehrmaet,s i.as sTo ifhnfe eo rpethdoe sirsn if bitiehlliidtsys b, oowfoe k ac niasyn n sioumtc pshro omsvoeel u fttihinoea nla sdoieslq uua taicov yytohf. THEORY AND REALITY
knowiedge. These essays are offered as a contribution to that
our
process.
NOTES
1 Throughout this book I use the term 'ethnography' in a general sense
that is broadly equivalent to 'qualitative method'.
2 Though, of course, this coexistence takes a variety of forms, from
advocacy of the combination of quantitatüve and qualitative methods
- each being regarded as having its own strengths and weaknesses- to
rcatment of the two approaches as incommensurable paradigms that
are valid in their own terms.
3 The two sorts of criücism I have idenüfied are sometimes found
Logether: see Gitin el al 1989.
4 By their very nature theories apply in many different situations, and
therefore probably have general relevance.
8
1
WHAT'S WRONG WITH
ETHNOGRAPHY?
The myth of theoretical description
In the past thirty years, ethnography and other forms of qualitative
method have moved from a marginal position in many social
science disciplines towards a much more central place. Of course,
in the case of social and cultural anthropology ethnography has
always been the predominant method; but it now has a strong
presence in sociology and social psychology, as well as in applied
areas like education and health. While this shift in methodo-
logical opinion among social researchers is to be welcomed in
some respects, it is important to emphasise that ethnography
offers no immediate solution to the problems that currendy face
social research. Indeed, I shall argue in this chapter that it suffers
from a disabling defcct itself: it is guided by an incoherent
coneption ofi ts own goals.
e rationale for ethnography is based on a criüque of quanti-
Lati, notably survey and experimental, research. The validity of
this research is challenged on a number of grounds:
That the structured character of the data collecüion process
involves the inposition of the researcher's assumptions about
the social world and consequently reduces the chances of
discovering evidence discrepant with those assumpüons.
2 That making claims about what happens in 'natural' setungs
on
the basis of data produced in settings that have been specially
set up by the researcher - whether experiment or formal inter
view is to engage in a largely implicit and highly quesionable
form of generalisation.
That to rely on what people say about what they believe and do,
without also observing what they do, is to neglect the complex
relationship between attitudes and behaviour: just as to rely on
11