Theorizing the Interview Author(S): Ray Pawson Reviewed Work(S): Source: the British Journal of Sociology, Vol

Home , Methodological dualism

Theorizing the Interview Author(s): Ray Pawson Reviewed work(s): Source: The British Journal of Sociology, Vol. 47, No. 2 (Jun., 1996), pp. 295-314 Published by: Wiley on behalf of The London School of Economics and Political Science Stable URL: http://www.jstor.org/stable/591728 . Accessed: 27/12/2012 03:20

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Wiley and The London School of Economics and Political Science are collaborating with JSTOR to digitize, preserve and extend access to The British Journal of Sociology.

http://www.jstor.org

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Ray Pawson

Theorizingthe interview

ABSTRACT

This paper attempts to breathe a little life into one of the most moribund corners of the methodological literature, namely the 'debate' on interview strategyand the supposed opposition between 'structured'and 'unstructured' approaches.From the very beginning, we tend to learn about interviewingas an issue concerning the prosand consof each respective strategy. The choice of interviewing style is thus presented as a matter of inclination towards standardizationversus sensitivity,enumeration versus emancipation,anonym- ity versusardour, and so forth. All such distinctionsare essentiallymethod-driven and have resulted in extensive technical literatures on how to achieve the chosen ends. Forgotten,therefore, in most of the literatureis the very purpose of the interview- namely to advance data in order to inspire/validate/falsify/ modify sociologicalexplanation. This paper proposes a theory-drivenapproach to the construction of the interview. It takes on board two contemporary approaches to sociological understanding, namely a realisttheory of explanation and a structurationisttheory of social being, and attempts to incorporate their principlesinto the basicstructure of the interview.The paper is illustrated with examples from the author's research with prisoners, and so hopes to inspire a donsand cons approachto the interview.

INTRODUCTION

There is a timelessquality to methodologicaldebate in sociology. Readers will recognize the mode instantly, if I give it the label of the 'polarity principle'. It operates as follows. Whatever the issue, be it a matter of fundamental strategy or the application of practicalskill, two camps of basicallyopposite persuasion will draw up and glare at each other, with the result that the development of the said method will be forever framed in a discourse of dualism. The reason for the methodologicalbifurcation is, of course, that most of the said polaritiesseem to be 'nested'.Thus, if we start with a broad epistemological opposition ('positivism' versus 'phenomenology'),this tends to have implicationsfor explanatory scope ('nomothetic'versus 'idiographic'),for data collection strategy ('quantitative' versus 'qualitative'),for population studied ('sample'versus 'case study')and so on.

Bri. Jnl. of Sociology Volumeno. 47 Issue no. 2 June 1996 ISSN 0007-1315 ) LotldonSchool of F:conomics1996

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions RayPawson 296 voices, of There are other methodological Imust not exaggerate. recognizablewhen strategywhich will be equally course.These espouse a has alwaysstruck me as principle'.This approach Ito refer it as the 'pluralist Newby's(1977) adjec- characterizedby Bell and beingmost memorably What tends to be argued methodological pluralism. truck tive,namely decent researchers have little proper, get-your-hands-dirty it is often hereis that in actualresearch practice supposed polarities,since of diverse withthese operate with a combination indeed advantageous,to have pooled sensible, that there are studies which It goes without saying interviewand methods. the formal and informal thesurvey with the ethnography, understandingof the produced a more comprehensive soand on, thereby the same example is always study. Oddly enough, institutionunder Moonies( 1984). you know it - Barker's polarity quoted,I'm sure a preference for the here is not to express tetchy intro- Mypurpose Indeed, this somewhat or the pluralist principle. of yet principle what it is, namelythe construction should be recognized for a plague of duction My task is thus to declare methodological dualism. reasoning is that another and the pragmatists.My the houses of the purists in fact leaves both between them, their opposition despitethe seeming of gulf approach to methodo- unchanged. The 'purist' methodologicaldebate which attempts to is the 'rational reconstruction' with logicalrule-making methodologicalapparatus consistencyof an entire these achievethe logical axiom. Disagreement on epistemological/ontological described somebasic to the nested oppositions axioms automatically leads approach, basic a-bit-of-this-and-a-bit-of-that Yet pluralists,with their beyond above. no methodologicalrefinements develop no new thinking, go as follows. actually Their argument tends to fuzzy mid-waycompromise. features, qualita- the for structural/institutional method is good investigationneeds Quantitative the meaningful stuff; our approachesare best for of both worlds. tive thing and make the best both,so let us do the decent by these antagon- been shaken but not stirred Sociologicalmethod has I want to promote a Against such a framework, ismsfor many a year. in order to promote to get out of the trenches, 'parleyprinciple'. In order to be genuine syn- development, there needs generalmethodological start is with the most of opposites. The place to thesisbetween the ranks and 'method'. Even of all - that between 'theory' as stultifyingbifurcation domains ('armchairtheorists' used to describe these have themetaphors theorists'versus 'underlabourers') to 'fieldworkers', 'grand to Mills( 1959) opposed One has to go right back intellectualapartheid. to note thering of at a dialogue. It is interesting Merton( 1957) for real attempts strikes a and of modern sociology, Giddens, thateven the great synthesiser the implications of his it comes to discussing state of repose when empiricalmethod. structurationtheory for with any competing theoretical structurationtheory, as as The concepts of perspectives be regarded should for many research perspective more. ( 1984: 362) sensitisingdevices, nothing

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorizingthe interview 297

FIGURE I: Structuredinterviews

Operationalization / \ Question

/ A. 2 2 \ \

gResearcher'sA g Subject's >

Analysis \ / Answer

A. 2 1/

Althoughthe scopeof thispaper is soundingever more grandiose, I do not pretendto furtherthe Mertonianor Millseanthesis here. I actually have a very modestambition, in respectof but one example.Methodo- logicalwriting on 'interviewing'typifies what I have been sayinghere (technically-driven,two main stylesand a mid-waycompromise). The paper suggests we begin to parley. What if we give theorists the responsibilityto designan interview?What might they come up with?

OLD ANTAGONISMS

In one wayor another,in orderto get theirdata, sociologists end up in talkingto people.Thus, despite possibly being the most inspected piece of socialinteraction, researchers remain at loggerheadson how to harness the flow of informationthat emergesfrom these dialogues.I refer, of course, to the battle lines between 'structured'and 'unstructured' interviewingand as a prefaceto attemptingto transcendthis distinction, I reducea few decadesof argumentationbetween the twoto the following coupleof paragraphs. Figure I representsthe flow of informationin the more formal, structuredapproaches. The subject'sideas and the subjectmatter of investigationare one and the same thing.The rationaleis to providea simple,neutral stimulus in orderto tapthe true'responses' or true'values'

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Pawson 298 Ray

FIGUREII: Unstructured! intenviews

all of individualsubjects. The usage of an identicalstimulus with is saidto allowfor propercomparison to be madeacross the respondents that entirefield of potentialviewpoints. Critics of suchan approachstress conceptualsystem is imposedentirely on the flow of the researcher's of information.The subject'sresponse is limited entirely to a set fragments.Set questionsand predeterminedresponse cat- operational the egories offer little opportunityto question,or even understand, researcher'schosen theoretical framework. II representsthe flow of informationin the unstructured Figure of (qualitative)interview. The subject'sideas and the subjectmatter are one and the samething. Data collection has the taskof investigation is creatinga conversationalsetting in whichthe informationprovided faithfulto the frameof referenceof the respondent.The investigator minimalsteerage of the researchtopic within broad areas of offers an discussionas theyseem appropriate to eachrespondent. Critics of such approachstress that the informationcollected in sucha situationis diverse anddiscursive and thus hard to comparefrom respondent to respondent. Researchersare accused of selectingfrom this massive flow of information and thusfitting together small fragments of the respondent'sutterances own preferredexplanatory framework. Whilst the data is into their is supposedto emergein 'mutual'understanding, the researcher'stheory on viewto the subject. neverclearly a Thisparticular opposition has proven more dogged and less prone to

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorizingthe interview 299 collapse into pluralismthan any other domain in the technicalrepertoire of sociology.The reason for this, of course, is the enhanced celebrationof the unstructured model as a feature of the development of certain fashionable research strategies which regard themselves not merely as 'qualitative'but as 'participatory'or 'emancipatory'(Oakley 1981; Barnes 1992). Pluralistthinking on the interview exists of course, but tends to play safe with a horses-for-courses approach - if you want factual information, go for the structuredapproach - if you want interpretative detail, go unstructured (Malseed 1987). Alongside this, perhaps, is the much used but little celebrated pluralist midway compromise, the semi-structuredinterview which recognizes that by offering respondents a chance to elaborate on their fixed-choice answers that both hard, comparableand rich, meaningful data can ensue. In advocatinga 'theory-driven'position within this debate, I will in fact seek out a midway position (c.f. Foddy 1993:73) which combines a 'structured'and 'unstructured'approach. However, I wish to do so in a manner which transcendsthe fuzzy mid-ground compromise and prom- ises more than the creationof a comprehensive,many-sided data set. The point of trying to synthesise these methods is to go beyond saying what they cover, and to show whyboth qualitativeand quantitativeinformation are needed in sociologicalexplanation and, above all, to show howit is to be melded together.

ENTER THEORY

The startingpoint for this effort is to rethink the 'task'of the interviewas well as the 'positioning'of the respondent. Perhaps the crucialdifference in what I advocate is a change in thinking about the subjectmatter of the interview (c.f. Pawson 1989, Ch. 10). Both 'mainstream'models tend to suppose that the subjectof the interview is its subject matter. The task is thus to ascertain(according to the favoured method) informationwhich is faithful to the subject'sthoughts and deeds. On the theory-drivenmodel theresearcher's theory is thesubject matter of theinterview, and the subject is thereto confirmor falsify and, above all, to refinethat theory. To many, the (italicized)statement above will seem a curiosity, since theoretical considerations are seldom taken to have such an immediate 'reach' into the world of data and the concerns of the subject. Nothing could be further from the truth. I want to illustrate this inevitable and intimate interrelationshipbetween theory and method with some of my own researchon the rehabilitativepotential of education in prisons. This is an ongoing project carried out collaboratively with 'corrections' researchersin the UK and Canada (Duguid 1981). It is an evaluation of some long-standing higher education courses carried out within prison walls, and seeks to discover whether attending such courses is associated with reduced reconviction rates. In order to answer such a question, we

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 300 Ray Pawson

FIGURE III: Baszcelements of realistexplanation

Context (C) suppose it is necessary to learn what it is about 'education'which might change an inmates reasoning about crime, and to discover what individual circumstancesand institutionalcontexts might prove favour- able to such a transformation.Now, as the reader will be able to imagine, we pursue a whole range of particular theories in exploring such questions. For the purposes of this paper, the detail of our meagre efforts in this direction are unimportant, since our hypotheses carry certain broad features which I believe to be common to the explanatorystructure of most substantive theory in sociology. It is these general features of explanationwhich must be attended to if we are to advancemethodologi- cal thinking on the interview. In my view, the starting point of any attempt to understand the syn- thesis of the quantitativeand qualitativeis to celebratethe potentialof the 'realist'approach to social investigation.Realism's head start over other attempts to codify the rules of sociologicalmethod is its commitment to 'ontologicaldepth' in explanation, that is to say-the notion that since social events are interwovenbetween variouslayers of social reality,then so must be any account of them. There has been a plethoraof attemptsto portraythe fine texture of this interlinkage,so much so that realismrisks becoming an incoherent sack-of-potatoesof a method. I cut a very long story short here by assertingthat in my book (Pawson 1989) realistexplanation can be boiled down to three key features (see Figure III).

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorizingthe interview 301 These three featurescan be woven togetherto form a fundamental explanatorystrategy for socialresearch and one thatis particularlywell suitedto gettingto gripswith the way the socialworld is put together. Explanatorypropositions are madeas follows The basictask of sociologicalinquiry is to explaininteresting, puzzling, sociallysignificant outcome patterns (O) between events or happenings or social properties.Explanation takes the form of positing some underlyingmechanism (M) which generates these outcomesand thus consistsof propositionsabout how the interplaybetween agency and structurehas consistuted these outcomes. Explanatory closure requires that,within the sameinvestigation, there is alsoan examinationof how the workingsof such mechanismsis contingentand conditional,and thusare only fired in particularhistorical or institutionalcontexts (C). As an exampleof realisttheory-making in action,let me demonstrate thisschema using the 'campus-in-a-prison'example. The startingpoint is the assumptionthat prison educationcourses do not 'work'towards rehabilitationin some undifferentiatedway. Attendingsuch a course involvesa myriadof differentevents and experiences. Explanatory work beginsby consideringcases in whichthere is a positiveoutcome (O) - i.e. thecessation of criminalactivity on release.The keytheoretical activity is to speculateupon the mechanisms(M) involvedin 'education'which mightprovoke a prisonerinto reckoningthat a way of life they once consideredjustified is justified no longer. In higher educationour weaponsare the rathergentile ones of reasoning,thought and reflection, andin a massivelyabbreviated way, I can givesome examples (in theory) of how these might sedimentinto an underlyingprocess of change. Educationmight be a spurto self-realizationand self-confidence(Ml), to economicpotential and career-building (M2), to increasedsocial skills and publicacceptability (M3); to moralchange and civic responsibility (M4), to cognitivechange and deepeningself-reflection (M5). Thesemechanisms are paraphrasedhere not becausethey are exhaus- tive and efficaciousor even particularlywise and worthy.Indeed, as everyone knows, they can be woefully far-fetchedin many prison contexts,where there are a wholerange of contraveningforces (M6) in operation.This brings me to the nextgreat explanatory imperative which is to considerthe impactof differentinstitutional and socialcontexts (C) on the processdescribed above. Any educationist would concede that one needs the appropriate'students' and 'climate'to sustain objectives. Theory thus has the job of speculatingon 'for whom and in what circumstances'such mechanisms might be influential. Prisonorganization itself, of course, is a responseto the different charactersand circumstancesof the inmates. Thus we have young offenders institutes (Cl), open prisons (C2), dispersal prisons (C3), trainingprisons (C4) and so forth as well as differentsecurity classifi- cationsfor inmateswithin each establishment. Such managerial thinking

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 302 RayPawson impingeson the successof a prisoneducation levels.Each of the courseat two different potentialmechanisms for reform aboveis goingto have througheducation more(or less)scope according to the 'typical'inmate. Thus by dint of profileof the the age (C5), offence (C6), record (C7) etc. certain custodial establishmentswill have an 'availability'of suitabletypes. Regime differences will and also bite at the institutionallevel since prisonsare also aboutsecurity, precise'ethos' of the surveillanceand control,the establishment(C8) will limit the chancesof of any rehabilitationmechanism success programme. incorporatedwithin an educational Of course,there is moreto snapshotis 'rehabilitation'than this. This littlerealist intendedprimarily to listthe kindof woulduse in a full 'ingredients'which one explanation(and evaluation). It thusacts as a tomy mainquestion about how to prelude the tracksuch ingredientsthrough data.Before we reachthat point, let into typical me addone furtherand entirely explanatoryassumption which I also take understandingthe as a prerequisitefor interview.This concernswhat Giddens 'knowledgeability'of the actor in calls the Peopleare processesof social transformation. alwaysknowledgeable about the reasons ina way which can for theirconduct but never carry total awarenessof the structuralconditions which entire set of prompt an action, nor the full set potentialconsequences of that action of prisoners (Giddens 1984). For instance, willenter an educationwith a clear areasonable understandingof whyit is choice from the (few) opportunities necessarilyappreciating available,without that certainof theirbackground criminalhistory, previous features(age, education,etc.) have madetheir morelikely. Nor willtheir reasons for candidature thewings, tryingeducation (sanctuary from choosing the lesser of several evils, a necessarilycorrespond good doss, etc.) to the outcomesthat can ensue interests,rehabilitation). In (developing attemptingto constructexplanations patterningof social activity, the for the researcheris thus tryingto developan understandingwhich includes hypotheses ingwithin a abouttheir subjects'reason- wider model of their causes and positioningof the consequences.This actorwithin sociological explanation is figurefour whichborrows from summarizedin Atthe Giddens(1984: 5). risk of repetition,let me stress representan that Figures III and IV entirelygeneral pictureof sociological instance,exactly the same explanation.For ingredients(ontological depth, the duality agencyand structure,contextually of conditionedcausal mechanisms, knowledgeableaction with unacknowledged consequences) conditionsand unintended can be found in explanationsof mobility(Goldthorpe et everythingfrom social al. 1980)to car parkcrime (Tilley tasknow is to say- if this is the 1993).The subject structureof 'theory'and 'theory'is the matterof the interview,what are the constructdata? implicationsfor the waywe

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions , Context-Mechanism - |

Theorzzingthe inte7view 303

FIGURE IV: Structurationtheory and the(partly) knowledgeable actor

Unacknowledged t Unintended conditions . __ , Knowledgeability __ > consequences of action + of action

. '

. , : .

. .

______. _ _ t

THE THEORY-DRIVEN IN I ERVIEW

Carried to the point of data collection, these explanatory imperatives prefigure a divisionof labourin the practice of interviewing, one based squarelyin a divisionof expertiseabout different aspectsof the topic under investigation.Between them the researcherand subjectknow a great deal about their subjectmatters, the trick is to get both knowledge domains - 'scholarship'and 'savvy'- working in the same direction. How does such a task break down? As a first approximation,we can say (using realist explanatory distinctions)that the understanding of contexts and outcomesshould be led by the researcher's conceptualizations. In relation to my working example, on matters such as the calculation of 'reconviction rates', the categorization of'offence' types, the measure- ment of 'educationalbackground', the phrasing of questionson 'custodial record' and so forth, the conceptual distinctions involved should be derived from the researcher'stheory and these meanings should be made clear to the respondent in the getting of information. Exploring explanatory mechanismsis another matter. In the example, these speak of the reasoning, choices, motivationswhich develop during prison education programmes. Typically, it will be the case that the researcherwill have a range of provisionalexpectations about what these may be. Equally typically, the 'hypotheses' will be 'theoreticallyover- determined' in that a whole range of potential mechanisms may be consistent with the outcomes postulated in the inquiry. Even in the 'mini-theory'of rehabilitationdescribed above, I managed to speculate upon potentialchanges in personal,economic, social, moraland cognitive mechanisms within the prison classroom. In short, in the realm of 'generative mechanisms', the researcher will often assume that the balanceof expertise lies with the informant in describingthe detailed way in which reasoning contributesto social change.

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 304 Ray Pawson

FIGURE V: Thetheory-driven interview

Teaches Learns conceptual and applies structu re conceptual structure

Motivational / accounts < ' Motivational proposed accounts refined

Here we reach the crux of my argument. In my suggestion of such a divisionof labour,the reader may be experiencing a sense of dejavu and a correspondingdisappointment. Do not the convential(purist or pluralist) models of the interviewacknowledge the difference between 'factual'and 'attitudinal'questions or between 'institutional'and 'affective'domains, and lay down a rather well-worn technicalapparatus for tacklingeach - namely the 'structured'and 'unstructured'interview? Well, yes indeed they do, but the whole point I am makingis that these distinctionsactually misunderstandthe division of labourbetween researcherand informant, and thus misspecify the requisite technical apparatus. By leading with theory, we can come to a better understandingof the divisionof expertise in the interview,which I try to capturein Figure V and which is distinctive in bringing to the fore two erstwhile hidden feature of data collection namely:a) the teaching-learning function and b) the conceptualfocusing

, . tunctlon. Fear not, dear reader. Although Figure V may look the demented scribblingof a city-centre traffic-flowplanner, it does in fact depict an informationflow which is common to all interviews.This flow needs to be understood clearly and then manipulated sensitivelyif we are to locate subject's knowledge into sociological explanations. The information highwayon the model remains a good old-fashioned structuredquestion

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorizingthe inte7view 305 and answer sequence running through the centre of the figure. Thankfully,the most common interviewingexperience is that if one puts a straight question, most of the time one gets a straight answer. This little miracle happens routinely because researcher and subject share a taken-for-grantedset of conceptual building blocks. Social interactionis premisedon this realmof the accepted-as-realwhich allows us to knowwe are talking about the same thing. (Giddens 1984: 331). This item-by-item, utterance-by-utterance,membership category-by- membership category understanding is, however, only the beginning of the story.Our everydayfamiliarity with conversational practices will always make interviews happen but not alwaysallow for the apposite data to be constructed.This is where the 'teacher-learner'function comes in. We are interested here in concepts to do with 'outcome'and 'context'elements in the explanatorystructure, and the issue is to consider how can we knowthat the subjectis attending to the researcher'sunderstanding of these items. The traditional(structured interview) answer to this problem is to rely on precisionin question wording and clarityin operationalization.Whilst the precise turn of a phrase is, of course, important,my basicobjection is that operational definitions alone are rarely sufficient to teach the subjectthe underlying research tack. In reducing the inquiry to variablesand values on variablesthey, in fact, construct meaning in a manner contraryto the way theory will have been devised. Theory hasa complex and deep structure(recall Figure I I I ) and basically the researcher will have come to learn the meaning of any individual concept therein, through its place in these elaborate propositional nets. Method-driven interviews traditionallypay little heed to this important source of conceptualclarity. So whilst researcherswill know full well they are askingquestions about a prisonerseducational background (C l ) as part of a proposition about how further education (Ml) in providing cognitive change (M2) might produce more potential for rehabilitation (°l) in inmateswho havebeen deprivedof earlyopportunities (C2), the inmatecan remain blithelyunaware of these purposes and meanings. Usuallyit is the case that this collateralinformation is smuggled in, ratherimplicitly across the pages of the questionnaire. What I am suggesting here is that the researcher/interviewerplay a much more active and explicit role in teachingthe overallconceptual structure of the investigationto the subject, for this in turn will make more sense of each individual question to the respondent. Inopracticethis means paying more attention to 'explanatory passages',to 'sectional'and 'linking'narratives, to 'flowpaths' and 'answer sequences',to 'repeated'and 'checking'questions and so on. It also means being prepared to take infinite pains to describe the nature of the information sought and thus a sensitivityto the struggles the respondent may have in using what are ultimately the researchers'categories. This function is depicted in Figure V (on the north-westernring-road). As every interviewer will know, respondents also travel these outer perimeters.So, as well as providing straightanswers to straightquestions,

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 306 Ray Pawson subjects ponder (mostly in silence)-'who is this person?', 'what is she after?','why am I being asked?','what have others said?','what should I be saying?', and so on. The theory-drivenmodel I am presenting here has a unique tack on such 'hypothesis-seeking'behaviour. The aim is not minimizeit (as in the structured approach), nor to wallowin it (as in the unstructured approach), but to channelit. That is to say, the battery of questions posed and explanatory cues offered should be understood as putting the subjectin a position which allowsthem to think (stillin silence, incidentally) -'yes, I understand the general theoretical tack you are exploring, this makesyour concepts clear to me, and applyingthem to me gives the following answer'.This partricularinformation flow is depicted in the 'north east' of Figure V. Elaborate as it may seem this in fact describes the thought process which underlies the typical question and answer sequences found in most detailed formal questionnaires and interviews. Elsewhere (Pawson 1989, Ch. 10) I have provided some workingexamples of how to facilitatethe teacher-learner function. However, a further step is needed in respect of those aspects of explanation to which intervieweeshave a privileged access, namely their own reasoning processes.This is where the 'conceptualfocusing' function comesin. Such a process is intended to describethe collection of data on explanatory mechanisms (M), the coverage of which is conspicuously absentin Pawson (1989). Thus the 'southern' ring-road in Figure V depictsan extension of interviewingprocess which allowssubjects to have theirown say (decidedly out loud) about how their thinking has driven themto particularactions. The key point, however, is that they deliver these thoughts on their thoughts in the context of and, (perhaps) as a correction to, the researcher's own theory. To explain - the overall structureof the researcher'squestions will, in general, contextualizethe areain which the subject'smake decision and highlight some potential decision making activity which goes on therein. The subject'stask is to agree, disagree and to categorize themselves in relation to the attitudinal patternsas constructedin such questionsbut also to refine their conceptual basis.It is at this point that mutual knowledge is really achieved. The subjectis saying in effect 'this is how you have depicted the potential structureof my thinking, but in my experience it happened like this . . .' In short, I am postulating a formula for 'attitude' questions (more properly, items in the cognitive and affective domainsgenerally) in which the respondent is offered a formal descriptionof the parametersof their thinking followedby and opportunity to explain and clarify this thinking. Torepeat, sociologicalexplanations offer hypothesesabout their subjects reasoning withina wider model of their causes and consequencesand the attraction of the particularmodel is that it reflects a division of labour whichis best able to put these pieces together. An 'example' is overdue at this point (and shall be delivered!). First, I should point out that what I describe as the 'formula' in the previous paragraph does not imply the existence of some singular and unique

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions T. neorzzang* * the tnterview. 307 techniquewhich capturesthe idea. The 'I'll show-you-my-theory-if- you'll-show-me-yours'strategy has echoes in a number of existing methods.Two that come instantlyto mind are vignettes (in which the stimulus stories are constructedto smuggle in the key theoretical parametersunder investigation,upon whichthe respondentis askedto reflect)and pilot interviews (whichsay - answerthese questions and please alsotell me whatyou thinkof'em).

DONS AND CONS

Mydetailed illustration comes (appropriately enough) from some pilot interviewingI did on a smallscale UK versionof the campus-in-a-prison projectat HMPFull Sutton. Towards the end of the studiesof the first cohortof men throughthe course,I madean attemptto drawan overall pictureof the men'saccounts about how (if at all)the coursehad changed theirattitudes, reasoning, outlook, etc. There are, of course,no standard questionnairesor attitudescales ready-made for sucha specificpurpose, so I hadto inventone. WhatI endedup doingwas modifying a 'discussion document'produced by the then NorthernRegional Education Officer whichtook as its taskto list and elaborateupon the potential'aims and objectives'of the prisoneducation service. The adaptationtook the form of rewritingeach statementof aspirationcontained in the document,so thatthey became a sortof attituderating questionnaire to whichthe men couldagree/disagree and so forth. As a researchinstrument, this could certainlybe improvedupon. It omitssome entire categories of potentialchange and I'm pleased to report that we are workingon a much more comprehensiveattack on the problemin the Canadianversion of the study. However,the example does have the basic methodologicalfeatures alluded to here. It was writtenby an 'insider'with an eye on encouragingpenal educators to look beyond getting their students through 'GCSE','City and Guilds'or whatever.It relatesthe classroomexperience to broaderconcerns about prisonand after. It contains(and this is the importantbit) the accumu- latedwisdom (or as I wouldprefer to say- 'theories')of practitionerson personalchange associated with educationalprogrammes in prisons.A little sub-plothere is that given its origins,which I made knownto my subjects,there was a 'whiffbof the HomeOffice about the constructionof the items.This, I recall,added a littlespice when I cameto get the mento completeand comment upon the questionnaire. The actualform of questionnairewas as follows.The studentswere presentedwith the list of statementsrepresenting possible goals of a prisoneducation course and they were asked to respondaccording to each item in respectof how the statementapplies to theirexperience of the FullSutton course. They were required to placeanswers in one of four categoriesas follows

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 308 Ray Pawson This applies to me to a considerableextent 1 to a moderate extent 2 to a slight extent 3 not at all 4 There follows a list of the statements and for each I record the mean response score using the scale as above. The course: Mean response a) helps inmatesto acceptthemselves and their feelings more fully | 3.4 |

b) helps inmatesto become more self-confidentand self-directing | 2.4 |

c) helps inmatesto become more acceptablepersons to society | 3.1 |

d) helps inmatesto accept more realisticgoals for themselves | 2.5 |

e) helps to change the moraloutlook of the inmates | 3.1 |

f) helps inmatesto become more flexiblein their opinions | 2 l

g) helps inmatesto behave in a maturefashion | 2.8 |

h) helps inmatesto change their maladjustivebehaviours | 2.8 |

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorizingthe interview 309

p) helps to providea basison which inmatescan build new life | 2.6 1

q) helps inmatesto achievecontrol over their actionsand choices | 2.6 |

Rather a lot can be learned by the simple device of orderingthe responses from those features which the men found consistentwith their own experience down to those which they considered inapplicable. As ever in data analysis, it is the patterns of response we are seeking to uncover and this can be aided by the device of superimposingsome breaks and boundaries within this rank order. In the following I distinguish those objectives which collectively met with i) considerable to modest agreement, ii) moderate to slight agreement and iii) slight to no agreement. I also insert a mid point axis (score 2.5) which can help us see the general balanceof sentiments.

| 1-2 | considerableto moderateagreement 'improve powers of concentrationand persistence' 'become more flexible in opinions' 'discernpreviously undiscovered talents'

| 2_3 | moderateto slight agreement 'experience success' 'acceptantof others and other points of view' 'self-confidentand self-directing' 'acceptmore realisticgoals' ...... (2.5) 'behavein a more mature fashion' 'correctpersonality characteristics in constructiveways' 'changetheir maladjustivebehaviours'

| 34 | slightto no agreement 'more acceptablepersons to society' 'change moral outlook' 'acceptthemselves and their feeling more fully' 'assumeresponsibility for their own lives' 'rejecttheir criminalpast'

It is possible to make some rough and ready sense of the above configuration by seeking to uncover the 'themes' which underlie the difference between those aspirations with which the men concur and those of which they are sceptical. It can be seen readily enough that the items with which the men concur concern the improvement in 'mental powers','learning skills', 'flexibility of viewpoints'and so on. In short, the

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 310 Ray Pawson connecting thread here is a recognition of personal change along a dimension that perhaps speaks for itself- namely academicrelated change. The roots of scepticism about the transformativecapacity of education seem more diverse. There would seem to be (at least) two distinctive features which underlie doubt. The first is when the items refer to public acceptability.The thinking here, presumablyis that all prisonersknow they are no longer free agents, expect a tough reception on release and do not expect things will be dramaticallydifferent, with or without a diploma. The second dimension which the inmates declare untouched by their presence in the academycan be thought of as items pertainingto personal character,especially those statementsgetting at their inner self and most specifically, of course, the only item on which there was unanimity, namely item (j) and its insinuation that education allows them to reject their criminalpast. What we have to this point is an unremarkable, not to say undis- tinguished, piece of attitudinalscaling which produces,incidentally, some rather unwelcome results - there being only the faintest whiff of 'rehabilitation'in all this data. Orthodox methodologicalthinking divides habituallyat this point. The quantitativeinstinct would be to get more formal- the pilot items could be beefed up, a proper factoranalysis could be attempted, and a rather larger sample could be constructed (have I mentioned that the above data is culled from seven inmates?) The qualitative instinct would be to ditch the lot as arbitrary number- crunching and to go for personal involvement as the high road to understandingpersonal change. It is possible to escape these weary old methodologicalstraightiackets by considering more closely the men's reasoning in response to being presented with this batteryof propositions.I can stillrecall vividly the Full Sutton students' outward reaction to this exercise two years on. They moaned, they groaned; a couple of them were on the point of refusing to complete the task at all (until I threatened them with more lectures on mobility tables). The roots of this discomfort were exactly the same as most people feel when they are asked to complete such exercises, but in this case MAGNIFIEDseveral times. That is to say, attitudinalstatements are normally regarded as irritating simplificationsand only with some generosity can one reduce the richness of life's experiences down to the pre-set categories. In this particularinstance, some of the simplifications were regarded as more than mere irritationsbut were seen as positively insulting (in certain respects which I will come to in a moment). The methodologicalpoint that shines through this, however, is that the questions perform a much more significantfunction than as the specific stimuli to respective responses.Neither are they an invitationfor respondents to muse on whateveraspects of their experience are central to them. Taken as a piece, these formal questions set a clear agenda which represents a body of theory, offering up the researcher's potential explanations for a closely circumscribedset of actions. Their key role,

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorizingthe interview 311 therefore, should be to involve the respondent in a closer articulationand clarificationof these theories. This can be done (and was done in this instance) by the simple device of getting the respondents to explain why they have plumped for the particularresponses to the particularitems. This is a common place enough tacticin semi-structuredinterviewing, but one that is never understood in the waythat I am presentingit here, namely -as a superbvehicle for the here's-my-theory-what's-yoursstrategy of data collection. What is induced by this process is a great deal of conceptual hair-splittingand this is precisely the kind of data which leads to better focused explanation. Let us look more closely at a couple of examples of this process at work. Question (j)about inmatesrejecting their criminalpasts because of contact with education got short shrift, yet the subsequent account of whythe statement is disregarded, prompts the inmates into a much more subtle level of reflectionon their own reasoning. The following extractsgive the accounts of four men on why they registered 'not at all' in answer to this question. As always,transcripts fail to give the underlying 'mood' of the answerwhichmightbe summarizedhelpfully here as 'furious','imperious', 'cool','cooler', respectively. - Butto reject yourcriminalpast, I'm notrejectingit. I'mnotrejecting what I've done, but you don't rejectit do you, you . . . you take and you . . . you step on from there and you try and learn from it. You don't go, well you don't know. Its a part of . . . its a part of you. - I know why overall I've scored so low its because its I . . I . . I do have thin thing umm . . . about personalresponsibility, you know I . . . I acknowledge that I'm in prison through my own fault, and umm . . . if I'm going to stop coming into prison it will be down to my own motlvatlon.. . I mean its (the question) assuming that its (the course) is gonna change somebody's whole outlook on life and behaviour and everything I don't relate to it, don't relate it at all. I mean I can see that the more educated you are the more you can get away I suppose. But I don't connect with it at all. In my case, when I commit a crime I know I'm doing wrong and I knowif I'mgoing to get caught, I'llgo to prison.So its not as though I'm rejecting it. A similartheme emerges in relation to the question of whether education can help inmates to accept 'themselvesand their feelings more fully'. I feel that I excepted myselfand my feelings before I cameonto the course, before I knew of the existence of the course. I fully accepted my feelings a long time before I came here. I agree that this course and education still could really help those people who don't reallyunderstand yourself (themselves).Firstly I

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 312 Ray Pawson

understand myself and I don't reallysee that (the course) leading me in to that direction. Really(this) is one thing I have to discover myself. What even these few clarificationsreveal is a tension in most of these prisoner'sbeliefs about education. It is recognized as 'improving'and yet they want to take credit for the improvement. They 'learn' but not as empty buckets filled with knowledge against their betterjudgment. It is recognized that education can lead to self-understanding but only becauseprison conditionsare alreadyconducive to intense self-reflection, since they provide many hours, days and years of opportunity for the same. This tension was perhaps best expressed by 'No7' who was most hostile to this particularphase of the researchbecause he felt the questionswere 'patronizing'and that they were full of'civil servicerhetoric'. He set out to swat down their 'preconceived ideas' with a series of'not at alls' in his written responses. Under follow-up questioning, he relents a little and finds that he was 'makinga nonsense of some of his own scoring.'Basically he back tracksbecause - I will go down the road of agreeing, because, err . . . I feel that education is a civilizing process ... it could well prove a contributing factor in the adjustment to acceptable behaviour. Change is something that comes within but you would be taking on board education.... it's a catalyst... more than a catalyst,as I've said before its a civilizingprocess Here is another man choosing his words carefullyand, being an educated sort, he does indeed know his 'catalysts'from his 'contributingfactors'. Actually,the most telling phrase he uses here is probably'taking on board education'and this is an image which comes through most strongly in all of the men's discussion. If we takeas thestarting point thatmany prisoners routinelyengage in self-scrutinyand choice-making then what a rigorousperiod of educationcan perhaps provide, is a meansof extending,deepening and affirming suchprocesses. Or to put this back into prison parlance. - It's not the course that's changed you as such, it's you've developed an interest inside you, you know. - By and large you've got your own . . . you've got your own way of working . . . and you can work in a number of directions . . . you're sort of given advice on which way to go and that, but at the end of the day its your choice. The sprinkling of metaphors in the above on 'interests inside you', 'taking on board of education', 'stepping on from there', contains importantmessages about the importanceand nature of cognitivechange as a potential mechanism for rehabilitation. The upshots of such reasoning will be explored in the researchto come. Here I should return

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions Theorazingthe interview 313 to the general methodologicalsignificance of this tale. I readilyadmit that the example came unfortunatelybefore the rationaleI am in the processof relating. To me it came as a (minor methodological) Eureka - after months of going round the houses, trading anecdotes about early educational experiences, the nature of crime, their likelihood of re- offence or rehabilitation,the influence of family, peers, teachers, Uncle Tom Cobbley and all - this simple formal schedule did the trick. All at once they talked about theirworld in mylanguage.

CONCLUSION

This paper ought to have brought on a strong sense of recognition to researcherswho will know that the processes described here are already part and parcel of the negotiation of meaning which goes on in any substantial interview. The paper will have worked if these same researchers believe that the conceptual framework elaborated here pro- vides a better methodological foundation than hitherto for understanding, controlling and developing these negotiations. In particularI have tried to rethink the boundary line between the researcher'sand subject's knowledge. In advocating this approach as one with general utility in data construction,I should make it clear that I am not simply putting the 'trick' or the 'technique'up for inspection.All this is not simplya matterof piling up a set of attitudinalstatements and getting them explained. What I am actually counselling is the informationfZow as depicted in the model in Figure V. Its key aspect is the creation of a situation in which the theoreticalpostulates/conceptual structures under investigationare open for inspection in a way that allows the respondent to make an informed and criticalaccount of them. Much more could be said about when, why and for whom one would adopt the approach. Here I only need stressthat it involvesa highly specificand carefully planned route marchwhich goes between the qualitativeand quantitativetraditions.

(Date accepted:April 1995) RayPawson Schoolof Sociologyand Social Policy Universityof Leeds

BIBLIOGRAPHY

Barker, E. 1984 The Making of a Moonie, Sociological Research, London: George Oxford: Basil Blackwell. Allen & Unwin. Barnes, C. 1992 'Qualitative Research', Duguid, S. 1981 'Prison Education and Disability,Handicap and Society7(2): 115- Criminal Choice', Canadian Journal of 24. Criminology23(1 ): 15-24. Bell, C. and Newby, H. (eds) 1977 Doing Foddy, W. 1993 ConstructingQuestiorzs for

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions 314 Ray Pawson

Interviewsand Questionnaires,Cambridge: Oakley,A. 1981 'Interviewingwomen: a CambridgeUniversity Press. contradictionin terms'in H. Roberts(ed.) Giddens, A. 1984 The Constitutionof Doing FeministResearch, London: Rout- Society,Cambridge: Polity. ledge. Goldthorpe, J. H. et al. 1980 Social Pawson,R. 1989A Measurefor Measures: Mobility and Class Structure in Modern A Manifestofor EmpiricalSociology, Lon- Britain,Oxford: Clarendon. don: Routledge. Malseed,J. 1987 'StrawMen', Sociology Tilley, N. 1993 'UnderstandingCar 21(4):629-31. Parks,Crime and CCTV',Police Research Merton, R. K. 1957 Social Theoo and Group Paper No 42, London: Home SocialStructure, Greencoe: Free Press. Office. Mills, C. Wright 1959 The Sociological Imagination,New York:Oxford Univer- sityPress.

This content downloaded on Thu, 27 Dec 2012 03:20:37 AM All use subject to JSTOR Terms and Conditions