Resistance and Adaptation to Criminal Identity: Using Secondary Analysis to Evaluate Classic Studies of Crime and Deviance

Qualitative data offer rich insights into the social world, whether alone or in tandem with statistical analysis. However, qualitative data are costly to collect and analyse. Moreover, it is a commonplace that only a portion of the data so labouriously collected is the subject of final analysis and publication. Secondary analysis is a well-established method in quantitative research and is raising its profile in application to qualitative data. It has a particular part to play when research is on sensitive topics and/or hard-to-reach populations, as in the example considered here. This article contributes to discussion of the potential and constraints of secondary analysis of qualitative data by reporting the outcome of the secondary analysis of a key study in the sociology of prison life, Cohen and Taylor's research on the long-term imprisonment of men in maximum security. The article re-visits Cohen and Taylor's original analysis and demonstrates support for an alternative, if complementary, conceptualisation, using archived data from the original study. Among the methodological issues discussed are the recovery of the context of the original fieldwork and the role of secondary analysis in an incremental approach to knowledge production.


Adaptation and Resistance in Long-Term Imprisonment
In their book Psychological Survival Stanley Cohen and Laurie Taylor examine adaptation to long-term imprisonment using the case of maximum security prisoners in Durham Prison.The book is important for several reasons, the most notorious being that, following its publication, the authors were barred by the British government from conducting follow-up research in prisons.The book remains unique in offering case study data about the most violent and dangerous prisoners in the British prison system.It attracted considerable media attention at the time, and is still used in teaching and research.
It is also very much a book of the times in which it was written: the research was done in the late 1960s and the book was published in 1972.Using methods associated with sociological naturalism (and the 'empowerment' orientation of some postmodernist approaches), it presents what could be called a 'sympathetic' account of the prisoners (although Cohen and Taylor feel their approach would more accurately be described as appreciative' in the sense suggested by Matza 1969).The book says little about the crimes the prisoners had committed, or their criminal career up to the point of receiving long sentences.Rather, it focuses on the ways the prisoners deal with their situation and depicts them as having a political and at times almost heroic adaptation to long-term imprisonment.
The analysis Cohen and Taylor developed was in line with the radical psychiatry of the time, represented by authors such as R. D. Laing and Timothy Leary, whom they quote, and with critical social reaction or labelling theory, a position which transformed American sociological criminology during the 1960s and was beginning to emerge in Western Europe.These positions were organised around an oppositional critique of institutions and a defence of those who were oppressed by the state.Long-term prisoners were, then, one more group whose rights were being crushed by state institutions, in particular the Prison Department of the Home Office.Cohen and Taylor could even warm themselves in a little reflected glory because, like the prisoners, they were being oppressed by the Home Office, which regulated their research and ultimately prevented their follow-up study (see Cohen and Taylor 1977 for an account of these events).
Notice that the book's title is 'Psychological Survival'.This itself carries a message.While the book discusses psychology a good deal, much of it is a critique of psychological analysis of criminal offending.Indeed, its appendix is largely an attack on psychological research being done at the same prison (and with some of the same prisoners) when Psychological Survival was approaching publication.The authors raise the suspicion that this project was intended to provide ammunition against their own findings.Now in so far as psychology is interested in how prisoners adapt to long-term imprisonment it may be regarded as ameliorative, oriented to helping the men adjust to their position.In contrast, the sociology of the time was transformative and aimed strongly at public opinion.It wanted to create an understanding -even a sympathy -for the prisoners.This was, then, a book which aimed at having an impact on public thinking about long-sentence prisoners, and using this to embarrass the prison authorities and pressure government.
That a general critique was intended rather than one aimed at specific policies is shown by the fact that when the book was written major changes were already under way in the handling of maximum security prisoners.Following the critical Mountbatten review, maximum security wings were being closed, with the prisoners dispersed in small groups to a large number of prisons.While a re-analysis of Cohen and Taylor's data shows material relevant to the competing policies of dispersal and concentration, their comments on the issue are brief and equivocal.This suggests that they regarded the specifics of policy as mere tinkering and that their interest lay in a general repudiation of the prison system.In the appendix attacking the psychological research at Durham, they suggest it was a sop to public opinion which had, since abolition of the death penalty, become concerned about the effect of long-term imprisonment on cognition and intellect.To quote the authors, 'we predict that... projects such as we have described will be used to provide reassurance about the nature of long-term imprisonment.They will do little to disturb the political and legal structure' (Cohen and Taylor 1972:207).
Cohen and Taylor were interested in the analogy between the situation of long-sentence prisoners and others placed in extreme environments where the routines on which normal life is based are disturbed.While one might assume this would lead to comparison with those in 'total institutions' like psychiatric hospitals, Cohen and Taylor more often make comparisons to the situation of polar explorers, African expeditionists and those undergoing long-term physical treatment in hospitals, that is, groups regarded as heroes or objects of sympathy.We can see a methodological effect here.The researchers referred their analyses to the prisoners themselves for comment.It was the prisoners who asserted 'the inadequacies of ... the psychologists' concern with specific sensory rather than general psychological problems (and) the sociologists' reliance upon large-scale surveys of medium-term prisoners' (Cohen and Taylor 1972:41).Cohen and Taylor's focus shifted from modifying existing analyses to accommodate the case of long-term prisoners to 'looking at the ways in which men in general might react to an extreme situation, a situation which disrupted their normal lives so as to make problematic such everyday matters as time, friendship, privacy, identity, self-consciousness, ageing and physical deterioration' (ibid., emphasis added).So they were approaching this group as examples of men in general rather than as instances of that unusual group of men who have murdered, raped, engaged in terrorism or committed violent robberies.
The role of cultures of resistance and adaptation is important in theories of the emergence and confirmation of deviant identity and was applied in the prison context in Clemmer's concept of 'prisonization' (1961), a form of secondary socialisation in which inmates learn to adapt to prison as a way of life.These adjustments have been documented along the range from withdrawal to continual rebellion.Cohen and Taylor complained that such analyses do not seek to understand what these adaptations mean to prisoners themselves.They accept the importance of the 'inmate code' and convict subculture (Sykes and Messinger 1960) in helping prisoners to get by, but want to stress 'the conscious, creative nature of the sub-culture' and 'the positive nature of the dogmas, mythologies, beliefs, modes of adaptation and feeling which are part of (the) day to day experience of people who find themselves in extreme situations' (Cohen and Taylor 1972:58).In this social constructionist approach it is worth noting that Cohen and Taylor emphasise the element of Berger and Luckman's argument (1971) concerning the way that constructing meanings are directed towards justifying one's experiences.
While this is an important way to capture the men's experience, it does seem to stray from why the men were prisoners in the first place.In fact the authors are explicit that they have little to say about the aetiology of the men's criminal careers.Yet these men are not all the same, and they do not appear from the data to have all found positive 'dogmas, beliefs and modes of adaptation' with which to 'justify to themselves' their experience.Indeed, some of them seem desperately lacking in an understanding of their predicament and one has to note that the book does not quote a single item of data in which the men express remorse about the crimes which led to their imprisonment.
Cohen and Taylor say that sensitivity to the men's position is no more than social scientists would apply to understanding the plight of the blind, the old or the poor.Such sensitivity is needed even if the prisoners got what they deserved, because, without appreciating how their everyday world has been broken by long-term imprisonment, their adaptation may be seen as somehow 'natural' rather than 'personally constructed as a solution to intolerable problems' (Cohen and Taylor 1972:44).It seems to us that if we want to understand this personally-constructed solution we need to understand what led to the man's imprisonment as well as the world he has constructed to get by once he is there.After all, would we attempt to understand the blind person's adaptation without discussing their blindness -whether it was partial or complete, whether the person was born blind or their vision deteriorated, and so on?If anything, the precursors of prison life are even more important for another reason.While other prisoners can choose friends from many other inmates, the small number of long-term prisoners, and their isolation in maximum security, restricts their companionship.Thus, in constructing their adaptation, ties to their past may play a greater part relative to other prisoners than current experiences.
The Durham E Wing prisoners were an 'elite', ostensibly the most vicious prisoners in the system.Their notoriety meant their names were often in the media.While this made for distance from fellow inmates and guards (unlike the affinities of staff and inmates found in the period's other notable British prison study, the Morrises' Pentonville 1963), the real division came about because the men regarded themselves as an intellectual elite and as cosmopolitans.They were mainly from London, while the guards were from the economically-depressed, rural North-east.The prisoners' intellectual claim takes more explanation.To some extent it is an artefact of method.The reason Cohen and Taylor were in the prison in the first place was not to do research.They were there to teach adult education classes to long-sentence prisoners.So they came in contact only with the more active and education-oriented prisoners.One prisoner used to read Plato by night and force other inmates to listen to his lectures on Plato the next morning; Cohen and Taylor say this was more to do with exercising dominance than hunger for knowledge.Whatever the motives, some prisoners were highly literate.One wrote 1,000 words of a novel every day, another wrote a 20,000 word paper about his army life.They read several books a week, tending to read everything by an author consecutively, for instance, the complete works of Freud or Shakespeare.The prisoners' letters to the researchers are often couched in academic language.Unsurprisingly, the men were keen readers of prison novels and memoirs.
Let us make a last point about the original analysis.It is true that the book contains a chapter titled 'Identities, biographies and ideologies'.Here one might expect to find an aetiological account of the men's individual criminal careers.However, the focus of discussion is the last word of the chapter title, 'ideologies'.The authors briefly report that the men had a variety of backgrounds and criminal careers.The descriptions of their crimes are general: for example, that there is a group of prisoners who run legitimate businesses but occasionally organise a major robbery.The kinds of crime they do are related to a series of 'ideologies' which are not derived from accounts provided by the men or other data but from accounts of anarchists and terrorists, concentration camp survivors, American outlaws of the Bonnie and Clyde variety, and members of the Black Power movement.This approach is even applied to those imprisoned for violent sexual offences: Society consistently attributes to sex offenders ... a set of wholly irrational motives: 'something must have come over him', 'he didn't know what he was doing'.In other words we deny them an ideology, insisting that their behaviour is initiated and sustained by sudden psychic impulses or aberrations, rather than proceeding from a self-conscious view of what is desirable in sexual behaviour.(Cohen and Taylor 1972:185) This neglects both the psychological conceptualisation of sexual offending and their own data.One would find it hard to identify adherence to ideology in some of the accounts in their data of the self-rejecting, isolated and compulsive behaviour of the sex offenders.

Methodology of the Original Study
The E Wing at Durham was a notoriously brutal environment.Its reputation was of a place where difficult prisoners were 'broken'.Following the Great Train Robbery the media made play with the idea of a new, more ruthless, violent and organised criminal.E Wing was used to test ways of dealing with this 'new' criminal type, including electronic surveillance, both by CCTV and sensors embedded in the floors, electrified door handles, lighting of cells 24 hours a day, dog runs, armed guards with gas masks, and a machine gun nest on the external wall.When the train robbers were taken to Durham it was widely assumed an escape would be attempted.The Chief Constable of Durham announced to a press conference: I am satisfied that Goody's [one of the train robbers] friends were prepared to launch something in the nature of a full-scale military attack, even to the extent of using tanks, bombs and what the Army describes as limited atomic weapons.Once armoured vehicles had breached the main gates there would be nothing to stop them.A couple of tanks could easily have come through the streets of Durham unchallenged.Nothing is too extravagant.(Quoted in Cohen and Taylor 1972:13) This rather idiotic statement was reported seriously by the media.Although the Chief Constable later claimed he said it to divert attention from the secret transfer of certain prisoners, it serves to indicate the paranoid attitude toward E Wing.Troops were called to help the E Wing guards.
The prisoners were classified as Category A, 'prisoners whose escape would be highly dangerous to the public or the police or to the security of the State'.Table 1 shows the main charges against 138 of the Category A group.By 1970 E Wing and other maximum security prisons held 225 men serving ten years, 218 serving over ten years and 159 serving life.E Wing witnessed hunger strikes, escape bids, riots and protests.It was so notorious that journalists formed a club with its own tie for those who had a steady line in stories about it.
The authors report that prisoners first suggested they should write up their classroom discussions as research.Cohen and Taylor hoped the book would be a 'survival manual' for future prisoners.Attendance at Cohen and Taylor's classes varied between two and twelve men; in total, fifty passed through it.They got to know ten intimately and an equal number fairly well.The class competed with the two hours a day the men could watch television, so the authors felt they did well to get these numbers.It is a small sample, but the authors strike a radical posture towards methodology.They say most academic research tells a 'chronological lie', presenting an orderly sequence of defining the research problem and hypothesis, using appropriate tests and then analysing the data.Instead Cohen and Taylor say they 'started without a problem, evolved a set of methods while we worked and ended up with a collection of observations, anecdotes and descriptions rather than a table of results ' (1972:32-3).They also say they 'talked more intimately to these men than to any other people we know' (ibid.:33).They initially worried about teaching dangerous criminals but, importantly for our analysis, they say 'as our belief in the stereotypes diminished, so our knowledge of the actual careers of the men became less critical' (ibid.:34).They used four research methods.First, there were unstructured group discussions in the classes, at the end of which points were summarised and sometimes discussed with the prisoners.Secondly there was correspondence between the men and researchers.Thirdly, literature extracts on subjects like prison, murder or sexual deprivation were read in class and discussed.Prisoners said which was most realistic and best captured their feelings.Fourthly, the authors regard their writing-up as a method because inmates commented on drafts.There was also a survey questionnaire administered by interview schedule to 100 long-term prisoners in other prisons.According to the preface the schedule was designed and administered by some of the inmates, but the book makes limited use of this data.
After the idea for writing up the classes as research emerged, Cohen and Taylor told the prison authorities what they were doing.The Governor endorsed their research, and it was submitted to the Home Office.The Home Office eventually rejected it, but Cohen and Taylor replied that by then the work had been done.The Home Office then banned them from visiting exmembers of the class and censored their correspondence with prisoners.It was worried about security and sensationalism, and criticised the small sample, lack of a control group, failure to use 'objective tests' and to monitor psychological changes over a longer time.

Secondary Analysis Methodology
Although secondary analysis of quantitative data is a common method, the same cannot be said of qualitative data (Heaton 1998).Yet the basic purposes of secondary analysis are similar whatever the data type: to pursue interests distinct from those of the original work or apply alternative perspectives to the original question (Hinds, Vogel and Clarke-Steffen 1997).Heaton suggests three analytic purposes: additional in-depth analysis; additional sub-set analysis; and new perspective/conceptual focus.Our work is of the third type, another example being Bloor and Macintosh's (1990) re-analysis of their own data on client resistance to surveillance in medical settings.Hammersley (1997) suggests that secondary analysis of qualitative studies of similar research populations can assess their generalisability, thus compensating for what is customarily a weakness of qualitative research.We also feel secondary analysis has a particular role in qualitative research concerning sensitive topics or hard-to-reach populations, because researchers can best respect subjects' sensitivities, and accommodate restricted access to research populations, by extracting the maximum from those studies which are able to negotiate these obstacles.Secondary analysis can protect the sensitivities of subjects and gatekeepers by ensuring they are not over-researched, and can position further enquiries so that they ask what is pertinent to the state of analytic development, building on, rather than simply repeating, previous enquiries.
In Britain the establishment of the 'Qualidata' Archival Resource Centre is a move towards providing the infrastructure for secondary analysis of qualitative data.But the period since Qualidata was established in 1994 has not exactly seen a stampede of researchers into secondary analysis, perhaps because research careers are made by discovering new things, not extracting the maximum from existing data.Also, although Qualidata encourages researchers to deposit their data, finds an archive to hold the deposit, advises depositor and archive on copyright, confidentiality and anonymisation, and catalogues the materials, it does not prepare them for use.Researchers still face practical problems in doing secondary analysis.Most of the material exists only as paper copies, not in electronic form; in fact, when we examined the archived holdings on two major topics, crime, and work and organisations, we found precisely one dataset in an electronic form.So researchers still must travel to get the material, photocopy it and either scan it to electronic files or, as with most of Cohen and Taylor's data, type it up from handwritten originals.Material does not appear to be organised in any standard way, either.
Qualitative researchers emphasise fieldwork relationships and the importance of taking account of the context in which data were collected.This context is not directly available in secondary analysis, except to the original researchers.Qualidata's guidelines (1997) suggest contextual information is especially important in longitudinal studies where rapport has developed between researchers and research subjects.We therefore searched for material that might help recover aspects of the context, and found the correspondence between the researchers and the research council, Home Office, and prison reform groups, valuable in this respect.Also, the fact that Cohen and Taylor took comments from prisoners on drafts should be taken into account.This makes selection and evaluation of data different from a situation where the researcher monopolises these processes.We do not know to what extent the researchers took note of inmate comments, although we do know from the records that disagreements were played out in ongoing correspondence.Nevertheless this procedure is a merit of the original methodology and we would venture that where an interpretation is attributed to it, points contested by subsequent researchers would require a higher standard of evidence.We agree with Mauthner et al. (1998) that 'data are the product of the reflexive relationship between researcher and researched, constrained and informed by biographical, historical, political, theoretical and epistemological contingencies' (1998:742), but we do not accept that attempts to go beyond using archived data for methodological exploration are 'incompatible with an interpretive and reflexive epistemology' (ibid.:743).Researchers are used to tracing the mediating effects of reflexivity in primary data analysis and we believe that the recovery of contextual features in secondary data analysis is a practical rather than an epistemological matter.Heaton (1998) suggests that in the light of the complexity of secondary analysis, research design and methods issues should be fully reported, including an outline of the original study and data collection procedures, a description of the processes involved in categorising and summarising the data for secondary analysis, and an account of how methodological and ethical considerations were addressed.We have already commented on the first and third point.Regarding the second, when we went to the Institute of Criminology at Cambridge, where these materials are archived, we found them roughly classified into seventeen different files.The material was in random order within each file, with no indexing of individual documents.We sorted this into data from the original study; correspondence; published documents relating to prisoners' unions, prison magazines and prison reform groups; reviews of the book and other press clippings.Of Cohen and Taylor's four methods there was no record of the classroom discussions (Stan Cohen confirms he kept detailed records of the meetings, which he describes as akin to today's focus groups, but these materials are not in the archive).There were some notes of the literature discussion method and selected correspondence with prisoners.This included letters written after the book's publication.We have used these only where the prisoner was in the original sample.All the data retains the real names of the prisoners, some of whom remain notorious.About a fifth of the completed surveys survive.We word-processed the material we classified as data, including the open response questions from the survey, correspondence from prisoners, and researchers' notes, along with all the data extracts that appeared in the book, and imported it into WinMAX.We analysed the closed response survey questions using SPSS.
Analysis focused on data relating to respondents' past lives, criminal careers and the crime for which they were imprisoned.While we used qualitative software to support our re-analysis, this does not mean the analytic procedure differed fundamentally from that which Cohen and Taylor employed.Like others, we used the software to manage the data, assist assignment (and revision) of codes to segments, and facilitate retrieval of coded segments (for an account of user strategies for qualitative software, see Fielding and Lee 1998).Analysis had still to be done by inspecting the results of given retrievals.The process of assessing the strength of data would not have differed from that used by Cohen and Taylor or most other qualitative researchers, either then or now.
However, there is a significant way in which software can enable analyses which take account of the complexities of qualitative data.By making the coding process mechanically easy (most packages support code assignment by 'drag and drop' functionality), researchers are encouraged to code all the data.Further, code revision is easy; changes can readily be made to the code assigned to single or multiple segments, to the codename itself, and so on.Reducing time required for mechanical aspects of coding frees it for more sustained work with retrieved data.Perhaps more importantly for the present discussion, software may also encourage researchers to validate the analysis against the complete corpus of data rather than to move relatively early to intensive analysis of selected sub-sets.It also facilitates assessment of the evidence for particular interpretations by providing content analysis facilities (for example, indicating the proportion of the data to which given codes were assigned).
These facilities remain tools to aid analytic decisions which must still be made by researchers.Nor can software eliminate the danger that researchers committed to a given perspective could simply ignore data contradicting it.However, using software can make analytic procedures more 'transparent' and reviewable.It allows aspects of analysis to be repeated and demonstrated; for example, through 'audit trail' facilities which go some way towards displaying how interpretations emerged in a series of moves around the dataset and coding scheme.We think such facilities will increasingly play a role in debate inspired by secondary analysis of qualitative data.

Psychological Survival: A Re-analysis
In looking again at this research we emphasise that we do not want to suggest that Cohen and Taylor somehow got it 'wrong'.A body of data can often support more than one analytic theme.Further, as Hammersley (1997) argues, the move away from epistemological foundationalism in social science means that the criteria for determining validity are contested.Secondary analysis is less a matter of proving an analysis 'right' or 'wrong' than of identifying what themes it has not explored.Since we wanted to see if the data supported additional themes, rather than challenge the interpretation that was made of particular data, the ontological status of the data was not in itself material to our analysis.Psychological Survival remains important in policy and conceptual terms and it is a mark of the richness of the original data that it can support further analytic burdens.
Cohen and Taylor's funding application for follow-up research comments that there were some unavoidable deficiencies which make further investigation desirable.These are: a) Size of sample (this was limited to the population of Durham E Wing); b) Nature of sample (only Category A prisoners were included from one wing); c) Duration of study (only 4-5 periods out of total sentences of up to 20 years were studied); d) Limitations upon access to prisoners and upon methods of data collection (visits were irregular and tape recording facilities were not available).
It seems that, despite their critique of standard methods, they came to accept several of the official criticisms of their research.The proposal says the original research 'demonstrated a link between generalised modes of adaptation (to the prison regime) and the criminal careers of the prisoners'.As we noted earlier, what it actually did was to theorise modes of adaptation based on examples from American research, and prison novels and memoirs.The researchers presented little data of their own concerning criminal careers.For example, they say this about 'traditional professional career thieves': These were men who led relatively respectable lives outside, running shops in the suburbs, maintaining good relationships with their neighbours, but who occasionally got together with a group of specialists to carry out a major job.They had a respect for the law which was quite different to that held by those who had confronted authority or attempted to buy it off.Their adaptation to prison tended to be fairly passive.(Cohen and Taylor 1972:156) As a description of a principal sub-group of the sample this summary of characteristics is at a high level of abstraction.We are not told how many of this group there were in E Wing or in the research sample.The terminology of the last sentence is weak: 'tended to be' and 'fairly passive'.
Similar observations were made in a review of the book: there are relatively few quotations from prisoners and nowhere are they balanced or compared the one with the other.Instead, at crucial points in the text, where one seeks for really substantive evidence, quotations are given from the work of other authors in different situations.
The review was by the head of training at the Prison Officers Staff College and was published in the Prison Service Journal.It was easy to dismiss criticism from such quarters, but the authors' response in a letter attacking the review would nowadays be taken as a rather unconvincing assertion of 'ethnographic authority'.They simply suggest they know more than the reviewer about what constitutes proper methodology.
To establish a link between mode of adaptation and criminal career, we should expect a detailed account of individuals exemplifying each category with information about their criminal career, and how they related to the prison regime.It is increasingly accepted that criminal careers do not display a monolithic trajectory; there are transition points where individuals may 'withdraw' or 'go further in', in which prison experience may feature (Sampson and Laub 1993).One question is, then, whether the archived data would have permitted an analysis which balanced the theme of prisoner resistance with a theme of prisoner rehabilitation and change.Among the items in the prisoners survey was the following sequence: (10) Do you feel your sentence was just?(11) (providing admits guilt:) Do you feel that if you had known that the possible punishment for your actions had been more severe or less severe, that it might have deterred you, or influence [sic] your action in any other way?These questions do relate to the circumstances in which the prisoners' crimes were committed and would be likely to elicit data about a criminal career and the prisoner's attitude towards it.In fact, 47 per cent responded 'yes' to Question 10, their sentence was deserved.
There were two parts to the survey data: responses from category A prisoners at one prison, of which seventeen completed records survive, and 'brief facts' about a further twenty-three respondents in other maximum security prisons.Here is an example of the latter: [Name] serving 20 years at [prison].Served --? Detained in security wing since 1968.Marriage broken under Category A restrictions.Is constantly in state of acute anguish over disruption and suppression of mail.Has child whose welfare is constant source of anxiety, and letters referring to child's welfare are often suppressed on pretexts [the blank and question mark following 'Served' are in the original].
Notice the attention given to the man's domestic situation.This is a feature of many of the descriptions: fourteen of the twenty-three comment on it.This data was made available to Cohen and Taylor by prisoners, who designed and administered the research instrument and produced a document summarising the findings.In other words, the prisoners thought it was particularly important to note the men's family situation, the strongest link to their past.One description relates this to the man's behaviour inside: 'Is constantly put into state of agitation (to the point of violence) by the domestic complications caused by disruption of communication.Said situation was damaging wife's health'.
Cohen and Taylor used the data relating to the family largely to criticise the way the prison authorities handled visits and correspondence.Another theme can be supported: family ties present alternatives to criminal lifestyles.'I receive regular visits from my parents now.This tenuous link with the outside and the past means a great deal to me.It always comes as something of a surprise to re-discover that there is an alternative system of values and an alternative way of life' (pre-publication letters).One respondent suggested the prison should run groups to discuss personal and family problems.A prisoner who expressed oppositional values and blamed the class system for many social ills asked the interviewer whether his imprisonment would affect his daughter's chances of attending university, suggesting that even embittered inmates remained oriented to conventional ways of 'getting ahead' and securing a good future for relatives.One prisoner in regular correspondence with Cohen observed that 'by imprisoning a felon, society must necessarily detach him from what is often the only socially conforming role that he voluntarily commits himself to, the family role.After all, most convicts show rather more responsibility towards their family than they do society'.This was part of a lengthy commentary on the importance of family ties, and how the authorities could use them more effectively in rehabilitation.After publication the prisoner commented on why the authors had not developed this argument in the book: 'You do not examine pressures on marriage or the strategies available to deal with them in anything like sufficient depth considering the importance this has for at least half the men in long-term imprisonment.Obviously it was an area you could not research very well as it is not really amenable to group discussion' (post-publication letters).This seems a plausible explanation of the many comments about family ties in the survey data, using the research instrument designed by prisoners, compared to the group discussion data Cohen and Taylor drew on for most of the book.
Cohen and Taylor were very much concerned with the 'present' of their sample, but the data includes numerous references to the men's past.These contain material relevant to several theories of criminal aetiology.A prisoner taking a correspondence course wrote: 'I can honestly assert that I have learnt and understood more about history in the last month than I did in all the years at school.I don't hold my teachers responsible for my ignorance, they really tried, but in retrospect I seemed very determined not to learn' (pre-publication letters).Similarly, an inmate comments on his friendship with a lady pen pal, who now visited him in prison: 'It's a strong relationship we have.She needs my letters as much as I need hers.They have managed to usurp the larceny in my heart, and I talk to her on paper as I've talked to no-one in my life' (postpublication letters).Such testimony would be significant, for example, in theo-ries emphasising the importance of childhood ties to authority figures and conventional others, such as Hirschi's control theory (1969).
There are also more straightforward comments repudiating past involvement in crime, such as this survey response: 'Past life of "glory" (?) and things like that, now I just don't want to know' (response to survey questionnaire).Cohen went to work in the United States for a time, where an inmate sent him a letter (included in the pre-publication letters) saying: I can understand America's attraction for you; my own inclination is very different.America has for me become associated with violence, which I find more and more repulsive (I'm not saying this for the censor's benefit).It's not just being involved in violence but even watching it on television.I don't suppose this abhorrence, which extends even to the hunting of animals, is going to be part of my make-up permanently.The drift to pacifism will no doubt go into reverse, once my mind has managed to rationalise violence.
The ambivalence of the comment suggests authenticity: one would certainly not write the second part if one was trying to impress the prison authorities byway of the censor.It also suggests the sense of contingency analysed by Matza (1964) in his work on delinquency and 'drift', and something quite different to a purely oppositional stance.
Another respondent wanted to see more questions in the survey about communications with the outside world and about the crime the person was imprisoned for.This relates to the expression of contrition, or at least regret, for what the prisoner had done.Such expressions were absent from the book, but can be found in the data.An extract from one letter comments, 'emotionally one is living in a desert, without even the love-substitute of the cloistered religious.Still, I mustn't grumble, I'm not in here for nothing.I try to keep that in the forefront of my mind but its not always easy to retain one's objectivity' (prepublication letters).A survey respondent expresses particular regret for the impact on his family: 'I realise that I have failed my family and my children'.Another takes contrition to an extreme: '[The] sentence for murder is much too lenient.Life imprison for ever, never out [should be the punishment].Yes, that goes for myself.I don't mind if I die in prison' (response to survey questionnaire).Another murderer said 'prison life is too easy for criminals.I expected to get kicked around.It encourages them to be criminals if its too easy' (response to survey questionnaire).A sex offender felt, as did 47 per cent of respondents overall, that his offence required treatment rather than punishment: 'Although he declared he did not need rehabilitation as he was not a criminal, he also said that if there was a treatment available for his crime he would benefit from it' (interviewer's note on survey questionnaire).
To say that expressions of remorse and willingness to change were present in the data is not to say they should be taken at face value.However, comments laying claim to such feelings are arguably as authentic as the data supporting the resistance theme which Cohen and Taylor developed, being extracted from the same data sources.Some of the data suggest men engaged in confronting their sense of self identity: I think that seeing ourselves ‚in time' must be one of the main arguments for the study of history.It makes us aware of our ephemeral nature, and perhaps gives us a fresh perspective.I tell myself, sometimes unsuccessfully, that my sentence is only the blinking of an eye in the cosmological passage of time.In December I shall have served eight years.
The prisoner also reveals insight into his own difficult character in a comment that moves between sophistication and a more down-to-earth sentiment: 'I am, I'm told, as contrary as ever.I reply to this charge that only by balancing thesis with antithesis can the synthesis of truth emerge.Of course, it could be that I'm just bloody-minded!' (pre-publication letter).To be self-critical in this way is not a mark of someone who has single-mindedly taken on cultural values of resistance to all authority.Likewise, this survey response suggests acceptance of the need to change before re-entering society: rehabilitation is about helping me to become a better person to go outside.To give me training and to give me help with personal problems.They should encourage instead of discouraging all kinds of handicraft that could be sold through prison authorities or otherwise, to help financially with dependants outside.I feel this would help to make men realise and accept their personal responsibilities.Some respondents were suspicious of glib claims to have changed, whereas 'rehabilitation means a real change in a person'.Another said simply 'the prison authorities help me to look at things, so I will be alright when I go back [home]' and another accepted that attitude change was necessary: 'rehabilitation is to change a man's views.To turn him into a fit and able person for society'.Another reports that prison has brought about 'a real change for the better.I am a different person.My attitudes have changed'.Such strong claims at rehabilitation are hard to accommodate in an analysis based on opposition and resistance.That prisoners are grappling with the meaning of their crimes is suggested by this respondent who was a member of a Middle Eastern elite until he killed his wife and attempted suicide, which left him blind: 'rehabilitation means education, if needed.Social education.People don't understand their crimes or understand social ethics.They do not understand social ills' (response to survey questionnaire).In fact, over half the survey respondents said rehabilitation would affect their behaviour (see Table 2).Some 65 per cent reported requesting help from the prison authorities with training needs, 59 per cent had requested help with domestic problems, and 59 per cent had requested help with education.While 47 per cent reported they had not received the help they wanted, 29 per cent had, and a further 18 per cent received partial assistance.These are not data suggesting an inmate group standing aloof from the authorities and the support they can offer.Criminologists are not alone in seeking to capture the prisoners' response to imprisonment using a heuristic based on rehabilitation and resistance.A survey respondent observed that 'those who are dominant in the prison are the sort of prisoners who fight the system.Because of this it gives them a certain influence over those who accept the situation'.The following extract from a postpublication letter also suggests an understanding of psychological mechanisms behind affiliation to prison culture: Prisoners, if they are to retain any integrity of self, must reject the label of being wicked and evil.They institutionalise the rejection within their prison culture; after all, it is bad enough being punished without having to accept the guilt of being wicked and evil as well.The prison culture thus serves the needs of its upholders: it defines prisoners as superior to their jailers, it demonstrates to the prisoner that he has been unfairly punished, and generally enhances the status and morals of the convict by a wholesale rejection of society at large.Some prisoners were helped to see things with such insight and detachment by the awareness they gained in Cohen and Taylor's classes.Again it is a matter of individuals receiving attention in a way no one had tried before, as this extract from a pre-publication letter shows: I have become a completely different and I think better person.Not only have I absorbed factual information, albeit incompletely, but I have also adopted to a large extent the social philosophy which you all [tutors] tend to purvey, in a gentle, undogmatic way.Having undergone what for me is an important political and ethical transition, it would be fatuous to deny that I was greatly influenced by you all, but I feel that I was not totally blind or uncritical in accepting many of the things that I have come to believe.If my beliefs are different and, dare I say more sophisticated, it is not simply because I think you [are] all friendly, amusing, liberal people but because with great patience things were explained to me in a way that they were never explained before.
A similar sense of being taken seriously in contrast to earlier experiences of education is conveyed by this remark: 'the classes at Durham gave me a great deal of confidence.One is always afraid of making a fool of oneself, especially with people one regards as social betters.Once I got to know you all I was not afraid to put my foot in it' (pre-publication letter).As tutors, Cohen and Taylor gave the men the chance to explore their prison experience without worrying about being evaluated.It is ironic that, to make their critical case against the maximum security regime, and emphasise the resistance elements of the culture, Cohen and Taylor downplayed the reformative impact of their own work with prisoners.
A further irony is that this work encouraged the prisoners to challenge Cohen and Taylor's analysis.After publication they received letters from prisoners criticising aspects of the book, some in great detail.We have already noted criticisms by reviewers connected with the prison system, but it is worth noting that prisoners themselves felt the book was unfair to prison staff: 'I have got reservations about your general attitude towards staff.It is almost as if you are scoring a few points for the "convicts".If this were a manual for selfsurvival you surely would have attempted to offer more insight into the attitudinal pressures of the warder's role ... [To refer to] "the dull peasant mentality of their provincial guardians" ... is very counter-productive' (post-publication letter).The prisoner maintained there were many prison officers who were as critical of the regime as were the prisoners, and while they would be sympathetic to the book's general analysis, they could be put off trying to change things if they felt abused by the book's portrayal of them.Some 35 per cent of survey respondents reported they co-operated with prison officers and a further 35 per cent said they sometimes did, leaving 29 per cent who did not.All but one accepted that prison officers had a legitimate job to do.
Another criticism by prisoners was that Cohen and Taylor were too quick to take prisoners' testimony at face value.The book reported an alleged violent separation of a prisoner from his wife and young daughter when they broke rules during a prison visit, but a post-publication letter warned: Be careful with prisoners' 'atrocity tales' ... Wives are not mucked about and children were allowed to sit on their father's laps at the time you were conducting your research ... Prisoners can kiss and cuddle wives at the beginning and end of visits.They can also hold hands while the visit is on.
A number of letters made similar points about incidents described in the book.
A final point about the researcher/respondent relationship is the dark suspicion raised in one post-publication letter: 'I've often wondered if my writing to you was one of the reasons why I'm still on the [Category] 'A list... Unless I'm being kept on as a form of punishment I can't think of another reason for it'.It is doubtful such correspondence would be more important than the original crime in keeping a man in maximum security, but the letter does raise the issue of a researcher's responsibility towards research subjects.As we have seen, Cohen and Taylor discharged this responsibility by campaigning to raise awareness of the circumstances of prisoners in the maximum security system, but we hope also to have shown that, in doing so, like all research, they told only part of the story.Important as adaptation through resistance is, it exists in tension with adaptations based on personal change, a tension manifest at the level of individuals as well as the level of culture.Re-visiting the data from this 92 groundbreaking study of long-term prisoners restores to the account themes which suggest less polarised, more ambivalent criminal careers (Tracy and Kempf-Leonard 1996).Secondary analysis is not just another means for researchers to engage in methodological reflection but can be an important way to exploit more fully research on sensitive topics and hard-to-reach populations.

Coda
Secondary analysis of qualitative research can contribute to a sense of collective and incremental endeavour in a field which is sometimes criticised for lacking these characteristics.So it has proved with this article.As noted in the Acknowledgements, we have benefited from the comments of Stan Cohen on our re-analysis of data from the research, and points he has made have been incorporated into the article.One point deserves greater prominence, in the light of its centrality both to the original analysis and our own effort.Stan Cohen writes that he believes our article, and earlier commentaries, somewhat exaggerate the degree to which he and Laurie Taylor saw the prisoners as engaged in a heroic political struggle.In that we have suggested an analytic balance needs to be struck between such 'resistance' and other adaptations, it is important to note Cohen's observation that he and Taylor 'certainly don't see all the prisoners as anything like this' (personal communication, 10 April 1999) and that their analytic stance was signalled by the decision to title the book' Psychological Survival' rather than 'Psychological Resistance'.Cohen would also refer readers to the later book Escape Attempts: The Theory and Practice of Resistance to Everyday Life (1976) for an elaboration of their position.For our part, we are content that secondary analysis has provided a means to extract further analytic purchase from research on a group seldom exposed to fieldwork.Those of a postmodern cast of mind may also find in this coda a sense of the polyvocality that is implicit in any secondary analysis of qualitative data.

Table 1 :
Records of 138 Category A Prisoners: List of the Main Charges of which the Prisoner was found Guilty on the Occasion of his Most Recent Conviction ** Grievous bodily harm, wounding, attempted murder.Source: 'The regime for long-term prisoners in conditions of maximum security', Report of the Advisory Council on the Penal System, London:HMSO, 1968.