RISK OF CRIME AND FEAR OF CRIME:
The Politics of Victimisation Studies
Criminologists work in a terrain fraught with difficulties. On the first and most immediate level the crime rate seems unconducive to human intervention. Nothing seems to work: each prized innovation, from community service to neighbourhood watch, seems to have little effect. Indeed, the only countries, outside of Japan, which can boast a declining crime rate would seem to owe their good fortune to demography not criminology: as the proportion of young males in their population temporarily declines. The second is that we write, as it were, in the middle of moral panic. The mass media portrayal of crime is extraordinary in its level of exaggeration: muggers fill the inner cities of the TV sets, rampant child abuse and serial murder headline our newspapers. Lastly, we have, until recently, been bereft of decent statistics: the dark figures of crime unknown to the police is variously estimated. To base criminological theory, or social policy for that matter, on the majority of official figures is an exercise in "guesstimates", and tealeaf gazing. Meanwhile, various groups with special pleadings regularly, and understandably, parade their 'statistics' to show that their section of the community needs resources or their agency has had such and such a success rate.
Criminal victimisation studies are a useful research instrument to deal with the problem of inadequate statistics and to more accurately pinpoint problems within society. Commencing on a large scale in the United States of the 1960's they reached Britain by the late seventies and have resulted in a series of British Crime Surveys (BCS) (see Hough and Mayhew, this volume). For a while it seemed that the problem of the dark figure of crime would be tackled. Indeed, Richard Sparks and his associates, in the introduction to their pioneering British victimisation study, summarised the decade of American research prior to their own with a note of jubilation: "Within a decade .... some of the oldest problems of criminology have come at last within reach of a solution." (1977, p.1).
As I will make clear in this article, I have little doubt that victimisation research represents a major advance in the techniques available to both criminology and the policy sciences? But there remains many problems which are only too easily skated over. Some of these stem from the nature of the mass survey per se, some from an over-eagerness to move from raw data to the computer keyboard, and some from a very weak analysis of precisely what questions such studies are asking.
A common and understandable tendency amongst criminologists, police officers, and other practitioners to the widespread fear of crime and the regular panics of the mass media is what we might term 'putting crime in perspective'. Thus it is often suggested that crime, although frequent, is a relatively minor irritant, given the range of problems with which the city dweller has to contend. The public, it is argued, suffer from a hysteria about crime fanned up by the newspapers and television. Moral panic abounds, particularly about mugging, sexual assault and violence which are out of touch with reality. People lock themselves in their homes because of their own irrational fears and the fear of crime becomes more of a problem than crime itself. Such an argument is backed up by evidence from sources such as the British Crime Survey, which shows that the 'average' person can expect "a robbery once every five centuries, an assault resulting in injury (even if slight) once every century, .... a burglary every 40 years .... and a very low rate for rape and other sexual offences" (Hough and Mayhew, 1983, p.15).
The "irrationality" of the public is demonstrated by tabulating real risks of crime against fears. Two groups, the elderly and women, are seen to be particularly disproportionately worried. Thus the BCS produces the following table which is very akin to those produced earlier in the US National Crime Survey and its at the centre of the fear of crime debate.
Table 1 BRITISH CRIME SURVEY: FIRST SWEEP
Fears for Personal Safety After Dark and Risks of 'Street Crime'
Question: "How safe do you feel walking alone in this area after dark?" (Ibid, p.25)
Furthermore, most crime is petty and the increase in crime is more of an epiphenomena of decreased tolerance of the public than anything else. Thus the authors of the BCS continue:
"The real message of the BCS is that it calls into question assumptions about crime upon which people's concern is founded. It emphasises the petty nature of most law-breaking - a point which also emerges from Criminal Statistics, but which is often overlooked. In showing that many crimes go unreported to the police or unrecorded by them, the survey also demonstrates the extensive scope for error when drawing conclusions about crime trends from statistics of recorded offences. Thus, the survey lends credibility to explanations of rising crime which have been dismissed in the past - that, for example, people's tolerance of petty crime may have declined, leading to increased reporting to the police: or that additional police resources and greater efficiency in recording practice have led to increased recording of crime.
Those incidents which go unreported usually do so for a very good reason: victims judge them too trivial to justify calling in the police. Of the offences uncovered by the BCS only a tiny proportion were crimes of serious violence, and very few were serious property crimes such as burglary or car theft. The vast majority were, for example, petty thefts, acts of Vandalism, and minor assaults. A corollary of this is that the risks which people face of being victims of serious crime are remarkably small.
It is far harder to convey these points convincingly than it is to talk in generalities about soaring crime rates, a breakdown in law and order, and the like. Serious crimes are cause for legitimate concern, however rare they might be. But these crimes - rape, serious wounding and robbery, for example - are a small minority of the total. The public should have a balanced picture of crime - especially in view of the likely consequences of sensational presentation: excessive anxiety about crime not only impoverishes people's lives, but also makes it difficult to secure rational discussion of criminal policy." (Ibid, pp-33-34).
Thus we have a related notion of crime as an exaggerated problem, of the rareness of serious crime, and of the irrationality of the fears of a sizeable section of the population. This is related to very definite policing conclusions:
"The revelation of so much more crime than is recorded in Criminal Statistics might suggest the need for further increases in police manpower. But it is doubtful whether the police could do much in respect of crime that is not reported to them; and while people might be encouraged to report more crime, it is debatable whether the police should be called in over each and every breach of the criminal law. Moreover, a substantial body of research indicates that it is difficult to enhance the police effect on crime. In particular, it is becoming clear that the effectiveness of the 'core' of policing - preventive patrol and criminal investigation - cannot be significantly improved by increased manning levels. For many sorts of crimes, people themselves might take more effective preventive action, either acting individually or together with others. The police could do more to promote preventive action of this kind, while the trend towards putting more officers on the beat may have the desirable effect of reducing fear of crime." (Ibid, p.34).
And this completes a package: which maintains that much crime - as the last Commissioner of Police for the Metropolis took constant pains to remind us - is 'opportunistic' (Newman, 1984, 1985, 1986), police action could scarcely deal with it, serious crime is rare - and has a high clear up rate anyway - and, finally, the role of the public is seen as central to crime control, whether it is in more locks and bolts, neighbourhood watch or victim support. (See Kinsey et al, 1986, Ch.4).
I would like to criticise this position, not from one which is dismissive of victimisation research, or, indeed, the very real achievements of the BCS but which attempts to point to new directions, as well as persisting problems in such work. 1 do this on the basis of the findings and problems thrown up by a series of local victimisation studies conducted in Britain in the last few years. (e.g. Kinsey, 1984, Hanmer and Saunders, 1984; Hall, 1985; Jones et al, 1986; Kinsey et al, 1986; Jones et al, 1987), particularly the Islington Crime Survey (ICS) and the Broadwater Farm Survey (BWFS).
The Dark Figure
The size of the dark figure of crime in the ICS was about 50% of ' total - that is only one half of crimes ever reported to police. of crimes were reported to the police by the public and only 3.8% directly detected by the, police.
Table 2. Major Reasons for Victim Non-Reporting to Police
If we look at Table 2 we can see how the dark figure was constituted, both nationally (BCS) and in the local survey (ICS). Quite correctly Hough and Mayhew have pointed to a large part of the dark figure, not only as being objectively trivial crime, but as being seen by the victims as such. But we must note that half of the unknown figure is not seen as trivial, and this rises to three quarters in an inner city area such as the London borough of Islington. It should be-noted that there is also a substantial section of the population who view their victimisations as non-trivial, yet perhaps realistically given the clear up rates in the inner city - see the police as unable to do anything. A large number of victimisations, therefore, genuinely belong to the dark figure as defined by the victims and are seen as a matter for the police. The dark figure, particularly in the inner city, does not shrink much.
But let us not leave the 'trivial' offences at this stage. There are many instances of anti-social behaviour which are, by and large, not within the scope of the criminal law: they are too trivial for any court room. Petty vandalism is one example, and much sexual and racial harassment another. But the accumulation of such incivilities can make people's lives a misery more than the reported instance of a 'true crime' - say a solitary burglary. In Islington 61% of white women under 24 and 72% of black women had been upset by harassment in the last twelve months. And, as we shall see, this under tow of incivilities can greatly show up in general fear of crime. (See Lea and Young, 1984, pp.54-58).
We should note also the often overlooked fact that victimisation statistics, like police figures, have dark figures - only, smaller. A proportion of questions will not be answered truthfully by respondents out of fear or embarrassment. Hough and Mayhew clearly acknowledge this when they comment on the extremely low figure of offences against women discovered by the BCS: only one rape in the 1982 Survey of England and Wales, and that attempted, whilst only 10% of assaults were classified as domestic. (Ibid, p.21). In contrast, more attempted rapes were found in Islington alone and 22% of assaults were domestic. Moreover, there is, of course, no doubt that these latter figures are underestimates. Even more significantly, a series of feminists studies such as Stanko (1985), Radford (1987) and Hall (1985) have found widespread sexual assault. Undoubtedly a large part of the greater accuracy of these figures relate to using sympathetic women interviewers, although the wording of the questions can make for different findings. There is a world of difference between simply asking the interviewee if she has been raped (as do most conventional surveys), and including a definition of rape in the question 'sexual intercourse without consent', which Hanmer and Saunders do (1984). There is, in short, different dark figures for different questions, and to relate the findings to say, fear of crime, depends on how one evaluates the causal relevance of the question asked. Further, a characteristic of many conventional victimisation studies (including the BCS and ICS) is to ask for victimisation over a period of the last 12 months. This is in order to make comparison with the police statistics. Of course, people's attitudes to crime are built up during their life. In order to capture this 'have ever' questions as used by Ruth Hall in her 1985 study are of great relevance. Many victimisation studies, thus, have a dark figure, not only partially in terms of the present, but also totally in terms of the past.
Up until now I have talked about the dark figure simply in terms of crimes not revealed by the interviewees. But there is a much more simple sense in which victimisation surveys have dark figures: namely in terms of non-response.
Table 3 COMPARATIVE RESPONSE RATES
The non-response rates in all these surveys are considerable, and apart from the Newham Survey (which is simply too low to be relevant), there is a fifth to one quarter of respondents whose victimisation is unknown. It goes without saying that such a large unknown population could easily skew every finding that we victimologists present. At the most obvious level: it probably includes a disproportionate number of transients, of lower working class people hostile to officials with clipboards attempting to ask them about their lives, and of those who are most frightened to answer their door because of fear of crime. Within the population surveyed we find a dark figure of crime equal to that known to the police. We would expect that crime within the unknown population to add at the very least 25% to this figure but because the population unsampled is in all probability, least likely to contact the police, the supplement would undoubtedly be greater than this.
It has been frequently noted that police statistics not only have a dark figure in terms of quantity, they are also a skewed measure qualitatively, because certain sorts of crimes against certain sorts of persons are disproportionally represented. In general property crimes are more likely to be reported than crimes of violence, crimes against high status groups more than against less powerful groups (eg blacks, women, the lower working class, youth), crimes committed by strangers than those committed within the family (see Kinsey et al, 1986). To give a potted contrast: a crime committed by a professional robber against a jeweller's shop is more likely to enter the statistics than domestic violence against a woman who is poverty stricken. We have seen how victimisation studies, like police statistics, have a quantitative dark figure. I would suggest that the qualitative skew is also present. For example: domestic crimes are less likely to be reported to interviewers and, in addition, have, in all probability, a greater frequency amongst those who refuse interview. None of this belittles victimisation studies: they do genuinely intrude into the dark area of crime and, in the case of the local and feminist studies, they reveal much higher levels of crime (particularly violence) against weaker groups. But they do not eliminate the dark area either quantitatively or qualitatively and every indication is that the exposure of these crimes would raise the known level of violence in our society.
The crime rate exposed by victimisation studies and likely to be revealed if we moved further into the dark area is, therefore, of a serious nature: it cannot be swept under the carpet as just less serious crimes to be added to the official statistics. Nor can it be represented as Alphonse Quetelet hazarded in the first part of the Nineteenth Century, as a simple fixed proportion of the official statistics - so that victimisation statistics are merely much of the muchness of official statistics. What victimisation surveys do is begin to shift the focus of crime away from what are traditionally police priorities to those which are the priorities of the wider public. As is well documented on both sides of the Atlantic, police priorities tend to create the dark figure with regards to certain offences and certain victims, which we have already indicated. Thus Smith and Gray (1983) in their important study of the London police note the difference made between 'ordinary people' and 'slags' and between 'real crimes' and 'rubbish' offences. The victimisation study, in the sense that it attempts to encompass all the victims in the community, is more democratic in its brief however imperfect it may be in its accomplishment.
The Focus of Crime
Crime is focussed both geographically in certain areas and socially in certain social groups. Crime figures which add together low and high crime areas are useful in assessing large scale service provision, but tend to obscure the pinpointing of crime within the population. Local crime surveys help to deal with the problem of defining areas of high crime, whilst sufficient high sampling, often with boosters, allow one to break the population down into the major social axes: age, class, race and gender, and what is most important their combination.
The Islington Crime Survey was conducted by the Centre for Criminology in an inner city area of London. Our study showed the substantial impact of crime on the lives of people in the Borough. A full third of all households had been touched by serious crime (eg burglary, serious robbery or sexual assault) in the last twelve months, and crime was rated as a major problem second only to unemployment. Crime shaped people's lives to a remarkable degree; a quarter of all people always avoided going out after dark, specifically because of fear of crime, and 28% felt unsafe in their own homes. As far as women are concerned, there is a virtual curfew of a substantial section of the female population - with over one half of women often or always not going out at dark because of fear of crime. Such a survey puts fear of crime In perspective. It is scarcely odd, for example, that 46% of people should admit to worrying 'a lot' about mugging, given that over 40% of the population actually know someone who has been mugged in the last twelve months, including themselves and their own family. Nor is it unrealistic to worry about burglary when its incidence runs at five times the national average and on some estates four out of five houses have been burgled in the last year.
The advocates of crime as moral panic point to the 'paradox' that women are more fearful about crime than men, although most studies show they have a far less chance of being victimised. Our survey suggests that their fears are perfectly rational. For women are, in fact, more likely to be-victims of crime than men. The reason for the shortfall in past findings is the nature of many of the crimes committed against women and their reluctance to admit them to a stranger engaged in a social survey. By the use of carefully trained researchers who were able to sympathetically conduct interviews we found a considerably higher rate of female victimisation. The reason for this is threefold: sexual assaults are almost exclusively a female 'prerogative' as is domestic violence., whilst additionally street robbery against women is greater than it is against men. Thus, in terms of non-sexual assault alone women are 40% more likely to be attacked than men. Sexual assault in Islington is fourteen times the national average - 20% of women interviewed knew someone who had been sexually assaulted or molested in the previous twelve months, and over 50% had experienced sexual harassment of a non-criminal kind. And all of this occurred in a situation where the women concerned took considerable greater precautions against crime than men. They were, for example, five times more likely to never go out after dark than men, three times more likely to always avoid certain types of people or streets, and, very significantly, six times more likely to always go out with someone else instead of alone. Is it surprising that women fear crime?
In terms of the over forty-fives, we found, as has been widely discussed, that they do indeed have a lower crime rate against them than young people However, we found that when assault did occur they were more likely to be injured - as are women compared to men incidentally - more likely to lose time off work, more likely to have an attack involving severe violence (kicking or use of a weapon), and the attack was more likely to have a greater effect on their lives. None of this supports the paradox of irrationality, often argued about older people and crime.
Crime by Subgroup
Our analysis, by focusing in upon sub groups, was able to precisely illustrate the extraordinary differences in the experiences of crime. For example, if we look at the Assault rates in Table 4 we see that over 45 year olds whites live in what amounts to a totally different universe with regards rime than do younger people. Young, white females are, for example, twenty nine times more likely to be assaulted than those over 45, and thirty times more likely to be sexually attacked. Note also how the most dangerous period for women differs by ethnic group: it is the youngest age group for whites, the 25-44 age group for blacks and the over 45 for Asians.
Table 4: Assault Rates by Age, by Race, by Gender
The above table illustrates the fallacy of talking of the problem of women as a whole, or of men, blacks, whites, youths, etc. A realist criminology must start from the actual subgroups in which people live their lives rather than wide categories which contain with them wide variations. The variations in the category 'white' for example is, in terms of the ratio of the lowest to the highest rates of assaults between subgroups, is 20 times, of blacks 11 times, of men 9 times, of women 29 times, of youth 7 times, and of the older age groups, 13 times.
The Fear of Crime
The problem of the fear of crime is usually formulated as the relationship between high rate (an objective fact) and fear of crime (a subjective attitude). Thus women - or the elderly - are seen to have objectively a comparatively low rate of victimisation and subjectively a high fear of crime. Young men, in contrast, have a very high risk rate and a lower fear. We have seen how, in inner city London, at least, these 'facts' could be disputed, but the data from the US National Crime Survey and the B.C.S. on a natural level point to this distinction. But let us for the sake of argument accept these findings and attempt to interpret them. First of all, what is a rational level of fear? Instead of ascribing irrationality to women would it not be more appropriate to ascribe irrationality to young men? Are their low fears of crime not merely a function of machismo culture which actively encourages fearlessness with regards to violence - and, even where fear exists, would never publicly admit so to an interviewer? Isn't their high death rate from motor car and motor bike accidents another instance of this irrationality? And lastly, would it not be more advisable to attempt to raise the fear of crime of young men rather than to lower that of other parts of the public? We could, for example, point to the very real dangers of knife wounds and grossly under- estimated morbidity from blows to the head, kicking, etc.
Of course, on a strict level of rationality, it is not up to the social scientist to suggest the 'proper' level of fear of crime for the public. The 'rationality' of fear of crime cannot be deduced from the risk rates. What can, of course, be said that, given your level of fear of violence (which is a person or group's own metric), then one is frightened of crime to an extent which is incommensurate with one's fear of, say, car accidents or air travel. To take the base line argument, fear of death is a religious and philosophical matter, it is not the province of social scientists to instruct the public as to their proper level of fear. Rationality can, however, be, so to speak, improved by the social scientist if the given fear of violence of a subgroup is accepted as the latter's own demesne, whilst granting that the lay person is often unclear as to the relative risk rates and dangers of different forms of violence and social encounters.
Let us turn to the other half of the equation, the supposedly 'objective' risks of crime which the mass victimisation studies expose. Most surveys ask the interviewee whether they have had a certain action committed against them and then subsequently ascribe to the victim report criminal status or not. They do not ask the public to define whether the act is or is not criminal (see the useful discussion in Hough and Mayhew, 1985 pp.4-5). This, as Hough and Mayhew correctly note, is of less a problem with regards to crimes such as burglary, where there is little public controversy, but is "much muddier for offences such as assault".
If one takes subgroups with very different definitions of assault, a sharp contrast is between young men and almost everyone else. I want to suggest that to ask seemingly objective questions on such matters as assault will generate a large number of positive answers, even although young men themselves will simply not consider many of these assaults as significant. A large number of incidents will merely be horseplay - of no significance whatever, either to the young men themselves or to anyone else for that matter. This is not to suggest that serious violence is not focussed amongst, say, lower working class young men. Rather it is to suggest that their violence figures are greatly inflated by crimes of little social significance. Furthermore, that this is a result of attempting to produce objective figures about a phenomenon crime which, of necessity, involves a subjective dimension.
The exercise, then, in relating an objective crime rate to a subjective level of fear is, from a realist perspective, flawed, because it:
i) assumes that rationality would involve each subgroup of society having a fear of crime rate proportional to their risk rate;
ii) it assumes that there is an objective crime rate irrespective of the subjective assessment of various subgroups.
What I want to suggest is that by attempting to set up an objective - legalistic - definition of crime independent of the subjective definitions of the various subgroups within society, victimisation research commonly trivialises that which is important and makes important that which is trivial. As we have seen any meaningful concept of crime rate must include the nation of human evaluation. What seems trivial to some would be serious to others, what might be serious to some is trivial to others. Men and women, in particular, may have different evaluation of what is serious and what is trivial. This is not to suggest that there is not a consensus over serious crime. There is, but the existence of disagreements at the margins must incline us to re-interpret; in particular to avoid reading off risks rates as simply a reflection of raw data. Further, I believe that this approach helps us understand other anomalies which have arisen within the literature of victimisation. For example, 'the education effect': where rates increase with years of education (see Sparks, 1981).
Up till now we have referred to the way in which both police statistics and victimisation studies tend to conceal certain sorts of crimes against certain sorts of people. We have seen how global figures occlude the geographical and social focus of crime and we have noted the probability of a systematic value bias in the statistics. But none of this has gone beyond the notion of risks rates as acting upon equal victims. It is, however, only when we place victims in their material context that we find the patterning of worries about crime becoming more understandable, just as is true of the different levels of tolerance of crime and violence.
The Myth of the Equal Victim
In What is to be Done About Law and Order, John Lea and I discussed the myth of the equal victim. Namely, that by discussing the notion of the impact of crime in terms of risk rates abstracted from the general material predicament of the victim we come to a totally false assessment of impact. Important here is the process of:
i) compounding: people who are victims of one crime tend to be victims of others;.people who are victims of one social problem tend to have other social problems laid on their doorsteps. (Hazel Genn in the book quite correctly identifies the process of multiple victimisation);
ii) differential vulnerability: people differ greatly in their ability to withstand crime. The yuppie may experience 'the positive burglary' which involves a creative insurance claim and a new stereo, a poor isolated elderly woman will suffer much more grievous harm (see Maguire, 1980). And, of course, the service provision which police, local council, victim support, etc provides varies similarly.
What a Realist method must do is move from abstract crime rates to the concrete predicaments which-people actually face. But let us take one move step forward and talk about the relationship of crime.
The Meaning of a Punch
If we take an objective approach to assault we can imagine a punch delivered with a given velocity and causing a certain level of bruising. We could then draw up tables which would show how such assaults were differentially distributed across the population, and we would relate this, perhaps, to fear of crime. Of course, something of this sort already occurs in present victimisation studies, although the level of objectivity is scarcely as exacting. The problem with this approach is that the 'same' punch can mean totally different things in different circumstances: it can be the punch between two adolescent boys - of absolutely no significance on the level of victimisation. It can be the punch of a policeman on a picket line or the punch of the picket against the police. 'It can be the drawn-out aggression of a violent man towards his wife. It can be the sickening violence of a parent against a small child.
Violence, like all forms of crime, is a social relationship. It is rarely random: it inevitably involves particular social meanings and occurs in particular hierarchies of power. Its impact, likewise, is predicated on the relationship it occurs within. We should continue to create our tables of victimisation but we would be wrong to believe that we can have a science of victimology which ignores the offender. For the very impact of the offence depends on the relationship between victim and offender.
Women and the Fear of Crime
Let us bring all the factors we have discussed together using the victimisation of women as an example. I have argued that the figures conceal crime risk rates and that the impact of crime is a function of risk, compounding, vulnerability and relationship. Let me expand briefly on this:
1. Concealment: The Invisible Victim Domestic crisis and sexual crimes are less likely to enter the statistics than property crimes, which leads to the systematic underestimation of crimes against women. The particular focus of crime on certain categories of women is concealed by global figures. The actual impact of known crime on women is underplayed by designating much of their victimisation as trivial.
2. Compounding: Women do not only suffer crime per se but also an undertow of incivilities and harassment which men do not experience. The impact of crime on women cannot be assessed without taking into account these incivilities.
3. Vulnerability: The relatively powerless situation of women: economically, socially and physically makes them more unequal victims than men.
4. Relationships: Crime is a relationship: crimes against women, and as indicated fifteen years of feminist research has indicated, crime is about patriarchy. Crime in the home occurs within a relationship of economic dependency: the woman - particularly if she has children - cannot walk away. It occurs also within an emotional bond which gives it all the more hurtful poignancy. Crime and incivilities against women in the streets reflect the overbearing nature of particular values. What a dramatic indictment of our civilisation it is that, in the inner cities of Europe, men can quite happily-walk the streets at night, yet a huge section of women are curfewed because of fear of crime.
It is easy to see then how crime has a greater impact on woman as well as, at the same time, women are more sensitive to violence. For in the last analysis many women react to the adversity of the world by creating a female culture which is opposed to violence, whilst men frequently react to adversity by creating a culture of machismo which is insensitive to violence and, indeed, in some groups glorifies it. Conclusion I have attempted a critique of a particular tendency within criminology which tends to underestimate the problem of crime and which tends to view fear of crime as more or less irrationally detached from risk rates. To do this I have applied a realist method which seeks to place crime and victimisation in its specific material circumstances and which underlines that crime involves both action and censure: behaviour and values. (See J Young, 1988). I have not had space to examine the policing implications of this position rather than to comment that this analysis suggests that present policing is woefully inadequate in dealing with crime, both in the public and the private arena.