Foreword | Acknowledgements | Tables & Figures | Executive Summary | Introduction | Levels of knowledge about crime and the criminal justice system | Attitudes to criminal justice professionals | Crime seriousness | Sentencing practice | Aims of sentencing | General discussion/concluding remarks | References | Appendices
2.1 Crime trends
2.2 Proportion of crime that is violent crime
2.3 Rate of household burglary
2.4 Cost of prison
2.5 Length of prison sentences
2.6 Reconviction rates among offenders
2.7 Offending while on bail
2.8 Respondents who were the best and least informed
2.9 Summary
Why should we be concerned about what the public knows about crime and the criminal justice system? Firstly, if public opinion is to influence criminal justice policy, then it is important to know the extent to which public perceptions are based on accurate knowledge. Secondly, it has been suggested that the effectiveness of the criminal justice system is dependent on public knowledge of the system. For example, people can only be deterred from offending if they are aware of what penalties may follow a conviction.
However, research on what the public knows about crime and the criminal justice system is an area that has generally been neglected by social scientists, who have tended to focus almost exclusively on public attitudes (Roberts & Stalans 1997). If, as Roberts (1992) suggests, the public subscribes to a crime-control model of criminal justice, then exaggerated beliefs about rising crime rates would foster the view that the criminal justice system has failed in its primary function, to reduce, or at least slow, the rate of offending.
Given the importance of placing the public’s attitudes to the criminal justice system in the context of their knowledge, one section of our survey concentrated on asking respondents a series of questions regarding their knowledge of crime-related facts and statistics. These questions focussed on crime rates, the cost of prison, the sentencing process and rates of re-conviction for some offences.
Surveys in the United States, Canada and Australia have consistently found that the public has an inaccurate and negative view of crime statistics, including the general beliefs that the crime rate is rising dramatically and that a high percentage of crime involves violence (Roberts 1992; Roberts & Stalans 1997). Similar findings have also been reported in Britain (Hough & Roberts 1998).
In our first question, we asked respondents whether the amount of crime reported to the police over the country as a whole had changed during the last two years (i.e. 1996–1998). This question was modelled on one asked of the British public in the British crime survey (Hough & Roberts 1998; Mattinson & Mirlees-Black 2000).
Police statistics[2] showed a 3.3% decrease in recorded crime between 1996 and 1998. However, it should be noted that there had been a slight increase of 2.7% in recorded crime over 1995–1996.
The overwhelming majority of the sample (83%) said that the crime rate was increasing over the last 2 years. Figure 2.1 shows that half said that there was ‘a lot more crime’ and a third that there was ‘a little more crime’. Only 13% and 4% respectively said that there was about the same amount of crime or less crime. If we take the correct answer as being that the crime rate is either decreasing or staying about the same, then 83% of our sample ‘got it wrong’.
Figure 2.1 Changes in recorded crime
Question: Please tell me whether you think that the amount of crime reported to the police for the country as a whole has changed over the past two years. Do you think there is a lot more crime, a little more crime, about the same, a little less crime or a lot less crime than 2 years ago?
Among survey respondents, those who substantially overestimated the amount of crime were more likely to:
Comparison with other research
This finding about New Zealanders’ inaccurate view of crime trends is of a similar magnitude to findings for the British and Welsh public. The 1996 British Crime Survey found that three-quarters of the sample surveyed believed that there was more crime than 2 years ago (Hough & Roberts 1998). This proportion had decreased in the 1998 survey, though over half the sample (59%) still thought that the crime rate was increasing (Mattinson & Mirlees-Black 2000). Yet crime figures for England and Wales had shown significant falls, in fact far greater than the decrease in the crime rate here in New Zealand.
New Zealanders perceive crime trends in their own neighbourhood more realistically than for the country as a whole. Respondents to the New Zealand national survey of crime victims (Young et al. 1997)[3] were asked whether they thought that in the last 12 months there had been more or less crime in their neighbourhood than before or whether it had stayed about the same. More than half (58%) thought that (at least in relation to their own neighbourhood), crime had remained the same, and only 16% thought that there was more crime, with 4% saying that there was a ‘lot more crime’. So, perceptions of crime trends in one’s neighbourhood appear to be more in line with actual crime trends in New Zealand than perceptions of crime trends generally.
Overseas research has found that the public has a distorted view of the amount of crime that involves violence and that violence is central to people’s perception of crime. When asked to estimate the amount of violent crime, the Canadian, Australian and British public overestimated by a considerable margin the proportion of crime that involved violence (Roberts & Doob 1988; Indermaur 1987; Hough & Roberts 1998). In addition, while 76% of respondents to Indermaur’s survey said that sentences were not severe enough, 70% were thinking of a violent crime when they responded to the question.
Yet most crime, in most western countries, consists of offences against property. New Zealand is no different. It is likely, therefore, that the media’s reporting of crime, in particular violent crime, has an impact on people’s perception of how much crime involves violence. In a small study we conducted of three daily newspapers during the data collection phase, we counted 419 crime stories – 91 related to crimes of murder.
Our next question in this series asked respondents to estimate the number of crimes out of 100 reported to the police which involve violence or the threat of violence. Police statistics show that in 1998, 8.8% of all (non-traffic-related) recorded crime was grouped as violent.[4] The definition of violence used excludes sexual offences, which make up an additional 1.5% of all recorded crime. Ministry of Justice statistics show that in 1998, 13% of all (non-traffic-related) convictions were for violent offences (Spier 1999). The Ministry of Justice definition includes most sexual offences.
We decided that any response that estimated that between 1% and 19% of crimes involve violence would be considered within the correct range. Respondents who estimated that between 20% and 49% of crime was violent were classified as those who ‘moderately overestimated’ the amount of violent crime. Respondents saying that 50% or more of crime was violent were considered to have ‘substantially overestimated’ the amount of violent crime.
Figure 2.2 shows that most respondents substantially overestimate the extent to which reported crimes involve violence. Excluding the 1.4% who said they did not know, two-thirds of the sample stated that 50% or more of crimes reported to the police involved violence or the threat of violence; with 50 violent crimes out of 100 being the most commonly mentioned (by 19% of the sample). One percent said that out of 100 crimes reported to the police, all involved violence or the threat of violence. A further 29% moderately overestimated the amount of violent crime, saying that between 20 and 49 crimes out of every 100 involved violence. Only 5% of the sample were within the correct range in their estimate of the amount of violent crime. The mean response was 53%.
Figure 2.2 How much reported crime is violent?
Question: Of every 100 crimes reported to the police, roughly what number do you think involve violence or the threat of violence?
Among survey respondents, those who substantially overestimated the proportion of violent crime were more likely to:
Comparison with other research
Table 2.1 contains the results from surveys in Australia (Indermaur 1987), Canada (Doob & Roberts 1983) and Britain (Hough & Roberts 1998; Mattinson & Mirlees-Black 2000) that asked respondents to estimate the amount of violent crime. Results from the current survey have been included for comparison purposes.
At the time of the Australian survey, 8% of crimes involved violence (based on estimates derived from official reports and crime-victimisation survey results) and responses falling between 0% and 19% were considered accurate. Respondents who said that 40% or more of crimes involved violence were considered to have made a large overestimate. Doob and Roberts (1983) suggested that less than 10% of crimes in Canada involved violence, and in Britain at the time of the survey, around 6% of crimes were violent or sexual in nature. Both Doob and Roberts (1983) and Hough and Roberts (1998) classified as accurate any responses which stated that between 0% and 9% of crimes involve violence. Respondents who said that 30% or more of crimes involve violence were considered to have substantially overestimated the amount of violent crime.
While a ‘large overestimate’ has been classified differently across all the surveys, it can be seen in Table 2.1 that in each of the overseas surveys around three-quarters of the sample substantially overestimated the amount of violent crime. While the proportion is slightly lower in the New Zealand survey (66%), it must be remembered that this category includes those who estimated violent crime to be half or more of all crime. If a large overestimate is taken as those who said that 40% or more of crime involved violence, then 75% substantially overestimated the amount of violence in the New Zealand survey. This increases to 85% when saying 30% or more of crime involved violence is taken as a large overestimate.
Table 2.1 Estimates of the amount of violent crime compared across surveys
| Indermaur (1987) | Doob & Roberts (1988) | Hough & Roberts (1998) | Mattinson (2000) | NZ Survey 1999 | |
|
Accurate |
9% |
3% |
3% |
3% |
5% |
|
Small overestimate |
17% |
15% |
19% |
19% |
29% |
|
Large overestimate |
73% |
74% |
78% |
79% |
66% |
While overseas studies suggest that the public has a distorted perception of the amount of crime which involves violence, the public also appears to overestimate the rate of burglary. Roberts and Stalans (1997) report that the Canadian public perceived that the rate of burglary is increasing at a faster rate than other types of crime. Doob and Roberts (1983) and Indermaur (1987) also found that the Canadian and Australian public overestimate the property crime victimisation rate. When asked to identify what sorts of crime problems exist in their neighbourhood, New Zealanders appear to be more concerned about burglary than any other offence (Young et al. 1997).
Survey respondents were asked to estimate the number of New Zealand households out of 100 that would be burgled over 1 year. Data gathered in the New Zealand National Survey of Crime Victims 1996 suggests that on average one house in 14 will be burgled each year or 7.3 out of every 100 New Zealand households will be burgled annually.
We classified the respondents into four categories reflecting their degree of accuracy in estimating the rate of burglary. We regarded estimates as correct if they fell within the range of one to 10 households out of 100 burgled each year. Figure 2.3 shows that only 15% of the sample gave a response that fell within the correct range for the rate of burglary in New Zealand. This excludes the 1.7% of the sample who stated that they did not know. A similar proportion slightly overestimated the rate of burglary at between 10% and 19% of New Zealand households, a third moderately overestimated the burglary rate at between 20% and 49% of households, and a third substantially overestimated the rate of burglary at 50% or more, with 11% giving 50 out of 100 as their response. As with violent crime, 1% said that out of 100 households all would be burgled over 1 year. The average response was 36%.
Figure 2.3 Knowledge of rate of burglary
Question: Of every 100 New Zealand households, roughly what number, on average, do you think will be burgled over one year?
Among survey respondents, those who substantially overestimated the rate of burglary were more likely to:
It is worthwhile noting that those who had been victims of crime were less likely to substantially overestimate both the rate of violence and the rate of burglary. While this might seem counterintuitive, it appears that the experience of being a victim, particularly of a relatively minor offence, gives people a more realistic perception of the extent of crime. The survey findings indicate that victims of sexual offending and domestic violence were more likely than victims of other crime (primarily assault and property crime) to substantially overestimate rates of both violence and burglary. In fact, respondents who had been victims of burglary or theft were much less likely than victims of other types of offences to say that 50% or more of New Zealand households would be burgled in 1 year.
Comparison with other research -
While no other survey that we are aware of has specifically asked the public to rate the number of households burgled over a year, Indermaur (1987) asked Perth residents their chances of having property stolen in the next 12 months. In parallel with the survey findings reported here for burglary, Perth residents substantially overestimated their risk of becoming a victim of a property crime. Based on Australian victimisation-survey data, the likelihood of being a victim of a property crime in Perth was 13%. Nearly half of the Perth respondents (49%) said that their chances of having property stolen in the next 12 months was 50% or more.
The New Zealand National Survey of Crime Victims (Young et al. 1997) found that the overall risk of becoming a victim differed markedly among some socio-demographic groups. It might be supposed, therefore, that those at a higher risk of victimisation may overestimate the level of crime for the country as a whole.
Interestingly though, while the survey results show that women more often than men overestimated the amount of crime (including violent crime and burglaries) Young et al. (1997) found that the same proportion of women as men were subject to some form of violent or sexual offending on one or more occasion. In addition, there were no significant differences between men and women in terms of their experience in being a victim of a property offence. However, women were generally more likely than men to report that they were very worried about all forms of crime, particularly in relation to violent offending.
Young et al. (1997) found that Māori and Pacific Peoples respondents were generally more likely than NZ European/Pakeha to be the victim of some form of violent offending, but this same pattern did not emerge with respect to individual property offences. However, both Māori and Pacific Peoples respondents expressed high levels of anxiety about crime. Similarly, the survey findings indicate that Māori and Pacific Peoples were considerably more likely to overestimate the amount of crime than NZ European/Pakeha.
The average cost of keeping a person in prison for one year is around $55,000. This cost is publicised from time to time in the media. One of the questions in the survey asked respondents what they thought this cost to be. They were presented with four options listed on a show card: $15,000, $35,000, $55,000 and $75,000.
Figure 2.4 shows that 41% correctly estimated the cost of keeping a person in prison for a year. Over a quarter (29%) underestimated the cost, and 31% said it cost $75,000 to keep a person in prison for a year.
Figure 2.4 Cost of keeping a person in prison for one year
Question: Using this card, what do you think is the average cost of keeping a person in prison for one year?
Among survey respondents, those who overestimated the cost of prison were more likely to:
Overseas research has consistently found little public awareness of statutory penalties for crimes (Roberts & Stalans 1997). Doob and Roberts (1983) found, for example, that the Canadian public tends to underestimate maximum penalties. The public also has an imprecise view of the severity of penalties actually imposed by their local courts. Research by the Canadian Sentencing Commission (1987) found that slightly over two-thirds of respondents underestimated the imprisonment rates for assault and burglary. The British Crime Survey (Hough & Roberts 1998; Mattinson & Mirlees-Black 2000) also found that most of the public underestimated the use of imprisonment for mugging, burglary and rape.
Sentencing for rape
Respondents were asked a series of questions that explored their understanding of the criminal justice system’s response to a person convicted of rape.
The statutory maximum penalty for rape
In the first of these questions, respondents were asked to identify the statutory maximum penalty for rape.
The correct answer is 20 years imprisonment.[5] Only 14% of respondents gave precisely this response. However, for the purposes of the analysis we accepted 5 years either side of 20 years as being within the correct range, that is, 15–25 years.
The majority of the sample (56%) underestimated the statutory maximum for rape at between 1 and 14 years imprisonment. Within this range, 10 years in prison was most commonly mentioned (by 29% of the sample). One-third of the sample was within the correct range. Ten percent said that the maximum penalty for rape was 26 years or more, with 7% of respondents stating that the maximum penalty was life in prison.
Figure 2.5 Knowledge of maximum sentence for rape
Question: What do you think is the maximum number of years in prison that a judge can sentence a rapist to?
Average sentence imposed for rape
Respondents were then asked to estimate the average number of years in prison a rapist would be sentenced to serve. The average prison sentence imposed for rape in 1998 was 94.6 months or approximately 8 years. Only 9% of respondents gave a response of 8 years. As with the maximum penalty, for the purposes of this analysis we accepted respondents who came within a range either side of the correct answer, in this case accepting 6–10 years as being correct.
Figure 2.6 shows that nearly half of the respondents (48%) underestimated the average number of years to which a rapist would likely be sentenced, while a similar proportion (44%) were within the correct range. Only 8% said that a rapist would be sentenced to serve on average 11 or more years in prison, and less than 1% said that a rapist would be sentenced, on average, to life imprisonment. The mean response was 7 years.
Figure 2.6 Knowledge of average sentence for rape
Question: If a rapist is sentenced to prison, how many years do you think he or she will be sentenced to on average?
Among survey respondents, those who underestimated the average penalty for rape were more likely to:
Actual time in prison for rape
In the third question in this series, we asked respondents how long they thought a rapist would actually serve in prison. We estimated that the correct response was approximately 5 years.[6] For the purposes of this analysis, we accepted responses 1 year either side of 5 years (that is, 4–6 years) as being within the correct range.
Figure 2.7 shows that nearly half of the respondents underestimated the actual number of years a rapist would spend in prison. Over a third (38%) of the sample gave a response which fell into the correct range. Only 15% of the sample overestimated the number of years a rapist would spend in prison. The mean response was 5 years.
Figure 2.7 Knowledge of actual time spent in prison for rape
Question: If a rapist is sentenced to prison, how many years do you think he or she will actually spend in prison on average?
Among survey respondents, those who underestimated the actual time spent in prison for rape were more likely to:
Comparison with other research
When Canadian respondents were asked to identify one of the offences carrying a minimum penalty, few were able to do so (Canadian Sentencing Commission 1987). More recently, Douglas and Ogloff (1997) found that psychology students in Western Canada overestimated the maximum penalties for crimes against the person but underestimated the maximum for crimes involving property.
The public also has an imprecise view of the severity of penalties actually imposed by their local courts. Research by the Canadian Sentencing Commission (1987) asked respondents to estimate imprisonment rates for assault and for breaking and entering, and slightly over two-thirds underestimated the rates. The British Crime Survey (Hough & Roberts 1998; Mattinson and Mirlees-Black 2000) also found that most of the public in both the 1996 and 1998 surveys underestimated the use of imprisonment for mugging, burglary and rape. Over half of their samples made large underestimates in the percentage of convictions for males aged 21 and over actually resulting in custody for each of the specified offences.
Life imprisonment sentence
In addition to asking our respondents about the sentences for rape we asked them to indicate how long they thought someone sentenced to serve a life sentence would actually spend in prison.
A life sentence of imprisonment is the mandatory penalty for murder and treason and is also discretionary in the case of manslaughter and dealing in Class A drugs. In 1998, 24 offenders were sentenced to life imprisonment in New Zealand (Spier 1999). Section 89 of the Criminal Justice Act 1985 provides that an offender serving life imprisonment is eligible for consideration for parole after serving 10 years of the sentence, unless a longer minimum period has been specified under section 80 of the Criminal Justice Act. Such minimum periods may be imposed in exceptional circumstances, and there is no statutory limit on the duration of the minimum period the court can impose. It is possible that such a period could be for the rest of the offender’s natural life. Once a life imprisonment inmate is eligible for parole, he or she must be considered for release at least once a year. Once released, such inmates remain subject to recall to prison for the rest of their life.
The average time spent in prison by an offender sentenced to life imprisonment is about 12 years. For the purposes of the analysis, we considered responses within 2 years either side of 12 years to be within the correct range. All other responses were classified as either underestimates of the time spent in prison for life (1–9 years) or overestimates (15 years or more).
About two in five respondents (44%) correctly estimated the time spent in prison by those sentenced to life imprisonment at between 10 and 14 years. A quarter of the sample underestimated the amount of time served in prison at between 1 and 9 years; however, only 5% said that it was 5 years or less and 18% that it was between 6 and 9 years.
A third estimated that a person sentenced to life imprisonment would spend 15 years or more in prison; 15% said that it was 15 years and 9% that it was 20 years. Only 1% of the sample said that a person sentenced to life imprisonment would actually spend the rest of their life in prison. The mean was 13 years.
Figure 2.8 Knowledge of time spent in prison for a life sentence
Question: If someone is sentenced to serve a life sentence of imprisonment, how long do you think he or she will actually spend in prison on average?
Among survey respondents, those who underestimated the time spent in prison for a life sentence were slightly more likely to:
Some overseas research has shown that the public tends to have an overly pessimistic view of the proportion of offenders who reoffend (Roberts & Stalans 1997). For example, Doob and Roberts (1983) found that Canadians tended to overestimate the likelihood of recidivism for both first-time property and first-time violent offenders.
We asked the respondents to estimate the percentage of criminals who would be reconvicted of any crime within 2 years.
In a New Zealand study of offenders sentenced in 1995, 51% were reconvicted of an offence within 2 years (Spier 2000). For the purpose of our analysis any estimate between 40% and 59% was considered to be correct. A response below this range (that is 1–39%) was considered an underestimate and anything over this (60%+) an overestimate.
Figure 2.9 shows that nearly half of our sample (46%) overestimated the reconviction rate. About a third (32%) correctly estimated the number of criminals out of 100 who would be reconvicted of any crime within 2 years as being between 40 and 59, and 21% underestimated the 2 year reconviction rate. In terms of those who overestimated their response, 13% said that 60 out of 100 criminals would be reconvicted of any crime within 2 years, 10% that 70 would be reconvicted and 9% that 80 would be reconvicted. The mean response was 54.
Figure 2.9 Knowledge of reconviction rates
Question: For every 100 people who are convicted of any crime, roughly what number do you think are reconvicted of any crime within 2 years?
Among survey respondents, those who overestimated the 2 year reconviction rates were more likely to:
Overseas research shows similar findings in relation to estimates of recidivism. Doob and Roberts (1983) found that only 12% of respondents were accurate in their estimates of recidivism for violent offenders and 21% were accurate for property offenders. Large overestimates were made by 45% of the sample with respect to recidivism of violent offenders and 34% with respect to recidivism of property offenders.
A study by Roberts and White (1986), specifically examining recidivism rates, again found that their sample considerably overestimated the proportion of first-time offenders who are reconvicted. Visitors to a Canadian Science centre were asked to estimate the percentage of offenders reconvicted within 5 years of release from their current sentence. They were asked about three offence categories: non-violent property offences, offences against the person and sex offences. Respondents estimated on average that 49% of first-time property offenders, 56% of those convicted of an offence against the person and 58% of sex offenders would be reconvicted (official rates are 27% for property offenders, 17% for offences against person and 14% for sex offenders). There were few differences in knowledge of recidivism rates as a function of socio-demographic variables.
When Roberts and White (1986) asked respondents to estimate the reconviction rates for second- and third-time offenders, their estimates did not vary from first- to third-time offenders, although in reality conviction rates rise dramatically.
Most people charged with an offence are bailed or remanded at large if their case is not concluded at first appearance. It is a fundamental principle of justice that no one should be held in detention simply because they have been accused of a crime. However, the law also recognises that there are cases where it will be necessary for people to be remanded in custody, pending trial, for reasons of public safety and the integrity of the trial process.
A number of sensational cases of violent offending while on bail has meant that this issue has been in the public eye. The last question in this series of knowledge questions asked the respondents to estimate the rate of offending while on bail.
Research by the Ministry of Justice shows that around 20% of people who are remanded on bail until their case is heard are convicted for committing an offence while on bail. This varies only slightly from year to year, and results from a recent study show that of those granted bail in 1997, 21% offended while on bail (Spier 2000).
About one in ten respondents said that 20 out of every 100 people who are charged with an offence and remanded on bail get caught for offending while on bail. Again, we classified respondents into various categories reflecting their degree of accuracy in estimating the rate of offending while on bail. We regarded estimates as correct if responses fell from 10 to 29. Other respondents were classified into two groups according to whether they underestimated the rate of offending while on bail (1–9) or overestimated the rate (30+).
Figure 2.10 shows that the majority of respondents (57%) overestimated the rate of offending while on bail. About a third of the respondents fell into the correct range. One in 10 said that 30 people get caught for offending while on bail, 8% that 40 get caught and 16% that the rate of offending while on bail is 50%. The mean response was 36.
Figure 2.10 Knowledge of offending while on bail
Question: For every 100 people who are charged with an offence and remanded on bail until the case is heard, roughly what number do you think get caught for offending while on bail?
Among survey respondents, those who overestimated the proportion of people who offend while on bail were more likely to:
We undertook further analysis to determine the characteristics of the respondents who were the best and least informed about crime statistics. We defined the least informed (or the most likely to perceive that there was a lot more crime than is the reality) as the one in three respondents who responded with substantial overestimates to at least four of five of the knowledge questions. These questions were about the direction of crime trends, proportion of crime that is violent crime, rate of burglary, reconviction rate and rate of offending while on bail.
Among survey respondents, those who were least informed about crime statistics were more likely to:
The best informed were the one in 10 who provided either correct answers or who responded with only moderate overestimates to at least four of the five questions. Survey respondents who were best informed about crime statistics were more likely to:
Overall, the survey findings support findings from overseas studies that show the public has an inaccurate and negative view of crime statistics.
Footnotes
2 Official New Zealand Police statistics, as released on 7 May 1999.
3 This survey presents the findings from the first comprehensive national survey of crime victims. The survey interviewed a random sample of the population aged 15 and over and asked them not only about the extent to which they had been subject to criminal victimisation during the period since 1995, but also about the circumstances and impact of those offences and their response to them, as well as asking about a range of other crime-related information.
4 This includes homicide, robbery, grievous assaults, serious assaults, minor assaults and other violent offences.
5 Section 128B of the Crimes Act 1961. Section 75 of the Criminal Justice Act 1985 states that a person convicted of rape may also be sentenced to preventive detention.
6 'Serious violent offenders' serve two-thirds of their sentence before being released. Therefore, given that rapists are sentenced to 94.6 months in prison on average, they would spend about 63 months or approximately 5 years in prison on average.