Sentencing in New Zealand: a
statistical analysis
Foreword | Acknowledgements | Executive Summary | Introduction | Chapter 2 | Chapter 3 | Chapter 4 | Chapter 5 | Chapter 6 | Chapter 7 | Chapter 8 | Chapter 9 | Chapter 10 | Chapter 11 | References | Appendix I | Appendix II | Tables | Figures
11.1 Outline of methods and limitations
The following results are based on the application of a multivariate method (logistic regression modelling) to determine the statistical factors influencing the probability of receiving each sentence type.
Multivariate methods estimate the effects of all statistical variables together, so that the independent contribution of each variable can be estimated. This is necessary as many of the variables are interrelated. For example, the seriousness of the current offence is higher on average for cases involving several charges, for offenders with a more extensive previous criminal history, and for offenders who are male, of Māori or Pacific ethnicity, or aged under 17 years.
Only factors that could be quantified from the available statistical data were included in the analysis. These include the type and seriousness of the major offence committed, the number of charges proved in the current case, the plea, the previous criminal history of the offender (number and seriousness of previous proved cases, rate of conviction, the time since the last case, and previous sentences) and the sex, age and ethnicity of the offender.
While these are important variables in determining sentencing, this does not imply that other factors are not important, only that they could not be measured from the available data. Therefore, this analysis is not intended as a comment on sentencing with respect to specific cases (which always involve unique circumstances influencing the choice of sentence), but rather is intended as a broad overview of the combined effects of various statistical factors on overall sentencing practice at a national level.
The models developed on recent (1995) data were used to predict the probability of receiving each sentence in three earlier years (1983, 1987 and 1991). These predicted probabilities are subject to statistical error (e.g. due to the limits of the accuracy of the model, from random errors due to small numbers in some categories, and from any changes within categories of some variables). Therefore, the analysis is only intended to indicate the relative magnitude of changes in sentencing practice.
The study is based on all proved cases involving imprisonable offences in each of the four years. The total sample size was just under 300,000 cases.
11.2 Overall changes in sentencing
The results indicate that very substantial changes in the use of different sentences have occurred since 1983. Some of this change can be explained by changes in the type of offence and offender dealt with by the courts, as measured by the statistical variables used in this study. In particular, increases in the average seriousness of offences dealt with by the courts and in the extent of previous offending for the average defendant have lead to pressure for an increase in the use of the more serious sentences, such as imprisonment and periodic detention.
However, much of the change in sentence use cannot be explained by these statistical trends. This suggests that changes in sentencing practice have occurred. That is, offenders with similar characteristics are now more or less likely to receive a specific sentence than they were in the 1980s. Changes in factors that could not be measured by this study, such as changes in the prevalence of aggravating and mitigating factors, may also have played a part in this change in sentencing practice.
A major trend over the last decade and a half has been the increase in use of community-based sentences. Overall, the use of community-based sentences has doubled as a percentage of the total proved outcomes. Approximately a third of this increase is estimated to be due to trends in the type of offence and offender being dealt with by the courts, while two-thirds is estimated to be due to changes in sentencing practice. The specific findings are:
The relative use of imprisonment decreased particularly between 1983 and 1987, following the introduction of stricter guidelines on the use of imprisonment (Criminal Justice Act 1985) and between 1991 and 1995, following the introduction of the suspended prison sentence (Criminal Justice Amendment Act 1993).
The increased use of periodic detention has largely occurred for the types of offences and offenders who would have received a prison sentence in the 1980s, suggesting that the decrease in the use of imprisonment has been balanced by the increase in periodic detention. The absolute increase in the percentage of cases resulting in periodic detention that is estimated to be due to changes in sentencing practice over the 1983 to 1995 period (an estimated +5.0%, the difference between the actual and predicted percentage for 1983) is somewhat higher than the decrease for imprisonment (an estimated -3.7%).
This estimated change due to sentencing practice is probably an underestimate, as the statistical factors that influence the increased predicted use of periodic detention include two factors (the number of previous periodic detention sentences and breaches of periodic detention) that themselves are influenced by changes in sentencing practice towards a greater use of periodic detention.
The peak use of periodic detention was reached in the early 1990s, after rapid growth in the late 1980s. Between 1992 and 1996 the use of this sentence declined again, although the probability of receiving this sentence is still much higher than it was a decade ago.
Table 11.1: Overall percentage of offenders receiving each sentence for each year, actual percentage and predicted percentage based on 1995 sentencing practice
Note: The predicted percentage is the percentage of offenders who would have received the sentence had the sentencing practices of 1995 been applied, as predicted by the 1995 logistic model. The difference between the predicted percentage for 1995 and other years indicates the proportion of the total change due to changes in statistical factors (e.g. the increase in average seriousness) while the difference between the actual and predicted percentage for each year indicates the proportion of total change due to changes in sentencing practice.
1 Sentences and court orders not included in the main analysis, including driving disqualifications, deferred or suspended sentences, community work or social welfare supervision for youth offenders, etc.
2 Discharge with or without conviction or admonished in the Youth Court (see Appendix I).
The reasons for the decreased use of monetary penalties and increased use of community service are not clear and cannot be established from this statistical analysis. The potential for net-widening in the application of community-based sentences as alternatives to imprisonment is discussed in section 11.4.2. It is also possible that community service is being more widely used now as a substitute for a monetary penalty where the offender would have difficulty in paying a fine.
Hall (1998) notes that the correct course for an offender unable to pay a fine is to 'reduce the fine to a level that is commensurate with the offender's means' (page D/381-2), but also that it is clear that the Courts consider that community service 'is also appropriate in cases where the offender lacks the means to pay a fine' (page D/468). This was true even when community service was a newly-introduced sentence option. An early study notes: 'the ability to pay a fine was the most common consideration amongst probation officers when distinguishing between offenders suitable for a fine and for community service' (Leibrich et al. 1984).
11.3 Effects of each statistical variable
11.3.1 Effects of seriousness and number of charges
The seriousness of the current offence, as measured by the Ministry of Justice seriousness scale, is strongly related to both the probability of receiving a prison sentence and the probability of a monetary penalty (as expected, given the definition of seriousness in terms of the average prison sentence imposed, section 2.2.2). As the seriousness of the major offence in the current case increases, the probability of imprisonment increases and the probability of a monetary penalty decreases (Figure 11.1). Similarly, as the number of charges proved as part of the current case increases, the probability of imprisonment increases and the probability of a monetary penalty decreases. The offence seriousness and the number of charges are among the most significant variables in the logistic regression models for prison sentences and monetary penalties.
Figure 11.1: Percentage of proved cases resulting in each type of sentence at five levels of offence seriousness, 1995
Offence seriousness score
The other sentences show a range of relationships with offence seriousness. The general pattern is that the highest probability of receiving a community-based sentence or suspended prison sentence is for offences of low to moderate seriousness (i.e. seriousness scores of >1-20 or >20-180) or moderate to high seriousness (>180-365), with a lower probability for offences of very low or very high seriousness (Figure 11.1).
For community service, the highest probability occurs for offences of relatively low seriousness (1-60). The use of periodic detention also peaks toward the low to moderate end of the range (20-60), although the probability of a periodic detention sentence is almost as high for offences of moderate seriousness (60-180) and periodic detention continues to be frequently used at the moderate to high end of the range (180-365).
Supervision and community programme show less of a relationship with offence seriousness. The probability of receiving these sentences is low for offences of very low or very high seriousness, but is fairly constant across the intermediate range, although supervision shows a somewhat higher probability for offences of moderately high seriousness.
11.3.2 Effects of offence type
Various types of offence are significant in the models for different sentences, in addition to the general effect of offence seriousness. To some extent these variables overlap. For example, most of the offences in the very high seriousness group (>365) are serious offences against the person (e.g. homicide, injury assaults, aggravated robbery, sexual violation), while other offence groups (e.g. traffic offences) are concentrated in the lower seriousness groups.
The probability of receiving a prison sentence is greater if the most serious offence in the current case is a violent offence. This is particularly so for the serious types of violent offences (serious offences against the person), but is also true for violent offences with lower seriousness scores, such as domestic violence and minor offences against the person (mainly common assaults and threats). This finding is consistent with the legal guidelines on the use of imprisonment for violent offences (Criminal Justice Act 1985). The decrease in the use of imprisonment noted in the previous section has had least effect on the imprisonment rates for serious offences against the person. As documented in a previous report (Triggs 1998), not only have imprisonment rates for the more serious of the violent offences not declined, but sentence lengths have increased.
Domestic violence is a key offence group for two other sentence types: supervision and community programme.5 Domestic violence is the most significant variable determining the probability of receiving these sentences. An offender is five times more likely to get supervision or community programme than another sentence if the current offence is domestic violence, compared to a statistically similar offender who has committed another type of offence. The use of supervision for domestic violence increased significantly in the early 1990s, at the same time as the use of imprisonment and other community-based sentences decreased.
If the current offence is a breach of periodic detention, then the odds of receiving a prison sentence or a periodic detention sentence are increased. However, the use of imprisonment for breaching a periodic detention sentence has decreased very significantly over the last decade and a half, whereas the use of periodic detention for this offence has increased. The implications of this finding are discussed further in section 11.4.1.
Offence groups that are mainly comprised of relatively less serious offences that do not involve violence (e.g. traffic, property and disorder offences) are more likely to result in the less serious sanctions, community service and monetary penalties. The use of imprisonment has decreased for these offences since 1983.
11.3.3 Effects of plea
A guilty plea appears to increase the probability of receiving community service, supervision or a monetary penalty, but decreases the probability of receiving a prison sentence or periodic detention.
11.3.4 Effects of gender
Gender 'is not in and of itself a justification for discriminating between offenders' (Hall 1998, page B173-4). Yet, the results of the multivariate modelling show that females are more likely than males to receive community service, community programme or no sentence and less likely to receive a prison sentence, periodic detention or a monetary penalty. Thus, gender differences in sentencing persist even after taking account of differences in the type and seriousness of the offence committed (e.g. the average seriousness of offences committed by women is lower than for men) and in the extent of previous offending (e.g. women have fewer previous convictions on average; section 3.1). Indeed, gender is the amongst the most significant variables influencing the probability of receiving a community service sentence or a monetary penalty.
The reason for these gender differences in sentencing cannot be determined from the statistical analysis. For example, it is possible that at least some of this difference may be due to factors that could not be measured in this study, such as systematic differences between the sexes in the average circumstances of cases (e.g. gender differences in the relative gravity of offending even within specific offence types or gender differences in the actions or circumstances of the average offender).
11.3.5 Effects of age
As noted in section 2.2.1, age is considered important as an indicator of rehabilitative potential, so that a youthful offender may receive a lesser sentence than a more mature offender. The Criminal Justice Act 1985 states that prison sentences should not be imposed on a person under the age of 16, except for a purely indictable offence, and discourages the use of imprisonment except as a last resort for young offenders. Periodic detention cannot be imposed on people aged under 15 years. The Youth Court cannot impose prison or community-based sentences.
These guidelines are certainly being followed in practice. Offenders aged under 17 have an extremely low relative risk of receiving any of the sentences considered here, but especially the more serious sentences of imprisonment and periodic detention. The relative use of these sentences has also decreased since 1983, following the two major legislative changes, the Criminal Justice Act 1985 and the Children, Young Persons, and Their Families Act 1989.
The latter Act had a major effect on the number of youth offenders who are dealt with by the court system, as most offenders aged under 17 are now dealt with by Family Group Conferences. One effect of this change is that a much smaller proportion of the proceedings involving youth offenders now result in a conviction or a proved outcome. The relatively few proved cases for youth offenders therefore tend to be for serious offences. Because of this, the actual percentage of cases resulting in a prison sentence has increased for youth offenders, even though the relative use of imprisonment (taking the seriousness of the offence into account) has decreased significantly.
Offenders aged 17-19 or 30 plus are less likely to receive a monetary penalty, but more likely to receive community service, supervision, or community programme than offenders aged 20 to 29 years.
11.3.6 Effects of ethnicity
As noted in section 2.2.1, people of different ethnic or cultural groups must be treated equally, although 'alternative means of rehabilitation which appropriately take account of different cultural or ethnic values should be utilised' (Hall 1998, page B/173). In particular, the community programme sentence is seen as a means to achieve this. Although the community programme sentence has never been widely used, Māori and Pacific offenders are about twice as likely to receive this sentence as Pakeha/Other offenders, once the effects of other factors have been taken into account.
Two other community-based sentences, periodic detention and community service, are also used relatively more for Māori and Pacific offenders, whereas these ethnic groups are less likely to receive a monetary penalty. The use of imprisonment did not differ between ethnic groups, once other factors had been taken into account (such as the differences between ethnic groups in the type and seriousness of offences committed and in the extent of previous offending; section 3.1).
11.3.7 Effects of criminal history variables
Various criminal history variables were tested in the logistic regression models, including the number of previous proved cases, the weighted sum of the past seriousness scores, the rate of conviction (number of proved charges per year), the time since the most recent previous case, and the previous sentences imposed.
Each of these variables shows some relationship to the probability of each type of sentence and various forms of these variables were significant in most of the models. However, the criminal history variables, with the exception of the previous sentence variables (see the following section), are not generally amongst the most significant variables, once the effects of other variables were taken into account.
The greater the number, seriousness and frequency of previous cases, and the shorter the time since the previous case, the greater the probability of a prison sentence and the lower the probability of a monetary penalty. The use of imprisonment has decreased since 1983 for offenders with any number of previous cases.
Criminal history is a distinguishing factor between offenders sentenced to periodic detention compared to community service. As the number of previous cases increases, the probability of periodic detention increases, whereas the probability of community service decreases. However, at very high levels of previous offending the probability of periodic detention decreases again, as imprisonment becomes more likely.
The probability of community programme also tends to be higher for offenders with a moderate to high level of previous offending, although criminal history variables are of limited significance in the community programme model (and the model as a whole is very poor). Supervision shows weak relationships with criminal history variables, although the use of supervision does tend to increase as the rate of conviction increases. Criminal history variables are of limited significance in the supervision model once other factors are accounted for.
Sentencing practice has changed significantly since 1983 with respect to criminal history variables. The use of periodic detention for offenders with several previous cases has increased, while the trend for imprisonment has been in the opposite direction. Also, supervision used to be less likely for persistent offenders, but now the probability of supervision is higher for offenders with several or many previous cases than for those with one or a few previous cases.
Community service has shown the most significant change in trend. In the 1980s the probability of receiving a community service sentence was fairly uniform across all levels of previous offending, but in the 1990s community service is much more likely to be imposed on offenders with one or very few previous cases. The use of community service for persistent offenders has also increased, but the magnitude of the increase has been much greater for those with no or a few previous cases. The use of monetary penalties, in contrast, has decreased for all offenders.
11.3.8 Effects of previous sentences
The most recent sentence prior to the current case, and also other past sentences, have a significant impact on the probability of the current sentence. This appears to be the strongest effect of any criminal history variable in most models, although the relationship between previous sentences and other criminal history variables are obviously not independent. That is, the previous sentence variables may be the most significant criminal history variable because they provide a useful summary of the criminal history, as the previous sentence is itself influenced by the number and type of convictions that have occurred in the past.
Nonetheless, the previous sentence variables did stand out as key factors in the models, even though a range of other criminal history variables were also tested. Therefore, the previous sentences (and particularly the most recent sentence) appear to have an additional, independent effect on the current sentence. The direction of the effect for all sentences is towards an escalation of the current sentence. Thus, the previous sentence increases the risk of the same sentence or a more serious sentence, but decreases the risk of less serious sentences, all other factors being equal:
The implications of this finding are discussed in the following section.
11.4 Interaction effects
11.4.1 'Fast-tracking' or sentence escalation
The results summarised in the previous section indicate that having a previous community-based sentence decreases the probability of receiving a monetary penalty and increases the probability of receiving a further community-based sentence or prison sentence. Given this finding, any increase in the use of community-based sentences may lead to a reinforcing cycle of sentence escalation, or 'fast-tracking' of offenders toward more serious sentences.
Therefore, the increased use of community service for offenders who would previously have received a monetary penalty may in turn have led to a greater use of community service and other community-based sentences for the same offenders when they are reconvicted. Following the next reconviction, the same process is repeated, and so on in a reinforcing feedback cycle, increasing the use of supervision and periodic detention and ultimately putting pressure on the use of imprisonment.
With the increased use of community-based sentences, a separate but related reinforcing effect is also likely. Increasing the number of people serving community-based sentences increases the number of people breaching these sentences. A breach of a community-based sentence, and especially a breach of periodic detention, puts the offender at higher risk of a periodic detention or prison sentence, thereby further reinforcing increases in the use of serious sentences.
The number of periodic detention sentences imposed increased by 2.8 times between 1982 and 1997. Over the same period the number of breaches of periodic detention increased by 2.4 times. By 1995, cases involving a breach of a periodic detention sentence accounted for 9% of cases resulting in periodic detention.
This cycle of escalation will only occur if offenders serving a community-based sentence have a similar or higher rate of reconviction than offenders who receive monetary penalties (i.e. if community-based sentences reduce the rate of reconviction then there would be fewer re-offenders to whom sentence escalation could apply). However, the analysis presented in Appendix II indicates that the sentence imposed is not a major factor in determining the probability of reconviction and that, if anything, offenders sentenced to a monetary penalty have lower reconviction rates.
One effect of sentence escalation is increased pressure on the use of imprisonment, as offenders are 'fast-tracked' up the penalty scale to imprisonment sentences. However, the effect of this has been countered by the change in the use of imprisonment relative to community-based sentences. In particular, the use of imprisonment for offenders who have breached a periodic detention sentence is now greatly reduced compared to the 1980s. In 1983, 32% of offenders who breached a periodic detention sentence received a prison sentence, compared to 16% in 1995. Conversely, 29% of offenders who breached a periodic detention sentence in 1983 received a periodic detention sentence, compared to 51% in 1995.
Similarly, although the likelihood of imprisonment is still relatively higher for offenders whose most recent sentence was periodic detention, the use of imprisonment for this group of offenders has decreased considerably. In 1983, 26% of offenders whose most recent sentence was periodic detention received a prison sentence as the current sentence, compared to 14% in 1995. The use of imprisonment following community service has also decreased, as has the use of imprisonment in general for offenders with several previous convictions.
These changes in the use of imprisonment have further increased the use of periodic detention and supervision, as offenders who would previously have gone to prison are now more likely to receive the more serious community-based sentences than in the 1980s.
This analysis indicates the need for an awareness of the potential flow-on effects of the increased use of community-based sentences for low-end offences. There are also other potential implications of this change in sentencing practice, such as the cost of administering additional community-based sentences and the potential loss of fines revenue. On the other hand, these factors must be weighed against the potential benefits of the increased use of community-based sentences, such as the involvement of the offender with the community, the value of work done on periodic detention or community service, and the benefits of programmes attended while on supervision or community programme. More research is needed on evaluating these factors.
11.4.2 Alternatives to imprisonment and 'net-widening'
Alternatives to imprisonment and why the prison population continues to increase
The original aims of extending the range of community-based sentences available and promoting the use of these sentence were to reduce the use of imprisonment and to encourage community involvement. The intention was both to reduce costs and to provide options that improved the outcomes for offenders and society (Penal Policy Review Committee 1982). Thus, community-based sentences were originally viewed as an alternative to imprisonment.
However, each of these sentences now appears to be considered as a sentence in its own right (as outlined in Hall 1998), although: 'Whether community service (or indeed any of the community-based sentences) should be seen as an alternative to imprisonment or as a sentence in its own right is an open question.' (Hall 1998, page D/467).
Despite the very widespread use of community-based sentences, the prison population has continued to grow. Because of this, new alternatives to imprisonment are often proposed. For example, the suspended prison sentence was introduced in 1993 and the home detention option (electronic monitoring) is about to be extended.
To a large extent, the growth of the prison population has been due to the increase in the number of serious offences being dealt with by the criminal justice system and the significant increase in the length of prison sentences imposed and served for serious violent offences (Triggs 1995, 1997, 1998). Trends in these serious offences have by far the largest effect on the size of the prison population due to the very long sentences served by these offenders.
On the other hand, the very substantial increase in the use of alternatives to imprisonment (community-based sentences and suspended prison sentences) should have had more of an impact on reducing the prison population than they have had. Although imprisonment rates for many offences and offenders have decreased, these decreases have been much smaller than the overall increases in the use of community-based sentences and suspended sentences. This finding suggests that net-widening has occurred, as discussed below. In addition, there appear to be feedback effects from the use of these sentences that actually increase the pressure on the use of imprisonment. These feedback effects include sentence escalation for community-based sentences (section 11.4.1) and the high activation rates for suspended sentences (Spier 1998).
Net-widening
The widespread application of sentences intended as an alternative to imprisonment to offenders who would not otherwise have received a prison sentence is known as 'net-widening'. One of the reasons it is so difficult to avoid net-widening is the size and nature of the target population of offenders.
The logistic regression model of imprisonment, described in Chapter Four, was able to identify offenders with close to a 100% probability of imprisonment, while other offenders had nearly a zero risk of imprisonment (Figure 11.2). However, this does not mean that most people who receive a prison sentence necessarily have a high risk of imprisonment. Forty percent of the people who actually get a prison sentence are derived from that group of offenders who have less than a 25% predicted probability of receiving a prison sentence, whereas people who have a predicted probability of over 75% account for just 17% of the total prison sentences.
This apparently counter-intuitive result is because the distribution of offenders is highly clustered around the low end of the scale (Figure 11.2). There are 72,828 offenders in the 1995 data-set who have less than a 25% predicted probability of receiving a prison sentence. Even at a very low average probability of imprisonment of 4%, these offenders account for 72,828*0.04=2971 prison sentences. In contrast, very few offenders have a high probability of receiving a prison sentence, therefore this group accounts for fewer prison sentences (1475 offenders with an average probability of 0.83 resulting in 1222 prison sentences). A further 1877 of those who receive a prison sentence have a 25-50% probability of imprisonment and 1327 have a 50-75% probability.
Thus, although the factors that lead to a high probability of imprisonment are clear (i.e. offences of very high seriousness, especially violent offences committed by persistent offenders), relatively few offenders fall into this high probability group, compared to the much larger number with a low to medium probability of imprisonment. This means a very large number of offenders are eligible to be considered for sentences that are alternatives to imprisonment at the low to medium end of the scale. 6
Therefore, even a very small shift towards community-based sentences can result in a very large number of these sentences being imposed, with relatively little impact on the number of prison sentences imposed. This is particularly so if, as for community service, the sentence is targeted as the lower end of the range, where the greatest numbers of offenders are.
Figure 11.2: Plot of the predicted probability of receiving a prison sentence versus the actual proportion receiving a prison sentence and the number of offenders at each level of probability
The introduction of the suspended prison sentence in 1993 provides a good example of the potential for net-widening. From Chapter 5, the factors influencing the use of suspended sentences are similar to the characteristics that indicate a moderate level of risk of imprisonment, with the exception of the greater use of imprisonment for male offenders. The risk of imprisonment can be predicted using the prison model. Offenders who received a suspended sentence in 1995 have an average predicted probability of receiving a prison sentence of 25%, about the same as the 27% probability for offenders who actually received a prison sentence of less than six months, but lower than the average predicted probability of 44% for those imprisoned for between six months and two years.
These findings suggest that offenders who receive a suspended sentence are similar to offenders who get short prison sentences. This result is consistent with the finding of previous research (Spier 1995-98) that the main reduction in prison receptions has occurred for prison sentences of less than six months rather than in the six months to two years range set down in the Act.
However, the previous research also shows that most suspended sentences have not been imposed in place of prison sentences, but rather in place of or in addition to community-based sentences. Why is this so, if the sentence has been targeted at the same risk of imprisonment level as those who get short prison sentences? The answer may lie in the difficulty of targeting such offenders accurately, as explained in the previous section. Thus, the type of offender who is most likely to receive a short prison sentence has a predicted probability of imprisonment of around 25% and therefore has a 75% probability of receiving a non-custodial sentence (mainly periodic detention or supervision).
This group of offenders with a probability of imprisonment of about 25% mainly comprises offenders in one of three groups:
As pointed out above, the majority of offenders have a risk of imprisonment of less than 25% (as predicted by the statistical model), yet this category supplies the largest number of actual prison sentences, due to the far greater number of offenders in this category. Even allowing for the significant improvement in sentence targeting once case-specific circumstances are taken into account, there remains some potential to impose a large number of sentences as an 'alternative to prison', by targeting the right general group of offenders, and yet not actually impose the alternative sentence on the specific offenders who would otherwise have received a prison sentence. The large number of people in the moderate risk category (i.e. the target group for alternative sentences) means that large numbers of the alternative sentence can be potentially imposed without making much of a difference to the number of prison sentences imposed.
In the case of suspended sentences this has a double effect. Not only is the reduction in the number of people receiving prison sentences less than hoped for, but the large number of suspended sentences imposed, in combination with high reconviction rates, means that a significant proportion of suspended sentences are activated, thereby adding to the prison population. Also, the offence leading to the reconviction may be more likely to receive a prison sentence, further adding to the increase in imprisonment.
FOOTNOTES
5.'Domestic violence', as defined for this study, includes only those offences recorded as 'male assaults female', 'common assault (domestic)' or 'assault on a child'.
6. The choice of imprisonment in individual cases will certainly be much more clear-cut than is indicated by the statistical results, given the case-specific circumstances of the offence and offender that could not be statistically measured for this study. Nevertheless, there is still likely to be a significant group of offenders for whom prison is an option but not a certainty (e.g. for those who have committed moderately serious offences and who have several previous convictions).
