Estimating The Effects Of Incapacitation Research Paper ⋆ Research Paper Examples ⋆ EssayEmpire

This sample Estimating The Effects Of Incapacitation Research Paper is published for educational and informational purposes only. If you need help writing your assignment, please use our research paper writing service and buy a paper on any topic at affordable price. Also check our tips on how to write a research paper, see the lists of criminal justice research paper topics, and browse research paper examples.

Overview

Among the manifold goals of penal confinement, incapacitation is intended to impose a period of “time out” from an offender’s criminal career, by deliberate removal of the opportunity for the offender to commit crime in the community for the duration of his or her sentence. From a societal standpoint, therefore, the key question for incapacitation is the following: On average, how many crimes are prevented in society by incarcerating an individual for a given period of time? If the criminal justice system is operating in an efficient manner, then those who are incarcerated will be the most serious and prolific offenders in the population, and the incapacitation effect of prison will be substantial.

After a period of over 20 years of empirical stagnation, questions concerning the incapacitation effect of prison have resurfaced, as indicated by a 2007 issue of Journal of Quantitative Criminology especially devoted to empirical research on the topic. This recent attention has been spurred, in part, by the unparalleled growth in the use of incarceration in Western society over the last three decades. This growth is most stark in the United States, which presently has an incarceration rate (including jails and state and federal prisons) over 750 per 100,000 residents. Yet questions concerning the incapacitation effect of prison are not the sole province of American criminology, as prison growth has been a more widespread Western phenomenon. In Europe as well, two-thirds of 35 countries recently surveyed experienced growth in their incarceration rates during recent decades.

The goal for a study of incapacitation is to estimate the number of crimes that an incarcerated offender would have committed, were he free in the community rather than confined in prison. This is taken as an estimate of the “incapacitation effect,” defined as the number of crimes averted by physically isolating an offender from society at large. This task would seem straightforward at first glance, but the problem on closer inspection is that this quantity is a “counterfactual” which can never, even in principle, be directly observed. Instead, the counterfactual outcome – how many crimes the offender would have committed if free – is unobserved and must be inferred from indirect sources. The challenge to providing a valid incapacitation effect on criminal offending, therefore, is one of overcoming this counterfactual problem.

In this research paper, an overview of the first generation of incapacitation studies is first provided. These studies employ a within-individual design to solve the counterfactual problem. This is followed by a description of the most recent round of incapacitation research, with a focus on studies that employ a between-individual solution to the counterfactual problem. Details are provided on one specific between-individual approach that entails propensity score matching. This research paper closes with a discussion of the ongoing empirical challenges to estimation of the incapacitation effect of prison.

Traditional Approaches To Estimating Incapacitation Effects

Spelman (2000) acknowledges two general kinds of incapacitation research. “Top-down” approaches use aggregate data to estimate the impact of growth in the prison population or in prison commitments on crime rates. Top-down estimates are often preferred by policy analysts for their ability to estimate the total effect of imprisonment on crime. Such estimates are, by necessity, agnostic about the specific mechanism by which prisons influence behavior – specifically, through incapacitation, deterrence, or rehabilitation. Indeed, by treating the crime prevention mechanism as a black box, top-down studies are unable to discriminate between competing explanations for prison effectiveness. “Bottom-up” approaches, on the other hand, attempt to peer into the black box of prison effectiveness by using individual-level data to estimate incapacitation effects. These studies typically address the incapacitation question by using arrestee or inmate samples to obtain self-report estimates of offending frequency in the months prior to criminal justice intervention, a quantity referred to in the criminal career literature as “lambda” (see Blumstein et al. 1986).

The remainder of this section focuses on “bottom-up” studies of prison effectiveness (Spelman 2000), or empirical attempts to estimate individual criminal offending rates which can then inform estimates of incapacitation effects. Estimation of the incapacitation effect of prison in such studies is fundamentally an actuarial exercise. The empirical challenge is one of overcoming the counterfactual problem, that is, providing a credible estimate of behavior that is not observed because an individual is locked up in jail or prison rather than free in the community. The key quantity for incapacitation is known as lambda, l, representing the annual offending frequency conditional on active offending, which can be taken as an estimate of the number of crimes avoided through incarceration. Zimring and Hawkins (1995, p. 81) observe that there are two approaches to estimating lambda:

T]o determine the level of crime that would have occurred if a particular group had not been confined, one must either study the criminal activity of the same group at a different time in their lives to estimate what that group would have done if not confined, or one must study the behavior of persons other than those confined to approximate the crimes avoided by imprisonment in the past.

In the former approach, the goal is to estimate a within-person counterfactual, while in the latter approach, the goal is to estimate a between-person counterfactual. The most influential incapacitation studies are universally of the first kind in Zimring and Hawkins’ (1995) typology. These studies provide within-person counterfactual offending rates from self-report surveys of arrestees or prison inmates. With this approach, targeted individuals are questioned about their criminal activity during the months leading up to their arrest or confinement. These estimates are taken as the number of crimes they committed on an annual basis when they were free (prior to incarceration), and by implication, the number of crimes they would have committed per year during the time that they were incarcerated. Therefore, the counterfactual offending rate for incarcerated individuals is their own offending rate in the months preceding their confinement.

The best known source of a within-person counterfactual is the Rand Corporation’s second inmate survey, a study of over 2,000 male inmates in California, Michigan, and Texas. The report by Chaiken and Chaiken (1982), for example, revealed that annual offense frequencies for ten different crimes were highly skewed, with active offenders at the median committing 15 offenses and offenders at the 90th percentile committing 605 offenses (these figures exclude drug offending). They estimated the mean offending rate to be either 187 or 278 crimes per year, depending on how ambiguous survey responses were treated. Based on these figures, incarceration was shown to have the capacity to substantially incapacitate criminal behavior. In fact, on the basis of this research, policy-oriented criminologists began to advocate “selective incapacitation” of high-rate criminal offenders as an explicit penal policy. In hindsight, however, research efforts using risk assessment tools to identify high-rate offenders and target them for longer prison sentences have been notoriously unsuccessful, calling into question the wisdom of a selective incapacitation policy.

These early bottom-up, within-person incapacitation studies had limitations, however. First and foremost is the question of the reliability of offender self-reports of their criminal behavior (Spelman 1994; Zimring and Hawkins 1995). It is unclear whether incarcerated offenders, particularly high-rate offenders, can accurately recall their prior criminal activity. Even if they can, they may not be motivated to report it honestly to interviewers. Spelman (1994), in fact, found evidence of both overreporting and underreporting in the second Rand inmate survey, but by individuals at different locations in the distribution of offending rates. Second, the presence and nature of “crime spurts” can introduce serious distortions in estimates of lambda (Blumstein et al. 1986) and, therefore, in estimates of incapacitation effects. Research has shown that offenders experience relatively short periods of high-rate offending immediately prior to incarceration, meaning that incapacitation effects will be severely overstated, especially if reporting windows are comparatively narrow. Moreover, a portion of the pre-incarceration crime spurt appears to be artifactual rather than behavioral, that is, a function of the way that individuals are filtered through the criminal justice system and selected for custodial sentences rather than reflective of genuine growth in offending frequency (Maltz and Pollock 1980).

Third, the early studies were possibly guilty of overly optimistic estimates of incapacitation effects. The Rand Corporation studies showed self-report offending rates to vary widely from one jurisdiction to another. For example, Chaiken and Chaiken (1982) reported the median total nondrug crime frequency among prisoners (not including jail inmates) in California, Michigan, and Texas to be 42, 17, and 9, respectively. Offending rates in other jurisdictions that attempted to replicate the Rand survey were comparably varied. Relatedly, the incapacitation estimates from this research tend to assume constant offending during the reference period, whereas Horney and Marshall (1991) demonstrated that most offenders were actively involved in crime only intermittently, and that failure to account for this intermittency would inflate estimates of the incapacitation effects of prison when offense rates are annualized.

Finally, existing bottom-up studies have employed samples of incarcerated offenders with questionable generalizability to contemporary circumstances (see Zimring and Hawkins 1995, for a similar critique). Prior studies base incapacitation estimates on inmates who were incarcerated near the beginning of an unprecedented expansion in US prisons. From the 1930s through the early 1970s, the incarceration rate hovered around 110 per 100,000 residents, then began a steady increase in the early 1970s and at present is in excess of 750 per 100,000. If the criminal justice system operates efficiently, officials will identify and incarcerate the most serious offenders – those who commit the most serious crimes and who do so at a high rate. Holding all else equal, however, rapid expansion in the prison population will also result in less active or less serious offenders entering prison. Contemporary incarceration thus might yield lower incapacitation effects than incarceration in earlier decades because of diminishing marginal returns as the criminal justice system reaches deeper into the offender queue. Therefore, incapacitation studies conducted in the 1980s and earlier may have limited utility for policy makers today, and may in fact overestimate the current incapacitation effect of prison.

Contemporary Approaches To Estimating Incapacitation Effects

Until recently, within-person counterfactual studies exhausted all empirical research in the incapacitation tradition. Two studies in the last 3 years have adopted the second approach in Zimring and Hawkins’ (1995, p. 81) typology, that of estimating the offending rate of “persons other than those confined to approximate the crimes avoided by imprisonment.” The studies by Owens (2009) and Sweeten and Apel (2007) represent only two of a new generation of incapacitation research. One strand of this research employs simulation techniques (Bhati 2007; Blokland and Nieuwbeerta 2007), while another strand takes advantage of a variety of aggregate “natural experiments” (Barbarino and Mastrobuoni 2008; Johnson and Raphael 2009; Kessler and Levitt 1999; Ramirez and Crano 2003).

Owens (2009) takes advantage of a change in the sentencing guidelines in Maryland that lowered the age at which an offender’s juvenile record can be used to add “criminal history points” for the purpose of criminal sentencing. This policy change effectively reduced the recommended sentence faced by certain groups of young adults with a juvenile criminal history. In her study, individuals who were sentenced after the policy change accumulated 2.8 arrests per year (with an implied Index offending rate of 1.5 offenses per year) during the time that they would still have been confined under the old sentencing regime.

Sweeten and Apel (2007) rely on propensity score matching to select incarcerated and non-incarcerated individuals who closely resemble each other on a wide variety of background variables, including criminal history. Using data from a contemporary, nationally representative sample of American youth, they limit their attention only to individuals who were not incarcerated prior to the reference conviction. They estimate lambda among the non-incarcerated sample to be about 9 offenses per year among 16–17-year-olds (95 % confidence interval = 6.2, 14.1), and about 6 offenses per year among 18–19-year-olds (95 % confidence interval = 4.9, 8.4). In the next section, the logic and method of matched sampling, as used by Sweeten and Apel (2007), is described in more detail.

Details On Matched Sampling

Matched sampling is a quasi-experimental method of estimating the number of crimes averted through incarceration. The study by Sweeten and Apel (2007) is the only application of this approach thus far. Specifically, the offending rate of a “non-captive” but high-risk comparison sample is taken as the counterfactual offending rate for an incarcerated sample during the period of its confinement. In other words, the incapacitation effect is the offending rate of a contemporaneous sample of individuals who are not incarcerated.

This particular approach has appeal because it addresses the counterfactual problem in a straightforward and transparent way. The intuition underlying propensity score matching can be elaborated by drawing comparisons with an experimental approach. In a hypothetical experiment, individuals would be targeted for incarceration at random. This ensures that the experimental and control groups are balanced, meaning that they are indistinguishable (in expectation) on all possible sources of nonequivalence which might confound incapacitation estimates. By virtue of randomization, the control group serves as a counterfactual for the experimental group, or an estimate of what the experimental group’s behavior would have been had it not been imprisoned. Randomization ensures that no additional adjustments are required, because all sampled individuals were “at risk” of being incarcerated with known probability; that is, at the start of the experiment, all were equally eligible to be assigned to the experimental group. In this scenario, the assignment mechanism is controlled by a statistically random process that renders it ignorable. The incapacitation effect of imprisonment then is simply the mean (or median) offending rate of the entire control group during the experimental group’s confinement.

Propensity score matching is a quasiexperimental approach that attempts to approximate the conditions of a randomized experiment by creating “synthetic” experimental and control groups that are balanced on a wide variety of confounding variables. The technique is particularly advantageous in situations where randomization is not possible, but the assignment mechanism is nevertheless well understood and can be explicitly modeled from variables that are readily available to the analyst. A well-specified model increases confidence that the assignment mechanism is ignorable conditional on the propensity score, a situation that satisfies what is known as the conditional independence assumption.

The initial step of this approach is estimation of the propensity score, which in the present context can be defined as the predicted probability of incarceration conditional on criminal conviction. To meet the conditional independence assumption requires intimate knowledge of the incarceration process. For instance, research on criminal sentencing suggests that the key determinants of incarceration are the characteristics of the instant offense the offender’s criminal history, and, to a much lesser degree, the offender’s personal characteristics and life circumstances, characteristics of the sentencing judge, and jurisdictional factors.

Once the incarceration process has been adequately modeled and the propensity score estimated, attention turns to the matching process. The goal at this stage is to identify, for each incarcerated individual, a non-incarcerated counterpart who is observationally equivalent. A “suitable” match is therefore defined as a non-incarcerated individual who has a propensity score which is identical (or if not identical, at least within a minimum distance known as a caliper) to that of his or her referent, incarcerated peer. A property of propensity score matching is that not all incarcerated subjects will necessarily have a suitable match, and conversely, not all non-incarcerated subjects will necessarily be matched. A feature of the common support condition is that only those individuals who most closely resemble one another on the propensity score – and by implication, the variables indexed by the propensity score – are actually matched and therefore contribute information to the estimate of interest. With matched samples in hand, the final step is to estimate the counterfactual of interest.

Ongoing Challenges To Estimating Incapacitation Effects

Existing studies examine the incapacitation effect of prison incarceration, with virtually no attention devoted to sentences of incarceration less than 1 year. (A recent exception is Sweeten and Apel 2007, although they do not distinguish prison from jail incarceration in their incapacitation estimates.) This is a glaring omission, as jail incarceration in the United States is far more widespread than prison incarceration. The jail population at any given time tends to be about one-half the size of the prison population.

However, the average daily jail population of about 750,000 individuals, from which the jail incarceration rate is calculated, underestimates by a very large margin the number of individuals who actually pass through the nation’s jails in a year, which is estimated to be a shade under 13 million. Taking account of the fact that just over 60 % of these individuals are awaiting trial still yields about five million who are incarcerated in jail for a crime. So while the prison incarceration rate of about 500 per 100,000 provides a fairly accurate estimate of the number of people who spend time in prison during a year, the true jail incarceration rate is over 1,500 per 100,000. Short sentences of incarceration are clearly the norm, yet virtually nothing is known about their incapacitation potential.

Incapacitation effects are also likely to be highly sensitive to the inclusion of drug offenses, which presents two empirical problems – frequency and replacement. First of all, drug crimes are less serious violations that occur at very high frequency, far higher than other types of crime. For example, Chaiken and Chaiken (1982) report from the Rand Corporation’s second inmate survey that among active drug dealers, the median offender sold drugs 100 times per year, while the offender at the 90th percentile did so on 3,251 occasions. The mean was between 880 and 1,299 crimes per year, depending on how ambiguous responses were handled. When drug offenses were excluded from the incapacitation effect, the average offender reported committing between 187 and 278 crimes per year (median = 15), but when drug offenses were included, the average offender committed between 614 and 933 crimes (median = 42). Thus, in samples which include large proportional representation of drug offenders, incapacitation estimates that include drug offending will be highly distorted.

A second problem with the inclusion of drug offenses in incapacitation estimates is that they are likely to be subject to much higher replacement rates than other kinds of crime. For instance, the removal of an active drug dealer from one street corner presents an opportunity for another drug dealer to take his or her place. If replacement in this scenario is 1-for-1, and the replacement drug dealer sells drugs at the same rate as the newly incarcerated drug dealer, then the incapacitation effect is effectively 0 from a societal standpoint. In reality, the replacement rates for specific crimes, as well as the offending rates among replacement offenders, are unknown. Nevertheless, the very nature of drug offending – drug transactions can be conducted briefly and at high volume, as well as a replacement rate that is likely to be high – renders it highly problematic for use in generating estimates of the incapacitation effect.

Incapacitation effects are likely to be sensitive to co-offending patterns. Consider two offenders who prefer to only commit crime in each other’s company, which means that their offending rates are equal. Incarceration of a referent offender in this co-offending dyad effectively removes the opportunity of both offenders to commit crime. In a manner of speaking, both offenders have been incapacitated, even though only the referent offender is imprisoned. On the other hand, if the co-offending partner continues committing crime at a rate which is unchanged, then the same number of crimes which would have been committed if the referent offender was not incarcerated will still continue unabated. In this latter scenario, the incapacitation effect is 0 from a societal standpoint (similar to the 100 % replacement rate discussed in the drug offending example above). The implication of co-offending networks for the magnitude of incapacitation estimates would thus be a very fruitful line of inquiry.

Conclusion

A growing literature has become attentive to the societal consequences of prevailing imprisonment policies. Increasingly, this literature cautions that the social costs of prison growth might outweigh the incapacitation benefits. One prominent line of research observes that confinement worsens an individual’s capacity to adopt a law-abiding lifestyle upon returning to the community. Such “collateral consequences” of incarceration include labor market stigma, slowed wage growth, and marital disruption, among other consequences. These have the potential to exacerbate long-term criminal offending. Thus, while prison sentences will temporarily incapacitate criminal behavior, they might worsen offending over the long run relative to noncustodial sentences. It will thus be important to provide a full accounting of the impact of prisons on criminal behavior – both for better and for worse.

It will also be important to put incapacitation effects into historical context. Prison growth leads to the confinement of individuals who pose steadily lower risk to society, on the margin and all else equal. Consistent with the “law of diminishing returns,” recent studies of prison expansion show substantial erosion in crime control over the last three decades as the scale of incarceration has grown (Johnson and Raphael 2009). To the degree that the marginal return with respect to crime does indeed shrink with unmitigated growth, incapacitation is likely to play an increasingly minor role in policy discussions concerning the use of prison.

Bibliography:

Barbarino A, Mastrobuoni G (2008) The incapacitative effect of incarceration: Evidence from several Italian collective pardons. Unpublished manuscript. Research and Statistics Division, Federal Reserve Board, Washington, DC
Bhati AS (2007) Estimating the number of crimes averted by incapacitation: an information theoretic approach. J Quant Criminol 23:35–375
Blokland AAJ, Nieuwbeerta P (2007) Selectively incapacitating frequent offenders: costs and benefits of various penal scenarios. J Quant Criminol 23:327–353
Blumstein A, Cohen J, Roth JA, Visher CA (eds) (1986) Criminal careers and “career criminals”, vol 1. National Academy Press, Washington, DC
Chaiken JM, Chaiken MR (1982) Varieties of criminal behavior. Report No. R-2814-NIJ. Rand, Santa Monica Horney J, Marshall IH (1991) Measuring lambda through self-reports. Criminology 29:471–495
Johnson R, Raphael S (2009) How much crime reduction does the marginal prisoner buy? Unpublished manuscript. Goldman School of Public Policy, University of California, Berkeley
Kessler D, Levitt SD (1999) Using sentence enhancements to distinguish between deterrence and incapacitation. J Law Econ 42:343–363
Maltz MD, Pollock SM (1980) Artificial inflation of a delinquency rate by a selection artifact. Oper Res 28:547–559
Owens EG (2009) More time, less crime? Estimating the incapacitative effect of sentence enhancements. J Law Econ 52:551–579
Ramirez JR, Crano WD (2003) Deterrence and incapacitation: an interrupted time-series analysis of California’s three-strikes law. J Appl Soc Psychol 33:110–144
Spelman W (1994) Criminal incapacitation. Plenum, New York
Spelman W (2000) What recent studies do (and don’t) tell us about imprisonment and crime. In: Tonry M (ed) Crime and justice: a review of research, vol 27. University of Chicago Press, Chicago, pp 419–494
Sweeten G, Apel R (2007) Incapacitation: revisiting an old question with a new method and new data. J Quant Criminol 23:303–326
Wermink H, Apel R, Nieuwbeerta P, Blokland AAJ (2012) The incapacitation effect of first-time imprisonment: a matched samples comparison. J Quant Criminol DOI 10.1007/s10940-012-9189-3.
Zimring FE, Hawkins G (1995) Incapacitation: penal confinement and the restraint of crime. Oxford University Press, New York