Academic literature on the topic 'Actuarial scales'

Actuarial scales like the Static-99R are widely used to predict an individual’s risk of sexual recidivism. However, current actuarial scales only provide rates of detected sex offenses over 10-year follow-up and do not account for all recidivism risk factors. Therefore, some forensic evaluators extrapolate, adjust, or override recidivism rates derived from actuarial scales to predict the lifetime risk of committed offenses that accounts for external risk factors, those not addressed by the actuarial scales. However, critics contend that altering rates from actuarial scales degrades their predictive validity. This article makes the case for extrapolating risk for time of exposure and for evidence-based external risk factors. It proposes using odds ratios (ORs) from case-control studies to adjust predictions from follow-up cohort studies. Finally, it shows how evaluators can apply ORs and their margins of error to sex offender risk assessment.

Assessing the predictive accuracy of actuarial risk assessment tools requires consideration of discrimination (the differences between recidivists and nonrecidivists) and calibration (the credibility of the recidivism rates associated with test scores or categories). Currently, there are no conventions for reporting calibration effect sizes for offender risk tools. This article explains one promising calibration effect size statistic (the Expected/Observed [E/O] index) and provides an illustrative example of how it can be calculated and interpreted. Briefly, the E/O index is the ratio of the expected number of recidivists to the observed number of recidivists. Guidance is provided for calculating the E/O index with fixed follow-up data as well as from survival data. This article also discusses alternative approaches to examining calibration and provides references to other studies using the E/O index to assess the calibration of offender risk scales.

Correlates of prison violence and the classification accuracy of an actuarial model were examined from retrospective review of the disciplinary records of former death row inmates in Texas ( N = 111) who had been predicted to commit future violence at trial and subsequently obtained relief from their death sentences between 1989 and 2008. Correlates of “potentially” violent infractions included age (inversely), intellectual capability (inversely), prior violent crime arrest, and gun-only weapon used in murder (inversely). An actuarial scale constructed from the sample was modestly (area under the curve [AUC] = 0.690) associated with combined violence on death row and in the broader prison population, as were scales constructed previously from other samples (AUC = 0.609 to 0.656). Although AUCs for serious assaults in three models were relatively high (AUC = 0.799 to 0.831), low base rates result in these actuarial scales having high false positive rates (e.g., 76%) in identifying offenders who will commit serious prison assaults.

In the UK the majority of life insurers employ relatively liberal underwriting standards so that people can easily gain access to life assurance cover. Up to 95%of applicants are accepted at standard terms. If genetic testing becomes widespread then the buying habits of the public may change. Proportionately more people with a predisposition to major types of disease may take life assurance cover while people with no predisposition may take proportionately less. A model is used to show the possible effect. However, the time–scales are long and the mortality of assured people is steadily improving. The change in buying habits may result in the rate of improvement slowing down. In the whole population, the improvement in mortality is likely to continue and could improve faster if widespread genetic testing results in earlier diagnosis and treatment. Life insurers would not call for genetic tests and need not see the results of previous tests except for very large sums assured. In the UK, life insurers are unlikely to change their underwriting standards and are extremely unlikely to bring in basic premium rating systems that give discounts on the premium or penalty points according to peoples' genetic profile. The implications of widespread genetic testing on medical insurance and some health insurance covers may be more extreme.

This article explores the predictive validity of two actuarial risk assessment instruments among mentally disordered offenders in Sweden: the historical part (H-10) of a historical, clinical, and risk management factors instrument (HCR-20) and the Violence Risk Appraisal Guide (VRAG). Actuarial scores were obtained retrospectively in two populations: one group of violent offenders with personality disorders and one with violent offenders diagnosed with schizophrenia. The predictive accuracy was evaluated with receiver operating characteristic analysis using a violent reconviction within 2 years from release or discharge as the criterion variable. Both scales fared better in the group of personality-disordered offenders than in the group of offenders with schizophrenia, and the H-10 fared better than the VRAG in both offender groups. The study found that historical data maintain a robust predictive validity in a population of personality-disordered offenders, whereas clinical and risk management factors may be of greater importance in offender populations in which major mental disorders are prevalent.

This research explored empirical dimensions of sex offender recidivism risk. Study 1 portrayed descriptive statistics and factor structure information concerning actuarial risk instruments and diagnoses derived from a sample of sex offenders being evaluated for civil commitment in Wisconsin. Study 2 used a sample from England and Wales to analyze the relationships between individual risk factors commonly found as items within actuarial scales. Factor structure results from Study 2 conceptually overlapped those found in the first sample, and variables developed from this factor structure predicted sexual reconviction as well. Results from these two studies are discussed in terms of separable components of risk for sexual recidivism and the roles those components may play in processes underlying sexual reoffense.

1. Salary scales have been widely used in actuarial literature about pension schemes, but they do not seem to have been developed beyond the idea first introduced by Manly (1901) and used in a series of papers following this, including McGowan (1901), Manly (1902, 1903 and 1911), and M'Lauchlan (1908). King (1905), Bacon (1907) and M'Lauchlan (1914) discuss the construction of a salary scale from records of individual employees. King made some valuable observations on how a salary scale may change with time if the observed population is not a stationary one, for example, because the firm is growing or declining, which Bacon also commented on, and M'Lauchlan went into considerable detail about the separation of different grades. Thomas (1913) gave an example of an organization with six ranks, within each of which there was a salary scale, and showed explicit probabilities of promotion in each year of age. His development comes closest to what I shall discuss below. Text books on Life Contingencies, such as Jordan (1952), Hooker & Longley-Cook (1957) and most recently Neill (1977), have followed essentially the definition introduced by Manly, as also have papers and text books on pension funds, such as Porteous (1936), Marples (1948), Heywood & Marples (1950), Crabbe & Poyser (1953) and Lee (1973). Curiously Spurgeon (1922) does not mention salary scales, although his book was written after they had come into use.

Recent research has demonstrated that antisocial attitudes and antisocial associates are among the better predictors of antisocial behavior. This study tests the predictive validity of the Measures of Criminal Attitudes and Associates (MCAA) in a sample of adult male offenders. The MCAA comprises two parts: Part A is a quantified self-report measure of criminal friends, and Part B contains four attitude scales: Violence, Entitlement, Antisocial Intent, and Associates. The MCAA scales showed predictive validity for the outcomes of general and violent recidivism. In addition, the MCAA significantly improved the prediction of violent recidivism over an actuarial risk assessment instrument alone. Discussion centers on the contribution that antisocial attitudes and associates make to risk assessment.

Includes abstract.
Includes bibliographical references.
This paper investigates the relationship between scale and efficiency through the application of Data Envelopment Analysis (DEA) to a set of South African private hospitals over the three year period from 2007 to 2009. As part of the investigation, this paper provides a description of the current research into scale and efficiency with a focus on definition and measurement. It also provides an introduction to DEA as a tool for measuring the relationship between hospital scale and efficiency. Based on the underlying set of private hospitals, this investigation found that scale efficiency improvements are likely to be possible.

Thesis (Ph. D.) -- University of Maryland, College Park, 2007.
Thesis research directed by: Criminology and Criminal Justice. Title from t.p. of PDF. Includes bibliographical references. Published by UMI Dissertation Services, Ann Arbor, Mich. Also available in paper.

Cette thèse a été faite sous la tutelle de Jean Proulx et R. Karl Hanson. Elle a évaluée par un jury composé de Franca Cortoni, Jean Proulx, R. Karl Hanson, Jean-Pierre Guay et Howard E. Barbaree. Suite à la soutenance, la thèse a reçu la mention "exceptionnelle", a été recommandée à la liste d'honneur du doyen et a été soumise pour le prix de la meilleure thèse de la FESP 2017.
Les outils actuariels servant à évaluer le risque de récidive criminelle des agresseurs sexuels ont souvent été critiqués pour leurs fondements « athéoriques ». En effet, ces derniers ont été constitués en rassemblant les caractéristiques les plus fortement associées à la récidive, sans qu’une théorie ne les unisse à priori. Si cette méthode a assuré une bonne validité prédictive à ces instruments, elle leur a insufflé très peu de résonnance clinique; plusieurs professionnels n’y voient qu’une liste de corrélats statistiques vides de sens. La présente thèse a entrepris de relativiser ces critiques en appliquant des modèles de facteurs latents aux différents items des outils actuariels. Les modèles de facteurs latents postulent que les comportements manifestes (observables) d’un individu renseignent sur ses caractéristiques psychologiques latentes (non observables) – de la même manière que les symptômes d’une maladie permettent d’inférer sa présence chez un patient. Puisque les items des outils actuariels correspondent à divers comportements ou caractéristiques criminogènes manifestes, il devrait être possible de les rassembler en dimensions pour identifier les principaux construits psychologiques latents associés à la récidive des agresseurs sexuels. Les articles de la présente thèse ont donc proposé d’identifier les dimensions de la Statique-99R et de la Statique-2002R, la famille d’outils actuariels pour agresseurs sexuels la plus utilisée mondialement. Trois dimensions ont été extraites par analyse factorielle : 1) la persistance dans les délits sexuels/paraphilies sexuelles, 2) la délinquance générale et 3) le jeune âge/victimes sexuelles non familières. La première dimension était exclusivement associée à la récidive sexuelle, tandis que les deux dernières étaient associées à tous les types de récidives. Leur validité convergente a ensuite été explorée afin d’évaluer à quelles caractéristiques psychologiques ces dernières référaient. La persistance/paraphilie a convergé avec des indicateurs d’intérêts sexuels déviants non coercitifs, alors que la délinquance générale a convergé avec une constellation de traits antisociaux. Finalement, le jeune âge/victimes sexuelles non familières a convergé avec l’intention claire de blesser ses victimes. Ces résultats ont mené au développement d’un modèle tridimensionnel du risque de récidive des agresseurs sexuels ayant plusieurs implications pratiques pour le domaine. D’une part, les outils actuariels gagneraient à utiliser des scores dimensionnels plutôt que des scores totaux. Par exemple, en retirant la dimension de persistance/paraphilie de la Statique-99R/2002R, il est possible de prédire efficacement la récidive violente non sexuelle et non sexuelle non violente des agresseurs sexuels, ce que l’instrument ne permettait pas précédemment. D’autre part, la connaissance des grandes dimensions de facteurs de risque est susceptible d’aider les évaluateurs à choisir et intégrer plusieurs mesures actuarielles. Cela se révèle particulièrement pertinent lorsque deux mesures du risque divergent, malgré qu’elles aient été conçues pour la même population. Finalement, nos résultats indiquent que les facteurs de risque statiques sont susceptibles de renseigner (imparfaitement) la pratique clinique quant aux besoins criminogènes des agresseurs sexuels. Bien que nous ne suggérions aucunement de substituer cette pratique à la cotation d’outils actuariels de troisième génération, elle pourrait se révéler pertinente pour les établissements n’ayant pas les ressources nécessaires pour coter de tels instruments. Au plan théorique, le modèle tridimensionnel comporte plusieurs avantages par rapport aux modèles à deux dimensions, traditionnellement constitués de la déviance sexuelle et de la délinquance générale/psychopathie (ex. : Doren, 2004). D’une part, les données empiriques supportent clairement la présence de trois dimensions du risque de récidive des agresseurs sexuels, et non deux. D’autre part, la troisième dimension permet d’intégrer une nuance importante au modèle, soit la distinction entre les intérêts sexuels déviants par leur objet de désir (ex. : pédophilie) et les intérêts sexuels déviants par leur aspect coercitif (ex. : sadisme sexuel). Cette distinction est primordiale, dans la mesure où ces intérêts sexuels ne prédisent pas les mêmes types de récidives. Finalement, notre modèle s’intègre au modèle étiologique du risque de Beech et Ward (2004), qui, lorsque complètement validé, pourra donner lieu à des efforts de prévention – une denrée rare dans le domaine.
Actuarial scales for the prediction of sexual violence have been criticised because they are purely based on atheoretical correlates. Therefore, they are unlikely to provide clinical and theoretical insight on the psychological traits and mechanisms that underlie criminal recidivism. However, according to latent variable models commonly used in psychology, patterns of behavior, thought, and emotion are caused by latent psychological constructs, such as extraversion and neuroticism. Because static and stable risk factors in actuarial scales are mostly behavioral, it should be possible to use them to infer the major psychological constructs responsible for recidivism risk. The current thesis applied latent variable models to nonredundant items from the Static-99R and Static-2002R, the two most commonly used risk tools for sexual offenders. Three dimensions were identified: 1) persistence in sexual crimes/paraphilia, 2) general criminality, and 3) youthful stranger aggression. To understand the psychological meaning of these dimensions, convergent and predictive validity analyses were conducted. Results indicated that persistence/paraphilia was related to dysregulation of sexuality towards atypical objects, without intent to harm, while general criminality was related to antisocial traits. Finally, youthful stranger aggression was related to a clear intent to harm victims. All three constructs predicted sexual recidivism with similar accuracy, but only general criminality and youthful stranger aggression predicted nonsexual recidivism. These results suggest that a tridimensional model of sexual offender risk is viable. That model has numerous practical implications. First, actuarial scales should sort items by constructs rather than rely on total scores. Total scores focus the predictive utility of risk scales to the specific outcome for which they were developed (usually sexual recidivism). When constructs are known, it is possible to improve the prediction of other outcomes by removing constructs unrelated to each of these new outcomes (e.g., removing sexual criminality items to improve the prediction of nonsexual recidivism). Second, construct-level approaches facilitate the integration of potentially conflicting risk scales. By understanding the constructs assessed by each scale, an evaluator can deduce which measures should be combined, and which should not. Finally, static risk constructs significantly correlate with psychological features that are found in dynamic risk scales. Consequently, scales composed entirely of static risk factors could – albeit imperfectly – inform the treatment needs of sexual offenders. These results could assist settings lacking the resources to implement dynamic risk tools. The tridimensional model also has theoretical implications. First, our model showed better statistical fit than classical two factor models based on sexual deviance and psychopathy (e.g., Doren, 2004), suggesting that there are more than two substantive dimensions related to sexual offender recidivism risk. Second, the presence of a third factor enabled an important distinction between noncoercive (e.g., fixated pedophilia) and coercive (e.g., sexual sadism) deviant sexual interests. Such a distinction is particularly relevant in the field of risk assessment, because they do not predict the same types of recidivism. Finally, our model can be integrated in Beech and Ward’s (2004) etiological model of risk, which, once fully validated, could enable primary and secondary prevention efforts.

In insurance risk theory, the surplus of an insurance company is modelled to monitor and quantify its risks. With the outgo of claims and inflow of premiums, the insurer needs to determine what financial portfolio ensures the soundness of the company’s future while satisfying the shareholders’ interests. It is usually assumed that the net profit condition (i.e. the expectation of the process is positive) is satisfied, which then implies that this process would drift towards infinity. To correct this unrealistic behaviour, the surplus process was modified to include the payout of dividends until the time of ruin. Under this more realistic surplus process, a topic of growing interest is determining which dividend strategy is optimal, where optimality is in the sense of maximizing the expected present value of dividend payments. This problem dates back to the work of Bruno De Finetti (1957) where it was shown that if the surplus process is modelled as a random walk with ± 1 step sizes, the optimal dividend payment strategy is a barrier strategy. Such a strategy pays as dividends any excess of the surplus above some threshold. Since then, other examples where a barrier strategy is optimal include the Brownian motion model (Gerber and Shiu (2004)) and the compound Poisson process model with exponential claims (Gerber and Shiu (2006)). In this thesis, we focus on the optimality of a barrier strategy in the more general Lévy risk models. The risk process will be formulated as a spectrally negative Lévy process, a continuous-time stochastic process with stationary increments which provides an extension of the classical Cramér-Lundberg model. This includes the Brownian and the compound Poisson risk processes as special cases. In this setting, results are expressed in terms of “scale functions”, a family of functions known only through their Laplace transform. In Loeffen (2008), we can find a sufficient condition on the jump distribution of the process for a barrier strategy to be optimal. This condition was then improved upon by Loeffen and Renaud (2010) while considering a more general control problem. The first chapter provides a brief review of theory of spectrally negative Lévy processes and scale functions. In chapter 2, we define the optimal dividends problem and provide existing results in the literature. When the surplus process is given by the Cramér-Lundberg process with a Brownian motion component, we provide a sufficient condition on the parameters of this process for the optimality of a dividend barrier strategy. Chapter 3 focuses on the case when the claims distribution is given by a univariate mixture of Erlang distributions with a common scale parameter. Analytical results for the Value-at-Risk and Tail-Value-at-Risk, and the Euler risk contribution to the Conditional Tail Expectation are provided. Additionally, we give some results for the scale function and the optimal dividends problem. In the final chapter, we propose an expectation maximization (EM) algorithm similar to that in Lee and Lin (2009) for fitting the univariate distribution to data. This algorithm is implemented and numerical results on the goodness of fit to sample data and on the optimal dividends problem are presented.