Showing papers by "David P. Farrington published in 1989"

Early predictors of adolescent aggression and adult violence

Abstract: The Cambridge Study in Delinquent Development is a prospective longitudinal survey of 411 London males from ages 8 years old to 32 years old. This article investigates the prediction of adolescent aggression (ages 12-14 years old), teenage violence (ages 16-18 years old), adult violence (age 32 years old), and convictions for violence. Generally, the best predictors were measures of economic deprivation, family criminality, poor child-rearing, school failure, hyperactivity-impulsivity-attention deficit, and antisocial child behavior. Similar predictors applied to all four measures of aggression and violence. It is concluded that aggression and violence are elements of a more general antisocial tendency, and that the predictors of aggression and violence are similar to the predictors of antisocial and criminal behavior in general.

Development of a New Measure of Self-Reported Antisocial Behavior for Young Children: Prevalence and Reliability

Abstract: Delinquent careers, typically, have been thought to begin in adolescence (ages 12 to 16) rather than during late childhood (7 to 11). The commission of offenses by elementary school aged youngsters, however, has attracted more attention from criminologists in recent years, which has resulted in some new insights. First, offenders now are said to engage in delinquent acts at an earlier age than was previously thought. Second, an early onset of offending, according to some studies, is predictive not only of longer offending careers but also of higher rates of offending during the career. Third, surveys show that a surprising proportion of elementary school age children commit offenses that do not lead to criminal prosecution, and thus do not show up in crime statistics (Loeber, 1987).

Self-Reported and Official Offending from Adolescence to Adulthood

Abstract: The research reported in this chapter had three major objectives. The first was to study the prevalence of different types of offending, as measured by self-reports and official convictions, of the same sample at different ages from adolescence to adulthood. The second was to investigate the relationship between self-reported and official offending of the same sample at different ages, and hence to advance knowledge about the concurrent and predictive validity of self-reports in comparison with official records. The third was to study the relationship between offending at one age and offending at another, as measured by both self-reports and official convictions, and hence to advance knowledge about the issues of continuity or discontinuity and specialization or versatility in offending.

Markov chain analysis and specialization in criminal careers

Abstract: This paper investigates whether the longitudinal sequences of types of offences committed in criminal careers can be treated as a stationary first-order Markov chain. The sample consisted of 698 English male prisoners who had each been convicted of at least two standard list offences. In agreement with the Markov chain hypothesis, the probability of switching from one offence to another remained constant over successive convictions. However, contrary to this hypothesis, the past history of types of offences helped in predicting future types of offences, suggesting that there was some specialization in offending. The forward specialization coefficient confirmed and quantified this. Sex offenders were the most specialized, and the most persistent offenders became increasingly specialized in fraud. A study of the extent to which one sentence followed another revealed the stepping-stone nature of the sentencing tariff.

A prospective test of a criminal career model

Abstract: In an earlier article in this journal, Barnett, Blumstein, and Farrington (1987) formulated a model that described the criminal careers of the multiple offenders in a cohort of London males that had been studied from their 10th to 25th birthdays. That model involved two subpopulations of offenders (denoted as “frequents” and “occasionals”), each characterized by a constant annual conviction rate (μ) and a constant probability (p) of terminating the career following a conviction. This article describes the results of a prospective and predictive test of the model using new data collected on the same offenders from their 25th to 30th birthdays. The original model accurately predicted the number of recidivists, the degree of recidivism risk, the total number of recidivist convictions, and the time intervals between recidivist convictions. However, the predictions for the frequents suffered some distortions introduced by a few “intermittent” offenders who seemed to have terminated their careers, but who re-initiated offending during the test period after a long gap.

Relative improvement over chance (RIOC) and phi as measures of predictive efficiency and strength of association in 2×2 tables

Abstract: In a 2×2 table, phi measures the strength of the relationship and relative improvement over chance (RIOC) measures the predictive efficiency. A simplified formula for RIOC is derived. From the linkage between RIOC and the adjusted standardized residual, the variance of RIOC is calculated, thereby allowing its statistical significance to be tested. A disadvantage of phi is that its maximum value can be less than 1. A corrected phi correlation, obtained by dividing phi by the maximum possible phi, is shown to be mathematically identical to RIOC. Hence, there is a linkage between the percentage of correct predictions and the percentage of variance explained. RIOC is unchanged even when the social costs and benefits of the different outcomes are differentially weighted. RIOC is also mathematically identical to a corrected kappa calculated in a similar fashion to the corrected phi. Therefore, RIQC has distinct advantages over phi and kappa.

Long-Term Prediction of Offending and Other Life Outcomes

Abstract: Many of the best predictors of juvenile offending are well known. Farrington (1987 a, b) reviewed numerous research projects showing the predictive power of: (1) early troublesome, dishonest, aggressive, or antisocial behavior; (2) poor parental child-rearing methods, such as cruel, passive, or neglecting attitudes, harsh or erratic discipline, and poor supervision; (3) criminal parents and delinquent siblings; (4) broken homes and early separations caused by divorce or parental conflict; (5) social deprivation, as reflected in low family income, large family size, poor housing, and an erratic parental employment record; and (6) school failure, as indexed by low intelligence, poor educational achievement, and truancy.