Abstract
Offending specialization continues to be a subject of empirical inquiry for scholars interested in criminal careers. Early research consistently spoke to the generality of offending profiles, but more recent work has revealed somewhat mixed findings. These results have emerged alongside newly developed and applied methods that detect and describe offending specialization. To what extent these methods shape divergent conclusions and/or provide overlapping insight remains unclear, however. Therefore, the degree to which more recent inquiries are actually studying the same operational definition of specialization is unknown. In order to consider this issue further, this study utilizes four frequently applied approaches with a single data set. The study indicates when and where findings converge and also describes any unique insights provided by each method. The work concludes with a discussion surrounding the utility of applying multiple strategies in assessing specialization in criminal offending.
Similar content being viewed by others
Notes
The fact that previous work has not explicitly compared a number of these methods might partly be due to data constraints. As forthcoming sections will detail, each of the methods requires data with certain attributes, such as knowledge of how offenses are sequenced, ample indicators of crime types (e.g., a number of items for violence and non-violence), and a healthy amount of criminal activity from which to derive patterns. Very few data sets meet these requirements.
Osgood and Schreck (2007) point out that the flexible coding available in the Multilevel IRT method would allow for categorization of multiple different types of offenses (e.g., violent, nonviolent, drug). Still, this would necessitate knowledge of anticipated categorizations ahead of time.
Such a circumstance, in which the relative and objective views of specialization are inconsistent, is at this point speculative. Thus far, analyses using this approach have not encountered this problem. Even so, this method is still emerging (thus, it has not been used across many data sets) and this divergence is a possible outcome under the modeling strategy.
Some may assert that if certain cases are excluded for particular analyses, then they should be excluded across all analyses for comparative purposes. The purpose of this study is to evaluate the relative benefits and drawbacks of these four methods, however. If the fact that one method does not exclude certain cases is one of its purported benefits, then it is appropriate to study the potential impact on the results. For instance, a forthcoming section describes how a hierarchy rule is often invoked with the FSC; in this way, it also excludes data. This is inherent in using these methods and researchers should consider this when assessing their utility.
It is possible that this choice of time window could impact the level of specialization/diversity uncovered in this study as specialization has been found to vary with age (e.g., Piquero 2000). The major emphasis of the current work is on comparing techniques, however, so this difficulty is attenuated somewhat in its impact on our objectives.
The category for “other offenses” was excluded because it largely represented administrative offenses.
Like others (e.g., Osgood and Schreck 2007) we believe that the selection of the most serious offense for a given event may remove interesting information from the analysis, but apply these methods in a way that is consistent with the manner in which they have been used previously.
Although the structure of the data make it somewhat difficult to fully and unambiguously assess the effects of utilizing only the most serious offense in calculating the FSC, we looked at a 10% random subsample of the 1,206 cases in these analyses to assess the extent to which these individuals were charged with more than one offense at a time. Across the time window of focus, individuals had roughly two offenses (mean = 2.16, SD = 2.74), on average, that were not captured in the construction of the transition matrices upon which the FSC is based. This indicates that utilizing only the most serious arrest/charge in a given incident may lead to some information loss in assessing specialization and diversity in offending.
The negative value for arson (−0.003) suggests that the overwhelming tendency is that arson will be followed by a dissimilar offense. This is not surprising—given the overall prevalence of that particular offense type in the data. Britt (1996), Farrington et al. (1988) and Paternoster et al. (1998) have previously discussed the nature of such findings in detail.
For reference, descriptive probabilities for various offense types based on the model estimates are also shown.
Osgood and Schreck (2007) mention that their method might be expanded to allow for more than one type of specialization (i.e., categorical rather than binary). This would still require that the list of possible offense items be reduced to broader categories, however, therefore losing some information.
As noted above, one limitation of the FSC is the fact that it is an aggregate measure, which may preclude its use in many theoretical tests for technical reasons. Additionally, this emphasis on theory requires that greater consideration be given to those methods that can seamlessly incorporate covariates and provide unbiased and efficient estimates of their impacts on specialization. Depending on the specific question, this suggests the utility of the diversity index, LCA, or IRT or those models developed by Britt (1996) and Deane et al. (2005).
References
Agresti A, Agresti BF (1978) Statistical analysis of qualitative variation. Sociol Methodol 9:204–237
Armstrong T (2008) Exploring the impact of changes in group composition on trends in specialization. Crime Delinq 54:366–389
Blumstein A, Cohen J, Roth JA, Visher C (1986) Criminal careers and career criminals, vol 1. National Academy Press, Washington, DC
Blumstein A, Cohen J, Das S, Moitra SD (1988) Specialization and seriousness during adult criminal careers. J Quant Criminol 4:303–345
Bouffard LA, Wright KA, Mufti LA, Bouffard JA (2008) Gender differences in specialization in intimate partner violence: comparing the gender symmetry and violent resistant perspectives. Justice Q 25:570–594
Brame R, Nagin DS, Wasserman L (2006) Exploring some analytic characteristics of finite mixture models. J Quant Criminol 22:31–59
Brennan P, Mednick S, John R (1989) Specialization in violence: evidence of a criminal subgroup. Criminology 27:437–453
Britt CL (1994) Versatility. In: Hirschi T, Gottfredson M (eds) The generality of deviance. Transaction, New Brunswick, pp 173–192
Britt CL (1996) The measurement of specialization and escalation in the criminal career: an alternate modeling strategy. J Quant Criminol 12:193–222
Bursik RJ (1980) The dynamics of specialization in juvenile offenses. Soc Forces 58:851–864
Bushway S, Brame R, Paternoster R (1999) Assessing stability and change in criminal offending: a comparison of random effects, semiparametric, and fixed effects modeling strategies. J Quant Criminol 15:23–61
Bushway S, Sweeten G, Wilson DB (2006) Size matters: standard errors in the application of null hypothesis significance testing in criminology and criminal justice. J Exp Criminol 2:1–22
Clogg CC (1995) Latent class models. In: Arminger G, Clogg CC, Sobel ME (eds) Handbook of statistical modeling for the social and behavioral sciences. Plenum Press, New York, pp 311–359
Cook TD, Campbell DT (1979) Quasi-experimentation: design and analytical issues for field settings. Houghton Mifflin, Boston
D’Unger A, Land K, McCall P, Nagin DS (1998) How many latent classes of delinquent/criminal careers? Results from mixed poisson regression analyses of the London, Philadelphia, and Racine cohorts studies. Am J Sociol 103:1593–1630
Deane G, Armstrong DP, Felson RB (2005) An examination of offense specialization using marginal logit models. Criminology 43:955–988
DeLisi M (2003) The imprisoned non-violent drug offender: specialized martyr or versatile career criminal? Am J Crim Justice 27:167–182
DeLisi M (2005) Career criminals in society. Sage Publications, Thousand Oaks
Denzin NK (1989) The research act: a theoretical introduction to sociological methods, 3rd edn. Prentice Hall, Englewood Cliffs
Farrington DP (1986) Age and crime. In: Tonry M, Morris N (eds) Crime and justice: a review of research, vol 7. University of Chicago Press, Chicago, pp 189–250
Farrington DP, Snyder HN, Finnegan TA (1988) Specialization in juvenile court careers. Criminology 26:461–485
Felson M (1998) Crime and everyday life, 2nd edn. Sage, Thousand Oaks
Felson M (2006) Crime and nature. Sage, Thousand Oaks
Flick UEW (1992) Triangulation revisited: strategy of validation or alternative? J Theory Soc Behav 22:175–197
Francis B, Soothill K, Fligelstone R (2004) Identifying patterns and pathways of offending behaviour: a new approach to typologies of crime. Eur J Criminol 1:47–87
Garner J, Maxwell C (2008) The crime control effects of prosecuting intimate partner violence in Hamilton County, Ohio: reproducing and extending the analyses of Wooldredge and Thistlethwaite. Department of Justice, Washington, DC
Haapanen RA (1990) Selective incapacitation and the serious offender: a longitudinal study of criminal career patterns. Springer, New York
Haapanen RA (1999) Longitudinal study of criminal career patterns of former California youth authority wards, 1965–1984. Interuniversity Consortium for Political and Social Research, Ann Arbor
Kempf KL (1987) Specialization and the criminal career. Criminology 25:399–420
Lo Y, Mendell NR, Rubin DB (2001) Testing the number of components in a normal mixture. Biometrika 88:767–778
Lombroso C (1911) Crime: its causes and remedies. Little, Brown and Company, Boston
Lussier P (2005) The criminal activity of sexual offenders in adulthood: revisiting the specialization debate. Sex Abuse 17:269–292
Lynam D, Piquero AR, Moffitt TE (2004) Specialization and the propensity to violence: support from self-reports but not official records. J Contemp Crim Justice 20:215–228
Mathison S (1988) Why triangulate? Educ Res 17:13–17
Mazerolle P, Brame R, Paternoster R, Piquero AR, Dean C (2000) Onset age, persistence, and offending versatility: comparisons across gender. Criminology 38:1143–1172
McCutcheon AL (1987) Latent class analysis. Sage, Newbury Park
McGloin JM, Sullivan CJ, Piquero AR, Pratt TC (2007) Explaining qualitative change in offending: revisiting specialization in the short-term. J Res Crime Delinq 44:321–346
McGloin JM, Sullivan CJ, Piquero AR (2009) Aggregating to versatility? Transitions among offender types in the short-term. Br J Criminol 49:243–264
Miethe T, Olson J, Mitchell O (2006) Specialization and persistence in the arrest histories of sex offenders: a comparative analysis of alternate measures and offense types. J Res Crime Delinq 43:204–229
Muthén BO (1998–2008) MPlus technical appendices. Muthén and Muthén, Los Angeles
Muthén BO (2002) Beyond SEM: general latent variable modeling. Behaviormetrika 29:81–117
Muthén BO, Muthén LK (1998–2006) Mplus user’s guide, 4th edn. Muthén and Muthén, Los Angeles
Nagin DS (2005) Group-based modeling of development. Harvard University Press, Cambridge
Nagin DS, Land K (1993) Age, criminal careers, and population heterogeneity: specification and estimation of a nonparametric, mixed poisson model. Criminology 31:327–362
Nagin DS, Tremblay RE (2005) Developmental trajectory groups: fact or a useful statistical fiction? Criminology 43:873–904
Nylund KL, Asparouhov T, Muthén BO (2007) Deciding on the number of classes in latent class analysis and growth mixture modeling. A Monte Carlo simulation study. Struct Eq Model 14:535–569
Osgood DW (2005) Making sense of crime and the life-course. Ann Am Acad Pol Soc Sci 602:196–211
Osgood DW, Schreck CJ (2007) A new method for studying the extent, stability, and predictors of individual specialization in violence. Criminology 45:273–312
Osgood DW, Johnston LD, O’Malley PM, Bachman JG (1988) The generality of deviance in late adolescence and early adulthood. Am Sociol Rev 53:1–93
Osgood DW, Finken LW, McMorris B (2002) Analyzing multiple-item measures of crime and deviance II: tobit regression analysis of transformed scores. J Quant Criminol 18:319–347
Paternoster R, Brame R, Piquero AR, Mazerolle P, Dean CW (1998) The forward specialization coefficient: distributional properties and subgroup differences. J Quant Criminol 14:133–154
Piquero AR (2000) Frequency, specialization, and violence in offending careers. J Res Crime Delinq 37:392–418
Piquero AR (2008) Taking stock of developmental trajectories of criminal activity over the life course. In: Liberman A (ed) The long view of crime: a synthesis of longitudinal research. Springer, New York, pp 23–78
Piquero AR, Paternoster R, Brame R, Mazerolle P, Dean CW (1999) Onset age and specialization in offending behavior. J Res Crime Delinq 36:235–274
Piquero AR, Blumstein A, Brame R, Haapanen RA, Mulvey EP, Nagin DS (2001) Assessing the impact of exposure time and incapacitation on longitudinal trajectories of criminal offending. J Adolesc Res 16:54–71
Piquero AR, Farrington DP, Blumstein A (2003) The criminal career paradigm. In: Tonry Michael (ed) Crime and justice: a review of research, vol 30. University of Chicago Press, Chicago, pp 359–506
Raudenbush SW, Johnson C, Sampson RJ (2003) A multivariate, multilevel Rasch model with application to self-reported criminal behavior. Sociol Methodol 33:169–211
Rojek DG, Erickson ML (1982) Delinquent careers: a test of the career escalation model. Criminology 20:5–28
Sampson RJ, Laub JH (2005) Seductions of method: Rejoinder to Nagin and Tremblay. Criminology 43:905–913
Schreck C, Stewart E, Osgood DW (2008) A reappraisal of the overlap of violent offenders and victims. Criminology 46:871–906
Schreck C, McGloin JM, Kirk D (2009) On the origins of the violent neighborhood: a study of the nature and predictors of crime-type differentiation across Chicago neighborhoods. Justice Q (forthcoming)
Simon LMJ (1997) Do offenders specialize in crime types? Appl Prev Psychol 6:35–53
Soothill K, Francis B, Fligelstone R (2002) Patterns of offending behaviour: a new approach. Research findings 171, full report. Home Office, London. http://www.homeoffice.gov.uk/rds/pdfs2/patternsrevisedr171.pdf
Soothill K, Francis B, Ackerley E, Humphreys L (2008) Changing patterns of offending behaviour among young adults. Br J Criminol 48:75–95
Sullivan CJ, McGloin JM, Pratt TC, Piquero AR (2006) Rethinking the “norm” of offender generality: investigating specialization in the short-term. Criminology 44:205–240
Sullivan CJ, McGloin JM, Piquero AR (2008) Treating the deviant y* in criminology: an examination of the assumptions of censored normal regression and some potential alternatives. J Quant Criminol 24:399–421
Tracy PE, Kempf-Leonard K (1996) Continuity and discontinuity in criminal careers. Plenum, New York
Vermunt JK, Magidson J (2004) Latent class analysis. In: Lewis-Beck MS, Bryman A, Liao FT (eds) The sage encyclopedia of social science research methods. Sage, Newbury Park, pp 549–553
Warr M (1988) Rape, burglary and opportunity. J Quant Criminol 4:275–288
Wolfgang M, Figlio R, Sellin T (1972) Delinquency in a birth cohort. University of Chicago Press, Chicago
Wright K, Pratt TC, DeLisi M (2008) Examining offending specialization in a sample of male homicide offenders. Homicide Stud 12:381–398
Acknowledgments
The authors thank Dr. Matt DeLisi. Dr. Wayne Osgood, and Dr. Christopher Schreck for their helpful comments on earlier drafts of this manuscript. They also appreciate the constructive suggestions offered by the anonymous reviewers and editors.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Rights and permissions
About this article
Cite this article
Sullivan, C.J., McGloin, J.M., Ray, J.V. et al. Detecting Specialization in Offending: Comparing Analytic Approaches. J Quant Criminol 25, 419–441 (2009). https://doi.org/10.1007/s10940-009-9074-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10940-009-9074-x