Although ordinary least squares (OLS) regression was once a common tool for modeling discrete count outcomes in criminology and criminal justice, the past several decades have seen an increasing reliance on regression techniques specifically designed for such purposes. Utilizing a practical example from the 1958 Philadelphia Birth Cohort, this article describes and compares various estimation strategies for modeling such outcome variables, including a discussion of the inappropriateness of OLS for such purposes and specific features of discrete count distributions that complicate statistical inference—overdispersion, non-independence, and excess zeros. Practical advice for selecting an appropriate modeling strategy is offered.
CameronA. C.TrivediP. K. (1986). Econometric models based on count data: Comparisons and applications of some estimators and tests. Journal of Applied Economics, 1, 29-53.
3.
CameronA. C.TrivediP. K. (1990). Regression-based tests for overdispersion in the Poisson model. Journal of Econometrics, 46, 347-364.
4.
CameronA. C.TrivediP. K. (2013). Regression of count data (2nd ed.). New York, NY: Cambridge University Press.
5.
CarkinD. M. (2014). Criminal career paths in the 1958 Philadelphia Birth Cohort: Onset, progression, and desistance (Doctoral dissertation). University of Massachusetts, Lowell.
6.
CarkinD. M.TracyP. E. (2015). Moffitt revisited: Delinquent and criminal career paths in the 1958 Philadelphia birth cohort. Journal of Law & Criminal Justice, 3, 14-39.
7.
GardnerW. P.MulveyE. P.ShawE. C. (1995). Regression analyses of counts and rates: Poisson, overdispersed Poisson, and negative binomial models. Psychological Bulletin, 118, 392-404.
8.
GreeneW. H. (1993). Econometric analysis. New York, NY: Macmillan.
9.
GurmuS. (1991). Tests for detecting overdispersion in the positive regression model. Journal of Business & Economic Statistics, 9, 215-222.
10.
GurmuS.TrivediP. K. (1992). Overdispersion tests for truncated Poisson regression models. Journal of Econometrics, 54, 347-370.
11.
HesterR.HartmanT. K. (2016). Conditional race disparities in criminal sentencing: A test of the liberation hypothesis from a non-guidelines state. Journal of Quantitative Criminology. Advance online publication. doi:10.1007/s10940-016-9283-z
12.
HilbeJ. M. (2014). Modeling count data. New York, NY: Cambridge University Press.
13.
HuberP. J. (1967). The behavior of maximum likelihood estimates under nonstandard conditions. In LecamL.NeymanJ. (Eds.), Proceedings of the fifth Berkeley symposium on mathematical statistics and probability (Vol. 1, pp. 221-233). Berkeley, CA: University of California Press.
14.
KleiberC.ZeileisA. (2008). Applied econometrics with R. New York, NY: Springer-Verlag.
15.
LambertD. (1992). Zero-inflated Poisson regression, with an application to defects in manufacturing. Technometrics, 34, 1-14.
16.
LeeL. F. (1986). Specification test for Poisson regression models. International Economic Review, 27, 689-706.
17.
LongJ. S. (1997). Regression models for categorical and limited dependent variables. Thousand Oaks, CA: Sage.
18.
MullahyJ. (1986). Speciation and testing of some modified count data models. Journal of Econometrics, 33, 341-365.
19.
ParkS. M.FisherB. S. (2015). Understanding the effect of immunity on over-dispersed criminal victimizations: Zero-inflated analysis of household victimizations in the NCVS. Crime & Delinquency. Advance online publication. doi:10.1177/0011128715607534
20.
RydbergJ.GrommonE. (2016). A multimethod examination of the dynamics of recidivism during reentry. Corrections: Policy, Practice, and Research, 1, 40-60.
21.
SturmanM. C. (1999). Multiple approaches to analyzing court data in studies of individual differences: The propensity for Type 1 errors, illustrated with the case of absenteeism prediction. Educational and Psychological Measurement, 59, 414-430.
22.
TracyP. E.Kempf-Leonard. (1996). Continuity and discontinuity in criminal careers. New York, NY: Plenum Press.
23.
TracyP. E.WolfgangM. E.FiglioR. M. (1990). Delinquency careers in two birth cohorts. New York, NY: Plenum Press.
24.
Ver HoefJ. M.BovengP. L. (2007). Quasi-Poisson vs. negative binomial regression: How should we model overdispersed count data?Ecology, 88, 2766-2772.
25.
VuongQ. H. (1989). Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica: Journal of the Econometric Society, 57, 307-333.
26.
WallaceD.EasonJ. M.LindseyA. M. (2015). The influence of incarceration and re-entry on the availability of health care organizations in Arkansas. Health & Justice, 3(3), 1-11.
27.
WhiteH. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica: Journal of the Econometric Society, 48, 817-838.
28.
William H. Fisher, Stephanie W. Hartwell, and Xiaogang Deng (2017). Managing Inflation: On the Use and Potential Misuse of Zero-Inflated Count Regression Models. Crime & Delinquency, 63, 77-87.
29.
ZeileisA. (2006). Object-oriented computation of sandwich estimators. Journal of Statistical Software, 16(9), 1-16.
30.
ZeileisA.KleiberC.JackmanS. (2008). Regression models for count data in R. Journal of Statistical Software, 27(8). Retrieved from http://www.jstatsoft.org/v27/i08/
31.
ZornC. J. (1996). Evaluating zero-inflated and hurdle Poisson specifications. Midwest Political Science Association, 18(20), 1-16.
