Selective Attrition and the Age-Crime Relationship

Robert Brame, Alex R. Piquero

Research output: Contribution to journalArticleResearchpeer-review

58 Citations (Scopus)


One of the most widely accepted findings in criminology is the strong curvilinear association between age and crime. Studies have indicated that involvement in criminal behavior rises throughout the teenage years until it levels off during the late teenage and early adult years and then declines throughout the remainder of the life span. Many of these studies, however, have relied on cross-sectional or repeated cross-sectional officially recorded data. In recent years, researchers investigating changes in self-reported criminal behavior among the same individuals as they age have discovered that involvement in criminal behavior actually declines during the adolescent years. Although there are a number of possible explanations for this phenomenon, we explore the extent to which this decline can be attributed to selective or non-random panel attrition.

Original languageEnglish
Pages (from-to)107-127
Number of pages21
JournalJournal of Quantitative Criminology
Issue number2
Publication statusPublished - 1 Jun 2003
Externally publishedYes


  • Age
  • Attrition
  • Censored bivariate probit
  • Crime
  • Identification problems
  • Missing data

Cite this