Weiman, 2001) and thus make unemployment more likely. The second is the limited ability of
incarcerated offenders to acquire cultural capital in the form of valued jobs skills (Waldfogel,
1994). In fact, as Irwin and Austin (1997) have pointed out, cultural behaviors and attitudes that
are adaptive for survival in prison are unlikely to be consistent with the work routines the
inmates will encounter once they have been released.
Hypothesis 5: The amount of time respondents spend incarcerated will be positively and
significantly associated with proportion of time spent inactive, and will partially mediate
the relationship between delinquency and inactivity.
Data and Methods:
The analysis is based on two samples from the National Longitudinal Survey of Youth,
spanning the years of 1979 through 2004. Both of these surveys were administered annually
from 1979 to 1994, and then every two years from 1996 to 2004. The first sample was designed
to be a representative cross-section of youth in the United States, and the second was designed to
oversample low-income and minority youth. However, when the data are adjusted using a
number of stratification and weight variables, they can be used in tandem to examine patterns
among youth as a whole. In our analysis, we limit our sample to males aged 14-21 at the base
year of the study. The dependent variable in our analysis (proportion of time spend inactive) will
be measured by examining the total number of years each respondent spent unemployed, out of
school, and not incarcerated during the 25 year period of the study. We then calculate the natural
log of the dependent variable, since it is highly skewed.
Our analysis involves several steps. First, we run several frequencies to demonstrate that
our sample roughly approximates the demographic characteristics of males in the United States.
10
Convention