Bl. Smith et Kr. Damphousse, TERRORISM, POLITICS, AND PUNISHMENT - A TEST OF STRUCTURAL-CONTEXTUALTHEORY AND THE LIBERATION HYPOTHESIS, Criminology, 36(1), 1998, pp. 67-92
The sentencing literature generally has been characterized by an inabi
lity to explain significant amounts of the variance in sentencing outc
omes. Two major theoretical explanations have addressed this issue: st
ructural-contextual theory and the ''liberation hypothesis.'' Structur
al-contextual theory suggests that the components of the justice syste
m traditionally work somewhat independently of one another. This theor
y suggests that variance explained in sentence outcomes will increase
appreciably when components function with greater interdependence-a so
-called ''tightening'' or ''coupling'' effect. Such tightening suppose
dly takes place when particular cases are given high priority for inve
stigation and prosecution. An example of this situation might be domes
tic terrorism. The liberation hypothesis suggests that the greater the
severity of an offense, the less likely judges or juries will feel fr
ee to follow their own sentiments regarding guilt and punishment. As a
consequence, the ability of legal variables to predict variation in s
entence length will be greater as crime severity increases. This study
compares a sample of officially designated terrorists matched with no
nterrorists convicted of the same federal offenses. OLS regression and
structural equation modeling procedures are used to compare the level
s of explained variance for the two groups. The results indicate stron
g support for the basic premises of both theories. Explained variance
for the terrorist sample is more than four times greater than the expl
ained variance for the nonterrorist sample. Further analysis shows tha
t explained variance is highest for terrorists who have committed a hi
gh-severity offense and lowest for nonterrorists who have committed a
low-severity offense. The subsequent addition of other predictor varia
bles available only for the terrorist sample further increases the exp
lained variance and provides additional support for the liberation hyp
othesis.