Prevention Science [PREV] PP247-344677 August 10, 2001 7:44 Style ﬁle version Nov. 04, 2000
Prevention Science, Vol. 2, No. 3, September 2001 (
Temperament Related to Early-Onset Substance Use:
Test of A Developmental Model
Thomas A. Wills,
and Karen Spera
We tested a theoretical model of early-onset substance (tobacco, alcohol, and marijuana) use.
A sample of 1,810 public school students was surveyed in sixth grade (M age 11.5 years) and
seventh grade. Temperament dimensions were related to substance use, and structural mod-
eling analyses showed indirect effects through self-control constructs. Good self-control had
a path to higher academic competence and had direct effects to less peer use and less ado-
lescent substance use; poor self-control had a path to more adolescent life events and more
deviant peer afﬁliations. Academic competence and life events had indirect effects to adoles-
cent substance use, through peer afﬁliations. Findings from self-report data were corroborated
by independent teacher ratings. Effects were also noted for family variables and demographic
characteristics. Implications of epigenetic theory for prevention research are discussed.
KEY WORDS: temperament; self-control; substance use; early onset; epigenetic theory.
Early onset of substance use is known to have
prognostic signiﬁcance. Follow-back and follow-up
studies have shown that substance use before the age
of 13 years is predictive of substance abuse problems
at later ages (Hawkins et al., 1997; Kandel & Davies,
1992; Robins & Przybeck, 1985). Thus it is important
to obtain a better understanding of the processes that
act to either promote or deter early use. Theoretical
models have suggested that temperament dimensions
Ferkauf Graduate School of Psychology of Yeshiva University,
Bronx, New York.
Department of Epidemiology and Social Medicine, Albert
Einstein College of Medicine of Yeshiva University, Bronx, New
Present address: Department of Epidemiology and Biostatistics,
George Washington University Medical Center, Washington, DC.
Present address: Psychology Department, Mercy College, Dobbs
Ferry, New York.
Correspondence should be directed to Thomas Ashby Wills,
Health Psychology Training Program, Ferkauf Graduate School
of Psychology, Albert Einstein College of Medicine, 1300 Morris
Park Avenue, Bronx, New York 10461; e-mail: wills@aecom.
may be particularly relevant for early onset (Tarter &
Vanyukov, 1994; Zucker, 1994). Temperament is de-
ﬁned as simple characteristics that are early appear-
ing, have some stability over time, and have a consti-
tutional basis (Pedlow et al., 1993; Rothbart & Ahadi,
1994; Rothbart et al., 1994).
Lifespan studies have
shown that temperament measures in childhood are
related to substance use at later ages (Cloninger et al.,
1988; Pulkkinen & Pitk¨anen, 1994). However, there
has been little evidence available to test theoretical
models of the relation between temperament and sub-
stance use, during the time when onset is actually
This research tested predictions derived from
epigenetic models of behavioral development (Scarr,
1992; Tarter et al., 1995). Epigenetic theory posits that
Temperament differs from personality in that temperament di-
mensions are early simple attributes that reﬂect the “style” of
behavior, while adult personality characteristics are complex or-
ganized domains including elaborated content involving action
tendencies, beliefs and attitudes, and orientations toward social
relationships, added with cognitive and social maturation and ex-
perience. The linkage of these concepts is discussed in several
places (Angleitner & Ostendorf, 1994; Molfese & Molfese, 2000;
Rothbart & Ahadi, 1994).
2001 Society for Prevention Research