Development of impulsivity and risk‐seeking: Implications for the dimensionality and stability of self‐control*

• DOI: http://doi.org/10.1111/1745-9125.12214
• Published date: 01 August 2019
• Date: 01 August 2019
• Author: Walter Forrest,Carter Hay,Michael Rocque,Alex O. Widdowson

Received: 12 January 2018 Revised: 11 February 2019 Accepted: 14 February 2019

DOI: 10.1111/1745-9125.12214

ARTICLE

Development of impulsivity and risk-seeking:

Implications for the dimensionality and stability

of self-control*

Walter Forrest1Carter Hay2Alex O. Widdowson3Michael Rocque4

1School of Social Science, University of Queensland

2College of Criminology and Criminal Justice, Florida State University

3Department of Criminal Justice, University of Louisville

4Department of Sociology, Bates College

Correspondence

WalterForrest, School of Social Science, Uni-

versityof Queensland, St Lucia, QLD 4072,

Australia.

Email:w.forrest@uq.edu.au

*Additionalsupporting information

canbe found in the listing for this arti-

clein t he WileyOnline Library at

http://onlinelibrary.wiley.com/doi/10.1111/

crim.2019.57.issue-3/issuetoc.

Abstract

In Gottfredson and Hirschi’sself-control theor y,introduced

in 1990, they contend that self-control is a unidimensional

construct that develops early in childhood and remains sta-

ble throughout the life span. According to findings reported

in recent research, however,these arguments are now being

challenged, with scholars pointing to ways in which self-

control may be multidimensional in nature and may change

beyond the period of alleged stabilization. In this study,

we draw on Steinberg’s dual systems model, introduced

in 2008, to consider this issue further. We examine that

model’s two key elements of low self-control—risk-

seeking and impulsivity—to determine whether they are

empirically distinguishable from one another and have dif-

fering developmental trajectories from childhood to early

adulthood. We also consider the consequences of changes

in risk-seeking and impulsivity for within-individual

changes in crime. We examine these issues with data

from the National Longitudinal Survey of Youth 1979

(NLSY79) for individuals from 10 to 30 years old. The

results of our analyses show support for a multidimen-

sional and dynamic conception of self-control—from age

Wethank the editor and anonymous reviewers for their comments on earlier versions of this article. We also thank TenkoRaykov

for his advice on the calculation of reliability coefficients. All remaining errors are our own.

512 © 2019 American Society of Criminology wileyonlinelibrary.com/journal/crim Criminology.2019;57:512–543.

FORREST ET AL.513

10 to age 30, risk-seeking and impulsivity are empirically

distinct and develop in divergent ways that are consistent

with the dual systems model. Changes in risk-seeking and

impulsivity also affect changes in crime, but their effects

vary with age and changes in the other element. We discuss

these findings and their implications for self-control and

the development of life-course criminology.

KEYWORDS

impulsivity, life-course, offending, risk-seeking, self-control

Self-control theory has generated considerable research with findings that show support for its central

proposition that low self-control leads to heightened crime and antisocial behavior (Pratt & Cullen,

2000; Wright, Caspi, Moffitt, & Silva, 1999). Confidence in that conclusion has prompted investiga-

tions of the theory’s other hypotheses regarding such things as the causes of self-control (Perrone,

Sullivan, Pratt, & Margaryan, 2004; Polakowski, 1994), whether self-control accounts for the impact

of other factors (Wright et al., 1999), and the possibility that self-control’s effects on crime are uni-

form across individuals and social circumstances (Hay & Forrest, 2008; Vazonyi, Pickering, Junger, &

Hessing, 2001). Research findings on these issues are more mixed, often highlighting a complicated

interconnection between self-control, crime, and other criminological factors. Overall, one conclusion

seems clear: Self-control is among the strongest known correlates of crime and deviance, but there is

much to learn about its role in crime causation.

Beyond these questions, two additional issues receiving attention involve the stability and dimen-

sionality of self-control. On stability, according to self-control theory, self-control is formed in late

childhood and remains stable thereafter—those with low self-control by approximately age 10 (rel-

ative to other 10-year-olds) should have low self-control at later points in life when compared with

similarly aged peers. On dimensionality, Gottfredsonand Hirschi (1990) argued that elements thought

to comprise the broader self-control concept “come together in the same people” (p. 91) to reflect a

unidimensional “stable construct.” The evidence on these two issues, which we describe in more detail

later in this article, has not entirely supported the theory. For example, although findings from longitu-

dinal studies reveal much self-control stability,change is evident also, especially in longer term studies

(Burt, Sweeten, & Simons, 2014; Hay & Forrest, 2006; Roberts & Mroczek, 2008). On dimension-

ality, although in many studies, researchers have identified a core self-control factor, they also often

have revealed multidimensionality—elements of lowself-control do not overlap entirely with the core

self-control factor, and independent effectson cr ime of lowerorder constructs sometimes exceed those

for global self-control measures (Longshore, Rand, & Stein, 1996; Ward,Nobles, & Fox, 2015; Wolfe

& Hoffman, 2016). Thus, taken as a whole, self-control does not seem entirely stable after childhood

and a global self-control construct includes independent dimensions.

In one line of recent research, scholars haveconnected these two issues (Burt et al., 2014; Steinberg,

2010; Steinberg et al., 2008). Specifically, self-control may be composed of distinct dimensions that

differ from one another in stability and change. Steinberg’s dual systems model of neurobiological

development, which increasingly informsbehavioral science research on risky behavior, is built on this

possibility. In the model, two dimensions of low self-control are emphasized, conceptualizing them in

much the same way that Gottfredson and Hirschi (1990) did. The first is risk-seeking, which involves

514 FORREST ET AL.

“the tendency to seek out novel, varied, and highly stimulating experiences” (Steinberg et al., 2008,

p. 1765), whereas the second is impulsivity, which involves an “inability to delay gratification” and

“spur of the moment” acts (Steinberg, 2010, p. 218). In contrast to self-control theory, Steinberg

argued that although these two tendencies may coexist within individuals, they have different neu-

robiological underpinnings and follow different developmental pathways. In particular, risk-seeking

should surge in early adolescence (in response to puberty-related neurological developments), whereas

impulsivity should gradually decline during adolescence (in response to long-term development of

the prefrontal cortex). Steinberg’s arguments have many implications, but one is especially relevant

to tests of the stability thesis: Tests that use global self-control measures (collapsing risk-seeking and

impulsivity into the same measure) may fail to capture the varied development for these two separate

dimensions.

And indeed, in most stability tests, global self-control measures have been used. In one recent

exception, however, support was shown for the dual systems perspective of self-control (Burt et al.,

2014). Burt and colleagues (2014) found that the development of risk-seeking and impulsivity differed

markedly in adolescence and early adulthood and that both constructs uniquely predicted changes in

crime. They interpreted these findings as supporting the need to disaggregate measures of self-control

because even if these two elements “come together for some individuals at some points in the life

course, it is also apparent that they do not come together for most” (p. 476).

Although insightful and important, those findings come from a single study conducted among 775

Black youths living in Georgia and Iowa. Our goal is to extend this line of research using a larger

and more representative sample of youth who were born in multiple years and monitored for a longer

period (through age 30), and in ways that address further questions regarding the empirical indepen-

dence, stability, and consequences for crime of risk-seeking and impulsivity. To that end, we analyze

11 waves of data from the National Longitudinal Survey of Youth (NLSY79) to examine patterns of

self-control from age 10 to 30. We seek answers to three research questions. First, can risk-seeking

and impulsivity components be empirically distinguished from a broader scale that includes the full

set of their items? An affirmative answer supports a multidimensional conception of self-control.

Second, are risk-seeking and impulsivity stable over time, or do they change? And if they change,

do developmental patterns for these two dimensions differ? The findings from our analysis shed

further light on the extent of stability, while revealing differences in stability that may emerge across

distinct elements of self-control. Third, are within-individual shifts in risk-seeking and impulsivity

independently related to shifts in crime during this period? And if so, do changes in impulsivity and

risk-seeking affect crime more so at some ages than at others? If so, this would support a multidimen-

sional and dynamic conception of self-control—one that explicitly acknowledges a process in which

key dimensions of self-control fluctuate over time in ways that produce corresponding changes in

behavior.

1PRIOR THEORY AND RESEARCH ON SELF-CONTROL

STABILITY

Gottfredson and Hirschi’s (1990) arguments on self-control stability are clear. From their perspec-

tive, all individuals begin life in a primitive state lacking self-control. Self-control can be developed,

however, when parents monitor behavior, recognize deviance when it occurs, and punish such behav-

ior. This approach teaches children that deviance has consequences and that they must regulate their

actions. Those without such training remain in a primitive state and advance into adolescence with

self-control deficits.