Self-Control, Opportunity, and Substance Use

Abstract

We examine the effect of self-control and opportunity on adolescent substance use. When theorizing about the role of opportunity, we believe the “sanction potential” of a given situation should be considered. Our results suggest the effect of self-control on substance use does not depend on the availability of substances in the home (high sanction potential), but friends’ substance use (low sanction potential) conditions the effect of self-control on adolescents’ smoking, drinking, and marijuana use. Therefore, adolescents with low self-control are more likely to use substances only when they are presented with attractive opportunities that are unlikely to lead to sanctions.

Acknowledgments

This research uses data from the National Longitudinal Study of Adolescent Health (Add Health), a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Persons interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123 W. Franklin Street, Chapel Hill, NC 27516-2524 ([email protected]). No direct support was received from grant P01-HD31921 for this analysis.

Notes

¹Although Gottfredson and Hirschi (1990) originally discussed the role of opportunity in their theory, following the publication of A General Theory of Crime they have consistently denied the importance of opportunity. For example, Gottfredson and Hirschi (2003:9) argue that “self-control can be measured and the theory assessed without undue concern for differences in opportunities to commit criminal, deviant, or reckless acts.” Although Gottfredson and Hirschi have argued that opportunity is of little consequence, others argue that opportunity is an important part of self-control theory. For example, Goode (2008b:11) argues that “the criminally inclined do not commit crime under any and all circumstances; they commit crime only after assessing their odds of success and failure, and that assessment takes place within the context of opportunity.”

²Although self-control does not seem to influence perceptions about the risks and costs of crime, there is some evidence that people who are low in self-control are more likely to believe that crimes will be rewarding or pleasurable (Piquero and Tibbetts 1996) and to perceive sanctions as unfair (Piquero et al. 2004).

³Previous research generally shows that self-control does not have a significant effect on perceptions of the certainty and severity of sanctions. A related issue is whether or not self-control conditions the effect of sanctions on deviant behavior. That is, do sanctions have the same effect on people with high and low self-control? Research on this point has been very mixed. Nagin and Paternoster (1994) found that sanctions have a stronger effect on those high in self-control, while Wright et al. (2004) found that sanctions have a stronger effect on those low in self-control (see, also, Tittle and Botchkovar 2005b) and Cochran et al. (2008) found the effect of perceived sanctions did not vary by level of self-control.

⁴We argue that opportunities differ in terms of their sanction potential, or the probability that offenders will be sanctioned for engaging in deviant behavior. Similarly, borrowing from rational choice theory, in a recent article Seipel and Eifler (2010) argue that opportunities vary according to their costs. That is, opportunities with a “high risk of being caught are costly and therefore unfavorable opportunities,” whereas “opportunities linked to a low risk of being caught … are less costly and therefore favorable opportunities” (Seipel and Eifler 2010:173). Similar to our argument that adolescents with low self-control will be tempted most by opportunities with low sanction potential, Seipel and Eifler (2010) argue that personality traits, like self-control, are most relevant in low-cost situations. In contrast, Seipel and Eifler (2010) argue that in high-cost situations rational choice or utility is more important.

⁵Not all of the adolescents included in the first wave of Add Health were included in subsequent waves. The vast majority of adolescents who were seniors in high school at Wave 1 were not re-interviewed at Wave 2. Also, some adolescents who were eligible for Wave 2 could not be located or were unable or unwilling to be interviewed again. In total, 13,568 adolescents were included in both Wave 1 and Wave 2. Analysis of non-response suggests that sample attrition introduces very little bias in estimates of smoking, drinking, and marijuana use, which we use for our dependent variables (Kalsbeek et al. 2001). Furthermore, the sample weights included in Add Health adjust for nonresponse at Wave 2.

⁶We are not the first to argue that delinquent or substance using peers represent a measure of opportunity (see, for example, Longshore and Turner 1998; Warr 2001). As Warr (2001:79) argues, “opportunity is not only temporally and spatially structured, but socially structured as well, and opportunities for crime have as much to do with relations among offenders as with those between offenders and victims.”

⁷Gottfredson and Hirschi (1990:89) argue that self-control consists of multiple elements, including immediate gratification, simple gratification, thrill seeking, few long term benefits, little skill or planning, and insensitivity, many of which we tap with the items included in our measure of self-control. Contrary to Gottfredson and Hirschi (1990), but consistent with previous research, when we conducted a factor analysis the self-control items did not load on a single factor. Not surprisingly, the behavioral items tended to group together on one factor, while the problem solving items tended to group together on another. Despite the results of the factor analysis, following Gottfredson and Hirschi, and much previous research, we combine the items into a single measure of self-control.

⁸Although it may seem odd to code the measure of self-control so that higher scores on the index indicate lower levels of self-control, for the analysis of interaction terms, coding self-control in this way makes the interpretation of the simple effects more meaningful.

⁹Unstandardized interaction terms often cause problems with multicollinearity (Aiken and West 1991). In order to check our results, we replicated the analysis using standardized (or “centered”) interactions. Standardizing the interactions eliminated any problems with multicollinearity. The pattern of results for the standardized interactions was the same as the results using the unstandardized interactions (reported in the text).

¹⁰The items we use to measure substance use are not normally distributed. Although ordinary least squares (OLS) regression does not assume a normally distributed dependent variable (Allison 1999a:130), when the dependent variable is positively skewed two assumptions of OLS regression may be violated: non-normally distributed errors and heteroscedasticity. It is unlikely either of these problems influences our results. First, Allison (1999a:130) notes that non-normally distributed errors are not a concern with samples over 200 (see also Fox 1991). Second, using robust standard errors, which we have done, helps address heteroscedasticity (Allison 1999a:127).

TABLE 2 OLS Regression Coefficients for the Effects of Opportunity and Self-Control on Smoking

We also rely on OLS regression because testing interaction terms with a nonlinear model (e.g., logistic regression, ordered logistic regression, negative binomial, logit and probit models) creates statistical problems that can lead to faulty conclusions (Ai and Norton 2003; Allison 1999b; Norton et al. 2004). In particular, with nonlinear models interaction terms will often “reverse” direction. In fact, when we replicated the analysis for smoking using a negative binomial regression, compared to the OLS results the interaction terms were completely opposite (the same thing happened when we replicated the analysis for drinking and marijuana use with tobit regression). After closely inspecting the data (i.e., a series of crosstabs), we determined the OLS results are correct. That is, the direction of the OLS interaction terms fit the pattern that is found in the data, whereas the results for alternative models do not. Therefore, although some might argue an alternative method should be used for the analysis, we think OLS is a more trustworthy method.

*p < .05; **p < .01. Adjusted standard errors in parentheses.

¹¹When interaction terms are standardized, “b1 is the predicted effect of X on Y when Z equals its sample mean” (Jaccard and Turrisi 2003:25), but we used unstandardized interaction terms (i.e., we did not center the independent variables before creating the product term). Thus, “in the product-term model … the coefficient for X estimates the effect of X on Y when Z is at a specific value, namely, when Z = 0” and, conversely, “the coefficient for Z estimates the effect of Z on Y when X is at a specific value, namely, when X = 0” (Jaccard and Turrisi 2003:24).

In some cases, a score of zero on an independent variable either does not exist or is not meaningful, in which case it is inappropriate to interpret the simple effects. All of the variables we used to create the product terms, however, have meaningful zero points. For availability of substances in the home, which is a dummy variable, zero indicates substances are not easily available in the home. For friends’ substance use, zero indicates an adolescent has no friends who use a particular substance. For self-control, zero indicates adolescents with very high self-control.

*p < .05; **p < .01. Adjusted standard errors in parentheses.

*p < .05; **p < .01. Adjusted standard errors in parentheses.

Additional information

Notes on contributors

Scott A. Desmond

SCOTT A. DESMOND is a Visiting Assistant Professor in the School of Public and Environmental Affairs at Indiana University Purdue University-Indianapolis. Although his research focuses primarily on adolescent religious development, and how adolescent religiosity influences juvenile delinquency, he also studies how neighborhood characteristics and self-control contribute to crime, delinquency, and substance use.

Alan S. Bruce

ALAN S. BRUCE is Associate Professor of Sociology and Director of the Criminal Justice Program at Quinnipiac University in Connecticut. His research and teaching interests include youth crime, crime theory, and crime and media.

Melissa J. Stacer

MELISSA J. STACER is an Assistant Professor of Criminal Justice Studies at the University of Southern Indiana. She received her Ph.D. from Purdue University in 2010. Her research interests include the social world of prisons, problems of prison inmates, and the impact of the federalization of crime on prison populations.