stability hypothesis - Springer Link

5 downloads 8171 Views 1023KB Size Report
Florida Atlantic University, Boca Raton, FL 33431-0991. AMERICAN ..... Finally, we use hierarchical linear modeling tech- ..... Comparing criminal career models.
TESTING GOTTFREDSON AND HIRSCHI'S "LOW SELF-CONTROL" STABILITY HYPOTHESIS: An Exploratory Study$ BRUCE J. ARNEKLEVt Florida Atlantic University JOHN K. COCHRAN University of South Florida RANDY R. GAINEY Old Dominion University ABSTRACT: In Gottfredson and Hirschi's (1990) A General Theory of Crime, criminal and analogous behaviors are argued to be the result o f a stable individual psychological trait referred to as "low self-control." In this article, we test the proposition that low self-control is a stable characteristic. We conduct a two-wave panel study that measures the self-reported self-control o f college students at two relatively close time points. Our test o f the stability hypothesis includes conducting four separate analyses o f the data (t-tests, correlations, individual change scores, and H L M analyses). In general we find that most o f the dimensions o f self-controL and the overall self-control construct, appear to be relatively stable across this short period o f time. However, given that the measurements were taken closely together and that we do not find correlations as strong as might be expected, at points the strength o f the stability is somewhat unclear.

INTRODUCTION In A General Theory of Crime, Gottfredson and Hirschi (1990) purport to account for the variation in individual offending patterns. According to the authors, individual offending patterns diverge from one another to the extent that individuals have different degrees of time-stable individual differences in criminal potential, referred to as "low self-control," and opportunity differences for criminal and analogous behaviors. $ Paper presented at the annual meeting of the American Society of Criminology, Chicago, November 1996. ? Direct all correspondence to: Bruce J. Arneklev, Department o f Criminal Justice, Florida Atlantic University, Boca Raton, FL 33431-0991. A M E R I C A N J O U R N A L OF CRIMINAL JUSTICE, Vol. 23 No. 1, 1998 9 1998 Southern Criminal Justice Association

108

TESTINGTHE LOW SELF-CONTROLSTABILITY HYPOTHESIS

In this article, we test Gottfredson and Hirschi's "self-control" stability hypothesis. Our efforts are rather modest in that we simply look at whether self-control (and/or its dimensions) remains stable over a short period of time. Our focus is on a group of individuals who, according to the theory, should have relatively high levels of self-control: a sample of college students. In this exploratory study, we test whether these individuals' reports of their own self-control changed from the beginning of the semester (time 1) to the end of the semester (time 2), especially given the fact that finals week was approaching during the second wave. In this regard, we touch on Longshore, Turner, and Stein's (1996, pp. 224-225) suggestion that research should examine whether social factors have the potential to change levels of selfcontrol. Given the sample we use and the brevity of time in which levels of self-control are measured, we should acknowledge that our test probably gives the theory's stability hypothesis an optimal chance for support. First, as Gottfredson and Hirschi (1990, p. 251) observe, "Some subjects are more suitable than others for questionnaire surveys." Therefore, unlike Longshore et al., (1996), who recently surveyed a sample of drug-using criminal offenders, we survey a population with presumably higher levels of self-control (i.e., college students). Arguably, those scoring higher on criminality should disclose less reliable and less valid responses: "the higher the level of criminality, the lower the validity of crime measures" and, by extension, measures of self-control (Gottfredson & Hirschi, 1990, p. 249). At least theoretically, our respondents should provide relatively more reliable and valid responses regarding their own levels of self-control. Second, Gottfredson and Hirschi (1990, p. 231) have also suggested that since "differences in selfcontrol remain reasonably stable from the time they are first identified...short-term predictability should be excellent." To our knowledge, this is the first longitudinal test of the stability of self-control as it is conceptualized by Gottfredson and Hirschi (1990).

A General Theory of Crime Gottfredson and Hirschi's (1990) A General Theory of Crime has attracted much attention in the recent criminological literature (Akers, 1991; Arneklev, Grasmick, Tittle, & Bursik, 1993; Barlow, 1991; Bartusch, Lynam, Moffitt, & Silva, 1997; Brownfield & Sorenson, 1993; Evans, Cullen, Burton, Dunaway, & Benson, 1997; Gibbs & Giever, 1995; Grasmick, Tittle, Bursik, & Arneklev, 1993; Hirschi & Gottfredson, 1993; Keane, Maxim, & Teevan, 1993; Longshore, Turner & Stein, 1996; Paternoster & Brame, 1997; Piquero & Tibbetts, 1996; Polakowski, 1994; Polk, 1991; Reed & Yeager, 1996; Wood, Pfefferbaum, &

ARNEKLEV/COCHRAN/GAINEY

109

Arneklev, 1993). The theory can be seen as a general theory of crime in that the authors attempt, with only two key concepts, to explain all criminal and analogous behaviors across time (i.e., history), place (i.e., culture) and groups (e.g., gender, race, and class). Gottfredson and Hirschi suggest that such behaviors are the product of available opportunities and levels of "low self-control." Though Gottfredson and Hirschi mention the role opportunities play in the genesis of crimes, the crux of their argument centers on low self-control, which is perhaps the cause of crime, (see Gottfredson & Hirschi, 1990, pp. 96, 110, 232) The fact that crime is by all odds the major predictor of crime is central to our theory. It tells us that criminality (low selfcontrol) is a unitary phenomenon that absorbs its causes such that it becomes, for all intents and purposes, the individuallevel cause of crime. (Gottfredson & Hirschi, 1990, p. 232, italics in original) Therefore, unlike a number of other general theories that have recently been offered (e.g. Akers, 1977; Braithwaite, 1989; Tittle, 1995; Wilson & Herrnstein, 1985), what is unique about Gottfredson and Hirschi's theory is that it seems to view individuals as being differentially predisposed to crime. According to the theory, low self-control consistently differentiates individuals from one another according to their "criminal potential" (Farrington et al., 1990). Because of this, the theory in effect ranks individuals in terms of the likelihood that they will engage in criminal and analogous acts (Gottfredson & Hirschi, 1990, p. 77FTN). Once established, by the ages of six to eight, low self-control is proposed to remain a stable dimension of one's personality (Gottfredson & Hirschi, 1990, p. 272).

Persistent Heterogeneity v. State Dependence Because Gottfredson and Hirschi argue that an individual's criminal potential remains unaltered across the life course, their theory stresses the idea of "persistent heterogeneity," as opposed to "state dependence" (see Nagin & Farrington, 1992). The state dependence explanation for crime suggests that recent life course events and circumstances are important causal factors affecting crime. In contrast, the persistent heterogeneity explanation states that involvement in crime is due to "time stable individual differences in criminal potential that are attained early in the life course" (Nagin & Farrington, 1992, p. 501). Supporting the persistent heterogeneity argument, Nagin and Farrington (1992, p. 501) find that "age of onset and persistence of offending are entirely attributable to time stable individual differences." In

110

TESTINGTHE LOW SELF-CONTROLSTABILITYHYPOTHESIS

fact, the persistent heterogeneity perspective is given further credibility by the scholarly literature that demonstrates that age of onset and persistence in offending are closely related (Guttridge, Gabrielli, Mednick, & Dusen, 1983; Hirschi & Gottfredson, 1983; Van Dusen & Mednick, 1983; see also White, Moffitt, Earls, Robins, & Silva, 1990; Moffitt, 1993a, 1993b Barnett, Blumstein, & Farrington, 1987; Greenberg, 1992). Early involvement in criminal and analogous acts is usually accompanied by a continuation in these types of behaviors and late desistance from them. Thus, low self-controlled actors' offending patterns simply follow the age/crime curve, though being persistently involved in criminal and analogous behavior earlier and at higher rates than others. They are "the early starters (who) are simply the most enduringly crime prone" (Nagin & Farrington, 1992, p. 504). Consequently, it may be that early life influences play a far more important role in crime causation than has been traditionally emphasized (McCord, 1991a, 1991b; Nagin & Farrington, 1992; White et al., 1990; Wilson & Herrnstein, 1985). Alternatively, from a state dependence perspective, Gottfredson and Hirschi's theory is viewed as too simplistic. According to this position, recent life-course events, including involvement in crime, are influential as facilitators and/or inhibitors of an individual's criminal potential (Nagin, Farrington, & Moffitt, 1995; Sampson & Laub, 1993). while not dismissing that time stable individual differences could have an enduring impact on antisocial behavior, this perspective highlights the trajectories and retrogressions of antisocial behavior as potentially key determinants of involvement in illegal behavior. These two perspectives lead to different predictions about the stability of low self-control. Evidence that low self-control is a stable construct lends support to Gottfredson and Hirschi's theory and the persistent heterogeneity argument. However, evidence that reveals that low self-control (and/or the dimensions of low self-control) varies across the life course cautions us to question the validity of Gottfredson and Hirschi's theory and supports the state dependence explanation of crime.

Recent Literature To be sure, some of the literature discussing Gottfredson and Hirschi's theory has been rather critical (Akers, 1991; Barlow, 1991; Polk, 1991). However, even these evaluations tend to "admit to the power, scope, and persuasiveness" of the authors' arguments (Akers, 1991, p. 201). In fact, much of the research testing the theory illustrates that it merits further examination. Grasmick et al. (1993), for example, find

ARNEKLEV/COCHRAN/GAINEY

111

that after operationalizing low self-control, each of its six dimensions that can be identified in the theory, for the most part, form a unidimensional construct as the theory contends (also see Polakowski, 1994). In addition, low self-control interacts with crime opportunity to predict force and fraud (Grasmick et al., 1993). Furthermore, low self-control predicts many of the "imprudent" behaviors that Gottfredson and Hirschi argue are analogous to crime (e.g., drinking and gambling) (Arneklev et al., 1993; Evans et al,, 1997; Gibbs & Giever, 1995; Keane et al., 1993; Polakowski, 1994). Finally, such imprudent outcomes also predict involvement in crime, as the theory suggests they should (Evans et al., 1997). While we would argue that replicating these studies is also important, what is missing from this body of research is a test of the stability of low self-control as it is conceptualized by Gottfredson and Hirschi.

Defining Low Self-Control According to A General Theory of Crime, there are six essential elements of low self-control (see Gottfredson and Hirschi, 1990, pp. 8990; also see Grasmick et al., 1993; Arneklev et al., 1993; Wood et al., 1993; Longshore et al., 1996). Individuals with low self-control tend to be impulsive, prefer simple tasks, have a high risk-seeking potential, favor physical (as opposed to mental) activities, be self-centered, and possess volatile tempers. Each of these characteristics causes people to be differentially focused on the immediate benefits of criminal behaviors rather than the more distant consequences associated with their commission. From Gottfredson and Hirschi's perspective differences in low self-control "remain reasonably stable with change in the social loca-

tion of individuals and change in their knowledge of the operation of sanction systems" (Gottfredson & Hirschi, 1990, p. 87, italics in original). This stability hypothesis is the focus of our test. We use a twowave panel design to test if low self-control is a stable characteristic.

MEASURES AND METHODS Sample Data for this project were gathered in two separate surveys conducted during the fall 1995 academic semester. The respondents in the study consisted of a convenience sample of 175 college students enrolled in various social science courses at three southern public universities in the United States. As is commonly known, convenience samples such as our own are inferior to probability samples but are useful in the context of exploratory research. In addition, Gottfredson and Hirschi's theory states that the general population can be ranked in

112

TESTING THE LOW SELF-CONTROL STABILITY HYPOTHESIS

terms of self-control, which suggests that researchers can attempt to derive samples of individuals with relatively higher levels (or lower levels; see Longshore et al., 1996) of self-control in that hierarchy to explore if self-reported self-control changes. Following this line of reasoning, we attempted to increase our sample size by including more respondents than we could obtain from a single university. The respondents in the sample were assured anonymity and that any identifying information that was gathered would be destroyed after the data were collected and recorded. The sample was reduced considerably (N = 127) due to attrition. The mean age of the respondents in the final sample was 22.8 years, with a range from 17 to 47. Of the sample, 1.6% were freshman, 7.1% sophomores, 55.1% juniors, and 36.2% seniors. The sample was 45.7% male and 54.3% female, and 79.5% white. Respondents in the sample read and responded to 24 identical questions designed to measure the six dimensions of self-control at two different time periods. Therefore, similar to the vast majority of tests of Gottfredson and Hirschi's theory (Arneklev et al., 1993; Grasmick et al., 1993; Longshore et al., 1996; Wood et al., 1993; but see Polakowski et al., 1994; Evans et al., 1997), self-control is measured by the respondents' self-reported assessments of their own levels of self-control. The first wave of the survey was distributed in the first week of the semester. The second survey, which contained identical items, was conducted at the end of the same semester. The time interval between the first and second waves was approximately four months. Respondents at time 2 were not told that they were answering the same time 1 survey, though a very few asked if they were.

Low Self-Control In this exploratory test, we use Grasmick et al.'s (1993) scale to operationalize the dimensions of self-control and the overall self-control construct. All responses to the 24 items on this scale were fixed on 4-point scales of Likert-type ordinal metrics: (1) strongly agree, (2) agree somewhat, (3) disagree somewhat, and (4) strongly disagree. Persons scoring high on the items have high self-control. Means, standard deviations, and alpha reliabilities for the scales are listed in Table 1. Principle component factor analyses for the items designed to measure each of the self-control dimensions justifies the creation of single scales for each component of self-control. Furthermore, the alpha reliability for the entire self-control scale at time 1 and time 2 is .8633 and .8946, respectively. For the 24 items, principal components analyses, which measure the entire self-control construct, justify the creation of a single

ARNEKLEV/COCHRAN/GAINEY

113

encompassing self-control composite for both time periods. The factor eigenvalues for the self-control composite at time 1 are 5.8498, 2.5249, 1.9891, 1.8314, 1.6328, 1.2774. The factor eigenvalues for the same composite at time 2 are 7.1476, 2.4127, 1.9059, 1.6875, 1.2987, 1.2306. According to the Scree Discontinuity Test, the most obvious break in both sets of eigenvalues is between the first and second factors.

114

TESTING THE LOW SELF-CONTROL STABILITY HYPOTHESIS

TABLE 1 Self-Control Scale Items at Time 1 and Time 2 (N = 127) ITEM IMPULSIVITY C O M P O N E N T I don't devote much thought and effort to preparing for the future. I often do whatever brings me pleasure here and now, even at the cost of some distal goal. I'm more concerned about what happens to me in the short run than in the long run. I much prefer doing things that pay off right away rather than in the future. SIMPLE TASKS C O M P O N E N T I frequently try to avoid things that I know will be difficult. When things get complicated, I tend to quit or withdraw. The things in life that are easiest to do bring me the most pleasure. I dislike really hard tasks that stretch my abilities to the limit. RISK-TAKING C O M P O N E N T I like to test myself every now and then by doing something a little risky. Sometimes I will take a risk just for the fun of it. I sometimes find it exciting to do things for which I might get in trouble. Excitement and adventure are more important to me than security. PHYSICAL ACTIVITIES C O M P O N E N T If I had a choice, I would almost always rather do something physical than Something mental. I almost always feel better when I am on the move than when I am sitting and thinking. I like to get out and do things more than I like to read or contemplate ideas. I seem to have more energy and a greater need for activity than most other people my age. SELF-CENTERED C O M P O N E N T I try to look out for myself first, even if it means making things difficult for other people. I ' m not very sympathetic to other people when they are having problems. If things I do upset people, it's their problem, not mine. I will try to get the things I want even when I know it's causing problems for other people. TEMPER COMPONENT I lose my temper pretty easily. Often, when I'm angry at people I feel more like hurting them than talking to them about why I am angry. When I am really angry, other people better stay away from me. When I have a serious disagreement with someone, it's usually hard for me to talk about it without getting upset.

All Likert items are answered on a 4-point scale of strongly agree (1), agree somewhat (2), disagree somewhat (3), and strongly disagree (4). Alpha reliability for the entire Self-Control Scale at Time 1 = 0.8638 Alpha reliability for the entire Self-Control Scale at Time 2 = 0.8946

ARNEKLEV/COCHRAN/GAINEY TIME 2

TIME 1 MEAN

STANDARD DEVIATION

115

FACTOR LOADING

MEAN

STANDARD DEVIATION

FACTOR LOADING

3.732 3.181 3.449 3.016

.526 .830 .731 .776

.61548 .67633 .80562 .79743 (Alpha = .6974)

3.701 3.094 3.425 3.110

.509 .840 .751 .715

.61918 .80816 .84692 .74471 (Alpha = .7527)

2.874 3.488 3.126 3.354

.816 .628 .787 .740

.78281 .74691 .69464 .72003 (Alpha = .7144)

2.945 3.472 3.197 3.315

.810 .640 .713 .709

.78822 .65302 .69660 .84934 (Alpha = .7381)

2.142 2.504 3.110 3.220

.932 1.068 1.018 .765

.77915 .89131 .81407 .70370 (Alpha = .8115)

2.283 2.559 3.118 3.339

.844 .957 .931 .737

.85202 .88802 .77139 .70716 (Alpha = .8202)

2.583 2.307 2,205 2.331

.912 .947 .946 .900

.81223 .80462 .81069 .75329 (Alpha = .8065)

2.693 2.260 2.181 2.291

.939 .884 .929 .892

.81246 .78844 .87602 .69989 (Alpha = .8064)

2.976 3.567 3.378 3.417

.904 .719 .776 .718

.73688 .75480 .79447 .82809 (Alpha = .7772)

3.039 3.520 3.315 3.362

.801 .744 .842 .763

.70800 .71410 .83363 .86318 (Alpha = .7859)

3.024 3.331 3.055 2.795

.947 .900 .970 1.018

.69750 .79572 .72921 .80892 (Alpha = .7529)

3.071 3.409 3.039 2.843

.919 .810 .929 .971

.73356 .83770 .81997 .79127 (Alpha = .8038)

116

TESTING THE LOW SELF-CONTROL STABILITY HYPOTHESIS

RESULTS To assess potential biases introduced by attrition, we computed a dichotomous measure where 0 -- follow-up data available and 1 = follow-up missing. We then assessed the correlation between this variable and various demographic variables and measures of self-control. Chisquare statistics were used for dichotomous independent variables, and t-tests were used for interval scale independent measures. There was a slight tendency for African-Americans to be missing at follow-up (chisquare = 3.16, p