NDSU Department of Child Development and Family Science
North Dakota Department of Human Services
June 1999 -- Volume 2, Number 3
Kevin M. Thompson
Associate Professor • Department of Sociology • North Dakota State University
• Minard Hall 402 • (701) 231-8938 • (701) 231-1047 (fax)
Parents and school authorities have traditionally believed that activity participation is an effective vehicle for preventing delinquency and substance use. However, empirical tests of this assumption have received equivocal support. This study re-examines this assumption using survey data from over 7,700 North Dakota 6th-12th graders. The data show that participation in musical activities was associated with lower odds of delinquency and substance use. The association odds were even stronger when music was coupled with sports, school organizations or both. Participation in sports was negatively associated with marijuana use and police contact while involvement in school organizations was associated with reduced substance use. Participating in after-school organizations had no effect on delinquency and substance use. The form of the significant associations was non-linear with the difference in slopes decreasing with an increase in activity participation time.
Parents routinely believe that one of the solutions to preventing deviant outcomes in their children is to encourage participation in extracurricular activities and organizations (Jensen & Rojek, 1998). Delinquents themselves contribute to this belief when they complain that there are "not enough places to go and not enough things to do" (Goldstein, 1990, p. 126). Indeed, when gang members have been surveyed (Knox, 1998), almost one-third of them report that they would have desisted from gang life if more opportunities had been available (e.g., sports, music, Boy's Club). This belief is further augmented by justice officials in their most recent action plan. This document delineates a series of characteristics embodied in effective delinquency prevention programs. These include programs that promote healthy adolescent development, provide adequate support and supervision, and offer youth a long-term stake in the community (U.S. Department of Justice, 1996). These are features typically embodied in conventional youth activities such as band, student council, volleyball, and YMCA after school programs.
Theoretically, involvement in conventional activities should suppress delinquency and substance use through four mechanisms. First, participation in activities changes the opportunity structure of delinquency and substance use (Hirschi, 1969; Empey & Stafford, 1991). This supposition is embodied in the folk belief that "idle hands are the devil's workshop." Second, conventional activities expose youth to pro-social peers, facilitating positive interaction and modeling. Third, participation in conventional activities fosters a sense of future orientation in youth (Hirschi, 1969). Conventional activities increase personal capital by allowing the practice of behaviors that bridge adult responsibilities, such as equating achievement with effort, making group decisions or emphasizing team cooperation. And fourth, conventional activities expose youth to adult mentors who work closely with youth and provide them with modeling and moral lessons (Tierney & Grossman, 1995).
Surprisingly, empirical research has not always supported an association between conventional activity participation and delinquency and substance use. When supported, the effects appear to be small. For instance, Hirschi (1969) found that lack of participation was only weakly associated with delinquency in urban boys. In a study of rural youth, Hindelang (1973) found that the relationship between participating in school activities and delinquency was non-linear, with the lowest delinquency rates reported by youth that were somewhat active rather than very active. Cernkovich & Giordano (1992) revealed that school involvement was negatively associated with delinquency in a sample of Ohio youth. Others have not found a significant association between participation and delinquency (Donnelly, 1981; Pakiz, Reinherz, and Frost, 1992).
Several investigators have examined whether the nature of extracurricular participation conditions the relationship with deviance. Regarding athletics, some scholars have intimated that some forms of extra-curricular participation may increase aggression. Seagrave, Moreau, & Hastad (1985) found that ice hockey players reported more physically violent delinquency than non-athletes. Agnew (1989) found that delinquency rates were lower among youth that participated in organized activities (e.g., volleyball) than activities that lacked supervision (e.g., street hockey). Landers and Landers (1978) noted that delinquency rates were lower when athletics was coupled with participation in service organizations than athletics alone or service participation alone. Research linking forms of school activity such as music participation to delinquency and substance use is lacking.
Findings regarding the effect of organizational participation outside of school have been mixed. Howell (1995) found that community-based after-school recreation programs reduced juvenile crime in the areas surrounding the recreation center, but this study employed the community rather than the individual as the unit of analysis. Several evaluations of the Big Brothers/Big Sisters program received mixed reviews. Two studies found that these community based mentoring programs did not reduce drug abuse (Fo & O'Donnell, 1975; McCord, 1992) while a more recent study found substantial reductions in youth drug use among mentored youth (Tierney & Grossman, 1995).
This paper addresses three research questions: (1) is there an association between activity participation and delinquency and substance use? (2) If so, is the level of delinquency and substance use a linear function of time spent in the activity, or does the curve of the slope vary between different levels of participation? (3) Does multiple activity participation predict delinquency and substance use better than a singular activity? If so, which coupled activities are most highly associated with a reduction in delinquency and substance use? These questions are examined using self-reported behavior among over 7,700 youth in North Dakota.
In 1992, the Search Institute of Minneapolis contracted with 14 school districts in Cass County to conduct a survey of youth in their schools. The survey was administered to 7,980 students in grades 6-12 (Search Institute, 1992). The intent of the study was to gather information from adolescents on a variety of problem behaviors. Care was taken to ensure that instrument administration procedures were standardized across the school districts. Written consent was obtained from parents prior to administration of the survey. The survey was administered to students in homerooms by school personnel. An estimated 5% of the students in these school districts were absent on the day the survey was administered. To ensure the students of complete anonymity, no names or identification numbers were used.
Because the reporting of problem behaviors carries the potential for unreliability, several checks were made on individual survey responses. The Search Institute discarded 247 surveys due to one of the following reasons: 1) reporting of inconsistent responses, such as a respondent reporting that they used LSD 20-39 times in the past year but reported no alcohol or marijuana use; 2) missing data on 40 or more of the 152 items, or 3) reports of unrealistically high levels of drug use and antisocial behaviors (e.g., a respondent reporting that they used PCP, LSD, and heroin more than 40 times in the past year). The number of discarded surveys represents less than 4% of the total number of surveys received. Typically, less than 5% of surveys are discarded for these reasons (Search Institute, 1992). The final sample yielded 7,733 usable questionnaires.
Activity participation
Respondents were asked, "During an average week, how many hours do you: 1) Spend in band, choir, orchestra, music lessons or practicing voice or an instrument (MUSIC)? 2) Spend playing sports on a school team (SPORTS)? 3) Spend in clubs or organizations outside of school (SCHORG)? 4) Spend in clubs or organizations outside of school (COMMCLUB)?" Response categories for each question were (1) 0 hours, (2) 1-2 hours, (3) 3-5 hours, (4) 6-10 hours, (5) 11 or more hours.
Delinquency
Two delinquency items were selected to assess respondent's level of delinquency. Respondents were asked, "During the last 12 months, how many times have you 1) gotten into trouble with the police, and 2) damaged property just for fun (such as breaking windows, scratching a car, putting paint on walls, etc.)? Because interest lies in the probability or log odds of encountering the police and engaging in an act of vandalism, these items were dichotomized. The police item likely represents a variety of forms of misconduct, including being taken into custody for a delinquent offense, being referred for a status offense or simply being asked to disperse. The vandalism item was chosen because it represents a common incident engaged in by youth when they are idle, weakly supervised, or lack community stakes in conformity.
Substance use
Respondents were asked about binge drinking and marijuana use in separate questions. For binge drinking, respondents were asked, "Think back over the last two weeks. How many times have you had five or more drinks in a row?" The item proceeded to describe what constituted a drink. Response categories were dichotomized to distinguish respondents who did not binge drink from those that did. To assess marijuana use, youth were asked, "How many times, if any, have you used marijuana (grass, pot) or hashish during the last 30 days?" Response categories were dichotomized to distinguish non-users from users. These items were selected because authorities are becoming increasingly concerned about the level of drinking and marijuana use among North Dakota youth.
Control variables
Because background characteristics are associated with conventional involvement and deviance (Jensen & Rojek, 1998), the following characteristics will be controlled in the analyses: age, gender, race/ethnicity, family living arrangement, fathers and mother's education level, and residence.
Variables for the analysis
Table 1 presents the variables and their univariate distribution for the analysis. The sample consisted of predominately white youths from two-parent families. The average respondent was 14.5 years of age and resided in a community with over 10,000 residents. Most respondents reported that their parents had attended college.
Table 1. Distribution of Variables (N = 7,733).
-------------------------------------------------------------------------
Variable Description N % Mean STD Range
-------------------------------------------------------------------------
Gender 0 = Male 3,793 49.2
1 = Female 3,922 50.8
-------------------------------------------------------------------------
Race/Ethnicity 0 = Non-white 310 4.0
1 = White 7,371 96.0
-------------------------------------------------------------------------
Living Arrangement 0 = Two parents 6,341 82.4
1 = One parent 1,359 17.6
-------------------------------------------------------------------------
Age Respondent's age
at time of survey 7,730 14.52 2.04 11-19
-------------------------------------------------------------------------
Residence 0 = Rural 1,799 23.5
1 = Urban 5,872 76.5
-------------------------------------------------------------------------
Father's Education Highest level of 6,791 4.35 1.18 1-6
education completed
-------------------------------------------------------------------------
Mother's Education Highest level of
education completed 7,064 4.29 1.10 1-6
-------------------------------------------------------------------------
Music
Participation Hours per week 7,720 2.04 1.24 1-5
-------------------------------------------------------------------------
Sports
Participation Hours per week 7,715 2.26 1.46 1-5
-------------------------------------------------------------------------
School
Organizations Hours per week 7,713 1.44 .81 1-5
-------------------------------------------------------------------------
Community
Organizations Hours per week 7,721 1.67 .96 1-5
-------------------------------------------------------------------------
Police Contact 0 = No 7,625 83.1
1 = Yes 16.9
-------------------------------------------------------------------------
Vandalism 0 = No 7,633 82.0
1 = Yes 18.0
-------------------------------------------------------------------------
Binge Drinking 0 = No 7,722 81.5
1 = Yes 18.5
-------------------------------------------------------------------------
Marijuana Use 0 = No 7,719 95.1
1 = Yes 4.9
-------------------------------------------------------------------------
The activity participation variables show that the most common activity engaged in by
these youth was sports, followed by music, community organizations/clubs, and school
organizations. The average respondent reported participating in sports slightly more than
two hours per week with 11% of the sample reporting participation levels of more than 11
hours per week. Consequently, of the reported activities, sports consumed the most time.
Endorsement levels for binge drinking, vandalism, and police contact were 17-18%.
Marijuana use was more infrequently endorsed with slightly less than 5% of the sample
reporting use.
Logistic regression analysis was employed to estimate the log odds of a youth engaging in delinquency and substance use as a function of their activity participation. This technique allows conversion of logit coefficients into odds ratios and probabilities. For continuous predictors (e.g., level of activity) this technique provides an indicator of how much more likely an outcome is when the predictor increases by one unit. When we dummy code a variable (e.g., gender), the odds ratios indicate how much more likely an outcome is for a specific predictor category as opposed to the category coded zero (see Menard, 1995).
Table 2 presents logit estimates and goodness of fit statistics for each of the four deviance measures. The central question in this table regards whether activity participation has an effect on delinquency and substance use after controlling for the youth's background characteristics (Model 2). Consequently, demographic characteristics were entered in Model 1 followed by the entry of the activity participation measures in Model 2.
Table 2. Logit estimates of delinquency and substance use.
-----------------------------------------------------------------------
POLICE CONTACT
---------------------------------------------------
- - - - Model 1 - - - - - - - - Model 2 - - - -
B (SE) Odds Ratio B (SE) Odds Ratio
-----------------------------------------------------------------------
Demographics
Age .22 (.02) 1.24* .20 (.02) 1.23*
Gender -1.06 (.07) .35* -.99 (.08) .37*
Race/Ethnicity -.12 (.18) .88 -.13 (.18) .88
Living Arrangement .63 (.09) 1.88* .56 (.09) 1.76*
Father's Education -.17 (.03) .84* -.13 (.03) .88*
Mother's Education -.03 (.04) .97 -.01 (.04) .99
Residence .14 (.08) 1.15 .07 (.09) 1.08
Activity Participation
MUSIC -.23 (.03) .79*
SPORTS -.09 (.02) .91*
SCHORG -.09 (.05) .92
COMMCLUB .06 (.04) .92
Constant -3.67 (.37) -3.07 (.38)
-2 Log Likelihood 5219.34 5145.23
-----------------------------------------------------------------------
VANDALISM
---------------------------------------------------
- - - - Model 1 - - - - - - - - Model 2 - - - -
B (SE) Odds Ratio B (SE) Odds Ratio
-----------------------------------------------------------------------
Demographics
Age .04 (.02) 1.04 .03 (.02) 1.03
Gender -1.65 (.08) .19* -1.55 (.08) .21*
Race/Ethnicity -.13 (.17) .87 -.15 (.17) .86
Living Arrangement .32 (.09) 1.38* .28 (.09) 1.32*
Father's Education -.05 (.03) .95 -.02 (.03) .98
Mother's Education -.03 (.03) .97 -.02 (.03) .99
Residence .11 (.08) 1.12 .08 (.08) 1.08
Activity Participation
MUSIC -.17 (.03) .84*
SPORTS .02 (.02) 1.02
SCHORG -.10 (.05) .90
COMMCLUB -.01 (.04) .99
Constant -1.18 (.35) -.77 (.36)
-2 Log Likelihood 5466.99 5425.06
-----------------------------------------------------------------------
-----------------------------------------------------------------------
BINGE DRINKING
---------------------------------------------------
- - - - Model 1 - - - - - - - - Model 2 - - - -
B (SE) Odds Ratio B (SE) Odds Ratio
-----------------------------------------------------------------------
Demographics
Age .36 (.02) 1.43* .35 (.02) 1.41*
Gender -.35 (.07) .70* -.21 (.07) .81*
Race/Ethnicity .03 (.18) 1.03 -.01 (.18) .99
Living Arrangement .55 (.09) 1.73* .48 (.08) 1.61*
Father's Education -.12 (.03) .88* -.08 (.03) .92
Mother's Education -.09 (.03) .91* -.06 (.03) .94
Residence .10 (.08) 1.10 .04 (.08) 1.05
Activity Participation
MUSIC -.28 (.03) .76*
SPORTS -.00 (.02) 1.00
SCHORG -.21 (.05) .81*
COMMCLUB .07 (.04) .93
Constant -5.96 (.37) -5.17 (.38)
-2 Log Likelihood 5669.43 5533.56
-----------------------------------------------------------------------
MARIJUANA USE
---------------------------------------------------
- - - - Model 1 - - - - - - - - Model 2 - - - -
B (SE) Odds Ratio B (SE) Odds Ratio
-----------------------------------------------------------------------
Demographics
Age .41 (.03) 1.51* .40 (.03) 1.50*
Gender -.36 (.12) .70* -.32 (.12) .73*
Race/Ethnicity -.17 (.28) .84 -.21 (.28) .81
Living Arrangement .80 (.13) 2.22* .71 (.13) 2.03*
Father's Education -.11 (.05) .89 -.05 (.05) .95
Mother's Education -.13 (.06) .87 -.09 (.06) .90
Residence .20 (.14) 1.22 .13 (.15) 1.13
Activity Participation
MUSIC -.15 (.06) .86*
SPORTS -.19 (.04) .83*
SCHORG -.37 (.10) .69*
COMMCLUB -.01 (.06) .98
Constant -8.20 (.66) -7.19 (.69)
-2 Log Likelihood 2341.83 2286.49
-----------------------------------------------------------------------
*p < .01
The data show that music participation was negatively associated with all four deviance
measures. The effect of music participation had its greatest impact on binge drinking,
followed by police contact, vandalism, and marijuana use. Each one unit increase in music
participation was associated with a decrease of .28 in the logit for binge drinking. Net
of a youth's background characteristics and other activities, music reduced the odds of
binge drinking by 24% (see odds ratio). Music participation had a much smaller effect on
marijuana use with a decrease of .15 in this logit and a reduction in odds of 14%.
Sports participation was unevenly associated with the deviance measures. Participating in sports was significantly associated with police contact and marijuana use but was not associated with vandalism and binge drinking. Sports participation had its strongest impact on marijuana use with a one unit increase in activity associated with a decrease of .19 in the logit for marijuana use and a reduction in odds of 17%. School organization participation was significantly associated with binge drinking and marijuana use, reducing the odds for participants by 19% and 31%, respectively. Neither of the delinquency measures was significantly associated with school organization involvement. Contrary to several published studies, community organization participation was not significantly associated with a reduction in delinquency or substance use, and in two cases the coefficients were positive.
Because some forms of activity participation were associated with deviance, it may be helpful to know how much activity participation is necessary to observe a marked reduction in the odds of deviance. This question requires an examination of the form of these relationships. Logistic regression allows us to translate the logits into probabilities, enabling us to assess the change in probability of deviance from one level of the independent variable to another.
Table 3 presents the estimated probabilities and change in probability from one level of participation to the next for the significant associations in Table 2. A reading of this table shows that the estimated probability of encountering problems with the police was around 22% for youth that did not participate in music and dropped to 5% for youth that participated over 11 hours per week. The slope of the curve varied, however, depending on the level of music participation. The largest change in probability was observed between youth who did not participate and those who reported participating 1-2 hours per week (6 ½%), suggesting that music had its greatest impact on police contact at the point of participation vs. non-participation. The difference in slope between youth that participated 11 hours or more per week and 6-10 hours per week was smaller (2.5%), indicating that the relationship was non-linear. In the next column, we see that playing sports had a much smaller effect on police contact than music participation. The total shift in slope from non-participation to maximum participation was only 4%, and at higher participation levels the slope changes were miniscule.
Table 3. Probability of delinquency and substance use and changes in probability.
-----------------------------------------------------------------------
POLICE CONTACT
--------------------------------------------------------
Music Sports
-------------------------- --------------------------
Change in Change in
Probability Probability Probability Probability
-----------------------------------------------------------------------
No Participation .224 .064 .182 .011
1-2 hours/week .160 .049 .171 .011
3-5 hours/week .111 .035 .160 .010
6-10 hours/week .076 .025 .150 .009
11+ hours/week .051 .141
Cumulative Change .173 .041
-----------------------------------------------------------------------
VANDALISM
----------------------------
Music
--------------------------
Change in
Probability Probability
-------------------------------------------
No Participation .224 .049
1-2 hours/week .175 .040
3-5 hours/week .135 .032
6-10 hours/week .103 .025
11+ hours/week .078
Cumulative Change .146
-----------------------------------------------------------------------
BINGE DRINKING
--------------------------------------------------------
Music School Organizations
-------------------------- --------------------------
Change in Change in
Probability Probability Probability Probability
-----------------------------------------------------------------------
No Participation .249 .075 .200 .039
1-2 hours/week .174 .056 .161 .033
3-5 hours/week .118 .040 .128 .026
6-10 hours/week .078 .027 .102 .022
11+ hours/week .051 .080
Cumulative Change .198 .120
-----------------------------------------------------------------------
MARIJUANA USE
--------------------------------------------------------
Music Sports
-------------------------- --------------------------
Change in Change in
Probability Probability Probability Probability
-----------------------------------------------------------------------
No Participation .067 .022 .063 .013
1-2 Hours/week .045 .014 .050 .011
3-5 Hours/week .031 .011 .039 .008
6-10 Hours/week .020 .006 .031 .007
11 + Hours/week .014 .024
Cumulative Change .053 .039
School Organizations
--------------------------
Change in
Probability Probability
-------------------------------------------
No Participation .057 .020
1-2 Hours/week .037 .013
3-5 Hours/week .024 .009
6-10 Hours/week .015 .005
11 + Hours/week .010
Cumulative Change .047
-----------------------------------------------------------------------
Music participation had roughly the same effect on vandalism as it did on police contact.
The probability of engaging in vandalism was around 22% for music non-participants and
dropped to around 8% for youth that participated over 11 hours per week. Again, the
largest reduction in slope was evident between youth who did not participate and those who
participated 1-2 hours per week. Changes in the slope of the curve for pairs of points at
higher levels of participation were smaller.
The binge drinking-music pattern yielded the largest change in slope from non-participation to heavy participation. The probability of binge drinking for non-participants was 25% and dropped to 5% for youth who participated 11 or more hours per week. Again the largest reduction was evident at the level between non-participation and participation of 1-2 hours per week with a change in slope of 7.5%. Participating in school organizations had a smaller effect on binge drinking than music participation with a cumulative change in probability of 12%.
For marijuana use, the cumulative effect of participation was quite small. While Table 2 demonstrated that these activities were significantly associated with a reduction in marijuana use, the difference between non-participants and active participants (11+ hours) was only around 5%. Again however, the form of these associations was non-linear.
The final question examines whether participating in multiple activities is associated with delinquency and substance use over and above the effect of singular activity participation. To estimate this effect, a variable was created contrasting seven forms of activity participation (e.g., music only, music + sports) against the reference category of non-participation for all three activities. Consequently, the logits in Table 4 indicate the effect of at least some level of activity participation on delinquency and substance use, contrasted with those that report no activity in all three categories. The Wald statistic (b/SE)2 was employed to assess the relative importance of each category in explaining delinquency and substance use.
Table 4. Logit estimates of delinquency and substance use for activity participation
(reference category = non-participation in all activities).
------------------------------------------------------- Activity Participation - - Police Contact - - (in descending order) B SE Wald ------------------------------------------------------- Music & Sports -.92* .14 46.27 Music & Sports & School Org. -.97* .15 43.10 Music Only -.74* .12 34.58 Music & School Org. -1.01* .18 30.35 Sports & School Org. -.67* .16 18.09 Sports Only -.38* .10 15.37 School Org. Only -.64* .17 13.76 ------------------------------------------------------- Activity Participation - - - Vandalism - - - (in descending order) B SE Wald ------------------------------------------------------- Music & School Org. -.98* .20 25.27 Music & Sports -.56* .12 20.63 Music & Sports & School Org. -.57* .14 17.44 Music Only -.49* .12 15.93 School Org. Only -.51* .18 7.86 Sports & School Org. -.36 .15 5.61 Sports Only -.07 .10 .55 ------------------------------------------------------- Activity Participation - - Binge Drinking - - (in descending order) B SE Wald ------------------------------------------------------- Music Only -.90* .12 51.41 Music & Sports & School Org. -.95* .14 46.76 Music & School Org. -1.11* .17 42.76 Music & Sports -.77* .13 34.54 School Org. Only -.61* .16 14.55 Sports & School Org. -.49* .15 11.50 Sports Only -.01 .09 .020 ------------------------------------------------------- Activity Participation - - Marijuana Use - - (in descending order) B SE Wald ------------------------------------------------------- Music & Sports & School Org. -1.95* .37 27.38 Music & Sports -1.18* .27 19.32 Music & School Org. -1.32* .32 16.59 School Org. Only -1.08 .30 12.99 Sports Only -.54* .16 12.02 Sports & School Org. -.88* .27 10.73 Music Only -.45 .18 5.81 ------------------------------------------------------- * p < .01
For police contact, the lowest odds were observed for youth who coupled sports with music
followed by youth who report participation in all three activities. Compared to youth who
reported no participation in any activities, participating in music and sports was
associated with a decrease of .92 in the logit for police contact. This translates to an
odds ratio of .60, meaning that coupling music with sports decreases the odds of police
contact by 60% over not participating in any activities. Participation in any activity,
whether engaged in solely or simultaneously, was significantly associated with a decrease
in police contact compared to non-participation. For vandalism, music was again an
important correlate, particularly when combined with another activity. When coupled with
school organization involvement, music reduced a youth's odds of vandalizing property by
63% over youth who reported no participation. Playing sports or coupling sports with
school organizations was not significantly associated with vandalism.
The correlates were different for binge drinking and marijuana use. Youth who participated in music and/or combined this activity with another activity were significantly less likely than non-participants in all three categories to binge drink. Youth who reported participation in all three activities reduced their odds of binge drinking by 62% compared to non-participants. Beyond these combinations, the effects of participation over and above non-participation diminished. Music participation had a different effect on marijuana use, requiring participation in another activity to achieve a significant association. Youth who reported participation in all three activities reduced their odds of marijuana use by 86% over non-participants. The implications of these findings are discussed below.
This study attempted to assemble better information about the association between activity participation and delinquency and substance use among North Dakota youth. Parents, school authorities, and community agency directors need to know whether activity participation reduces the probability of youth deviance, as folk wisdom would have us believe. If so, these individuals also deserve to know which combinations of activities can effectively divert youth from deviant activities. Finally, these groups need to know how much time youth need to devote to an activity to effectively prevent delinquency and substance use.
These findings demonstrate that delinquency and substance use were most consistently associated with music participation in the expected direction. Music participation was most strongly associated with avoiding police contact and with binge drinking and was less strongly associated with vandalism and marijuana use. When coupled with other activities, the associations between music participation and deviance were even more impressive.
Sports and school organization participation were more weakly and unevenly associated with delinquency and substance use. Playing sports was associated with lower odds of police contact and marijuana use but was not associated with vandalism and binge drinking. Athletes apparently surmise that marijuana use can hinder performance but heavy alcohol consumption can not. Interestingly, this claim has routinely resonated among at-risk youth in the community when they have been invited to sound off about solutions to youth problems in the community.1 Participating in school organizations was associated with lower odds of binge drinking and marijuana use, but had no effect on either of the delinquency items. Participation in clubs or organizations outside of school was not significantly associated with any of the deviance items.
1The author chairs a city-wide Task Force to address cause of at-risk youth behavior. We have had several groups of at-risk youth inform us at our meetings that athletes are as likely as other groups of youth to binge drink.
The form of these associations appears to be non-linear with the largest change in each slope evident between two pairs of points - non-participation and participation at a minimal level (one to two hours per week). Beyond participating one to two hours per week, the change in slope between different pairs of points diminished. Consequently, while the probability of delinquency and substance use was lowest at the highest level of participation, these data suggest that a minimal level of participation is adequate to efficiently reduce the probability of deviance.
Some narration regarding the limitations of these data is necessary. First, because of the item wording, these data cannot tell us which specific forms of activity participation are most highly associated with a reduction in deviance. We do not know, for instance, whether the odds of delinquency and substance use were lower for singing than playing an instrument. Nor can these data inform us whether the form of sports participation conditions the association between delinquency and substance use (e.g., hockey vs. gymnastics). Second, readers may wonder whether age or gender conditions the form of these associations. They do not. An examination of the interaction terms for these two variables showed that the effects of activity participation on delinquency and substance use did not vary by age or gender. Finally, the cross-sectional nature of the data restricts stronger declarations regarding the causal ordering of these variables. There is little doubt that a history of delinquency and substance use could deter or prevent a youth from playing in the orchestra or being elected to student council. Nevertheless, some activities could be attractive to youth who display an aggressive or reckless disposition (e.g., hockey, playing drums, YMCA basketball). Consequently, it may be prudent to re-examine these questions using panel data.
What are the implications of these findings theoretically and practically? First, activity participation was not associated with delinquency and substance use in a uniform manner. Theories that promote general activity involvement as a means of preventing deviance need to specify that some forms of activity are more effective than others. This may be particularly disconcerting to directors of community organizations (e.g., Boys Club, YMCA) who promote their agencies as a vehicle to reduce troublesome behavior. These agencies may be effective in stimulating and changing other components of adolescent development, but these data demonstrate that they do not disrupt patterns of delinquency and substance use.
In pondering prevention or intervention strategies, parents and school officials should consider encouraging participation in band, choir, or orchestra. When coupled with other activities, music involvement can reduce the odds of delinquency and substance use by 43-86% over non-involvement, net of a youth's background characteristics. While an examination of music's intervening properties is beyond the scope of this paper, it is possible to speculate why musical activities might be associated with lower odds of deviance. It could be that music, more than other forms of participation instills a stronger sense of discipline, patience, and self-control in youth; qualities that are inversely associated with deviance. Music might also provide youth with exposure to a larger array of pro-social peers, thereby facilitating the modeling of positive behaviors. Finally, youth who are not predisposed to delinquency and substance use may self-select for music participation. Controls for pertinent background characteristics should have reduced this possibility, but it is possible that deviance and music are related to some uncontrolled underlying factor. In any event, parents and school authorities would be wise to take stock of music's deterrent properties and encourage musically inclined youth to simultaneously participate in other activities as well.
Agnew, R. (1989). Leisure and delinquency. Social Problems. 36. 332-350.
Cernkovich, S.A. & Giordano, P.G. (1992). School bonding, race, and delinquency. Criminology. 30. 261-291.
Donnelly, P. (1981). Athletes and juvenile delinquents: A comparative analysis based on a review of literature. Adolescence. 16. 415-432.
Empey, L. & Stafford, M. (1991). American Delinquency: Its Meaning and Construction. Belmont, CA: Wadsworth.
Fo, W.S. & O'Donell, C.R. (1975). The buddy system. Effect of community intervention on delinquent offenses. Behavior Therapy. 6. 522-524.
Goldstein, A.P. (1990). Delinquents on Delinquency. Champaign, IL: Research Press.
Hindelang, M.J. (1973). Causes of delinquency: A partical replication. Social Problems. 21. 471-487.
Hirschi, T. (1969) Causes of Delinquency. Berkeley: University of California Press.
Howell, J.C. (1995). Guide for Implementing the Comprehensive Strategy for Serious, Violent, and Chronic Juvenile Offenders. Washington, DC: U.S. Department of Justice. Office of Juvenile Justice and Delinquency Prevention.
Jensen, G.F. & Rojek, D.G. (1998). Delinquency and Youth Crime. Prospect Heights, IL: Waveland Press.
Knox, G. (1998). An Introduction to Gangs. Chicago: New Chicago School Press.
Landers. D.M. & Landers, D.M. (1978). Socialization via interscholastic athletics: Its effects on delinquency and educational attainment. Sociology of Education. 51. 299-303.
McCord, J. (1992). A thirty-year followup of treatment effects. The American Psychologist. 33. 284-289.
Menard, S. (1995). Applied Logistic Regression Analysis. Sage University Paper series on Quantitative Applications in the Social Sciences, 07-107. Thousand Oaks, CA: Sage.
Pakiz, B. Reinherz, H.Z. & Frost, A.K. (1992). Antisocial behavior in adolescence: A community study. Journal of Early Adolescence. 12. 300-313.
Seagrave, J. Moreau. C. & Hastad. D. N. (1985). An investigation into the relationship between ice hockey participation and delinquency. Sociology of Sport Journal. 2. 281-298.
Search Institute. (1992). Search Institute Profiles of Student Life: Attitudes and Behaviors. Minneapolis, MN.
Tierney, J. P. & Grossman, J. B. (1995). Making a Difference: An Impact Study of Big Brothers/Big Sisters. Philadelphia. PA: Public/Private Ventures.
U.S. Department of Justice. (1996). Combating Violence and Delinquency: The National Juvenile Justice Action Plan. Office of Juvenile Justice and Delinquency Prevention. Washington, DC: U.S. Government Printing Office.
[ Home ]
The North Dakota Journal of Human Services, June 1999