Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Dec 1.
Published in final edited form as: Drug Alcohol Depend. 2013 Jul 24;133(2):10.1016/j.drugalcdep.2013.06.010. doi: 10.1016/j.drugalcdep.2013.06.010

Can the school context moderate the protective effect of parental support on adolescents alcohol trajectories in urban Chicago?

Fernando H Andrade 1
PMCID: PMC3818413  NIHMSID: NIHMS509639  PMID: 23891034

Abstract

Background

Research explaining school effects on alcohol use is scare. This study examined the interactive effect between family support and school characteristics (size, poverty, and sector) on adolescents alcohol use trajectories in Chicago.

Methods

Longitudinal and multilevel data were from the Project of Human Development in Chicago Neighborhoods and the Common Core of Data (National Center for Educational Statistics). The sample consisted of 2205 adolescents in 558 schools. A three-level hierarchical linear model was used to estimate multilevel growth curve models and school effects on alcohol trajectories.

Results

In addition to the strong relationship between parental support and alcohol trajectories; the results also found school effects on the average baseline of alcohol use and the rates of change across time. Interestingly, high levels of parental support were more effective in preventing alcohol use in public schools, while adolescents attending private schools with low levels of parental support were more likely to consume alcohol. Similarly, students attending public schools with higher rates of poverty who enjoy higher levels of parental support were less likely to consume alcohol compared to students with lower parental support attending lower rates of schools poverty.

Conclusion

Key findings highlight the importance of the interaction between parental support and school characteristics meaning that protective factors provided by parents could be reinforced or diminished by the school context.

Keywords: parental support, school context, alcohol trajectories

1. Introduction

Two of the most influential contexts in the development of children and adolescents are the family and the school. Despite the importance of schools, research on alcohol has broadly focused on the effect of individual and family characteristics on adolescents alcohol use (Fothergill and Ensminger, 2006; Allen, 2003; Guilamo-Ramos et al., 2005; Bachman et al., 1991; Foxcroft and Lowe, 1991). Within the family context, parental support stands as one of the most important family factors that can prevent alcohol consumption. Family support has been extensively discussed by Maccoby and Martin (1983), who broadly found that parental support and parental control play a substantial role in the socialization process of the child and adolescent.

In a more focused study of alcohol, Foxcroft and Lowe (1995) found that students who perceived more family support were less likely to engage in drinking than students who perceived their parents as authoritarian, negligent and indulgent parenting. Other studies have also found a negative relationship between family support and alcohol consumption (Caldwell et al., 2004; Piko, 2000; Stice et al., 1993).

Within the school context literature, there is a growing trend to account for school contextual effects when studying several areas of social development such as normative and pro-social goals (Meece and Eccles, 2010). However, research about school context effects on substance use is scare. Most of the studies using the term “school effects” refer to adolescents perceptions of their schools characteristics. These types of studies are referred to as “individual-level school-related exposure” (Fletcher et al., 2008). Outside this line of inquiry, to the best of my knowledge, there is only a handful of studies that explore how the school context might be affecting substance using behaviors among youth (Kima and McCarthy, 2006; West et al., 2004; Cleveland and Wiebe, 2003; Kumar et al., 2002).

One of these studies focused on alcohol use explored how the effect of extracurricular activities on alcohol use varies by school characteristics (Hoffmann, 2006). The findings of that study suggest that students attending schools with higher levels of average student SES composition, lower rates of African American and Hispanic composition, and Catholic schools were more likely to use alcohol. Hoffman also examined cross level interactions finding that extracurricular activities have a stronger relationship with alcohol use in students attending low minority and wealthier schools. Another study about school effects on adolescent substance use and health issues by West et al. (2004) shows that schools with higher composition of students who were disengaged from academic activities, knew fewer teachers and schools rated with poor ethos had students who were more likely to drink more alcohol and smoke. All these findings suggest the need for continuing research, especially considering the school context; which can further inform policies and prevention programs related to adolescent alcohol use.

Thus, this paper expands our understanding of school effects on alcohol use among adolescents by adding a longitudinal-multilevel component examining interactions between an aspect of the family context –parental support–and school characteristics. The main focus is on how the protective effects of parental support on adolescents trajectories of alcohol consumption vary depending on three school characteristics: school size, school school sector (public versus private) and levels of poverty. Given the importance of large urban populations, the aims of this paper are explored among adolescents living in Chicago, which is one of the largest urban settings in the U.S., where drug use problems and other risk behaviors such as violence tend to be prevalent. To address this research topic, three research questions are articulated below.

(1) Does the relationship between alcohol consumption and family support depend on the school size?

A comprehensive literature review suggests there are no studies linking school size and alcohol use. However, it can be argued that the middle and small size schools might generate a protective environment against alcohol use. This idea is based on research accounting for positive effects on adolescents behaviors related to alcohol use. For instance, Lee and Smith (1995) found that middle size schools could motivate more student engagement. Other studies have also reported that smaller schools might encourage more: connectedness with school (McNeely et al., 2002), adolescents participation and engagement (Silinsa and Mulfordb, 2004), and attachment with school (Crosnoe et al., 2004). These protective factors seem to be present also at the school level, for instance West et al. (2004) found that school engagement protects against smoking and alcohol use. Thus, it can be expected that adolescents attending larger schools might be more likely to consume alcohol over time. In addition, it can be expected that the protective effect of parental support could be lower in larger schools compared to middle size and small schools.

(2) Does the relationship between alcohol consumption and family support depend on the school sector?

The literature accounting for the effects of school sector on alcohol use among adolescents is also extremely rare. However, one study by Valois et al. (1997) showed differences on unadjusted rates of alcohol, tobacco and illicit drugs consumption between public and private schools. The results were inconsistent depending on the type of substance. In the case of alcohol, only lower rates of consumption were found among females attending public schools compared to females attending private schools. Following these results, and similar to the case of school poverty levels, it could be hypothesized that the protective effect of parental support could be weaker in private schools compared to public schools.

(3) Does the relationship between alcohol consumption and family support depend on the school poverty levels?

Based on the findings reported by Hoffmann (2006), it seems that the effect of school average SES is inconsistent and varies depending on individual factors such as gender. However, it could be expected that schools with higher levels of poverty have lower levels of alcohol consumption. This is based on the idea that schools with lower levels of poverty concentrate students with more socioeconomic resources and wealth. The availability of resources might open more opportunities to consume alcohol. Thus, it can be expected that lower levels of school poverty would generate environments where adolescents might be more likely to engage in alcohol use. It could be also argued that the protective effect of parental support could be less effcientin school with lower levels of poverty. This research question –because of methodological issues explained in the methods section– could be addressed only in public schools.

To address these research questions and given the focus on school–attending urban adolescents, analyses in this paper were based on data from the Project of Human Development in Chicago Neighborhoods (Earls and Visher, 1997; Liberman, 2007). These data are ideal to address the research questions because the study sample is representative of one of the largest urban sectors in the U.S. Also the data are longitudinal and multilevel containing three waves of measurement complemented with school-level information (National Center for Education Statistics, 1998–2007).

2. Methods

2.1. Sample

The PHDCN is a study focused on the pathways to juvenile delinquency, adult crime, substance abuse, and violence (Earls and Visher, 1997). the longitudinal component of PHDCN followed approximately 6,228 randomly selected children and adolescents. This longitudinal component consisted of three assessment waves from 1995 to 2001. The sampling design of the longitudinal study was a three stage cross-sectional stratified cluster sample of Chicago neighborhoods. For a more detail description of the sample design used in PHDCN see Sampson et al. (2005).

The analytic sample was restricted to cohorts 9, 12 and 15 (2205 students nested in 558 schools at wave 1. The distribution of the number of students per school ranged from 1 to 28 with median 2, it had a mean of 4 students per school and standard deviation 3.8). The rest of the cohorts were excluded because they did not have reports of drinking alcohol or participants finished school at the beginning of wave 1. About 50% of the sample was female and almost 49% percent were Hispanic, 37% African American, and 14% white. One out of five students belonged to families with an annual income lower than $10,000 at wave 1, while 40% lived in families with incomes over $30,000. The school data came from the Common Core of Data (National Center for Education Statistics, 1998–2007). The author of this paper gathered school data from the CCD. Students attended both public (69%) and private schools with average size of 710 students, and the percentages of low income students (n=1849 attending 385 public schools), ranged from 12% to 100% with a mean of 79%.

Missing observations ranged from 2% to almost 30% depending on the type of measurement and waves. To address the attrition problem, and under the assumption of missing at random, missing data were treated using multiple imputation procedures (MI). The MI was carried out using the Sequential Regression Imputation Method (Raghunathan et al., 2007).

2.2. Measures

Outcome

The outcome measure was a dichotomous variable accounting for the use of alcohol in the past twelve months. (Earls et al., 2006e,f,g). Participants were asked: “How many days did you have alcoholic beverages to drink, not just a sip or taste during the last 12 months”. The responses ranged in a nine ordinal categories from 0: never, 1–2 days, 3–5, 6–11, 12–24, 25–50, 51–99, 100–199, and 200 or more days. Participants who responded never were coded as 0: not consumed in the past year, the rest of the categories were collapsed into 1: consumed alcohol in the past year.

Time

Time was represented by the age of the adolescent which is centered on the individual mean age of the participant across the three waves. The reason to center this variable is to have zero represent the mean age. Age is expressed as the deviation from the individual mean across waves.

School measurements

Three school characteristics were used in the analysis: sector expressed as a dummy (1: public, 0: private). School size accounts for the number of students attending the school and the percentage of low income students attending the school. The percentage of low income students was defined and constructed only for public schools based on school data from the CCD that accounted for the percentages of students in the school receiving reduced or free lunch at schools. Data about reduced and free lunch were only available for public schools. Both size and percentage of low income students were standardized (mean 0 and standard deviation 1).

Family support

The participant’s family support score came from six items from the Provision of Social Relations Subject (PSRS) measured in wave 1 and 3 (Earls et al., 2006d, 2007). The family support score was categorized into three levels based on two cut o3 points −1 and +1 standard deviations: low, some, and high family support. The highest level was used as the reference category.

Individual covariates

To account for other risk factors that have been found to explain adolescent alcohol consumption the following individual covariates were included. Adolescent’s ethnicity was coded into a three level categorical variable that consisted of White, Hispanic and African American and the comparison category was White. The gender of the participant was expressed as a dummy variable called female (1: female). The socioeconomic status of the adolescent’s family was captured by three measures: parental education, parental occupational prestige and family income. Parental education was constructed as the maximum level of education in the family at wave 1: Less than high school, some high school, completed high school, some tertiary education non college degree, and bachelor degree or higher. Family income was coded into a three level categorical variable: less than $9999, $10000–$19999 (comparison group), and more than $20000. Parental occupation was accounted by the average standardized score of occupational prestige indexes across wave 1, 2 and 3. Higher values of the index represent higher prestige (Earls et al., 2006c,d, 2005).

Sensation seeking trait was represented by the score, at wave 1, from the Emotionality, Activity, Sociability, and Impulsivity Temperament Survey -EASI (Buss and Plomin, 1984; Earls et al., 2006a). Higher values represented higher degrees of sensation seeker trait. The religiosity scale was a standardized score based on items from the Family Environmental Scale FES (Moos and Moos, 1994; Earls et al., 2006b), which reflects the family’s religiosity (higher values represent higher degrees of religiosity). Peers influence was coded into a categorical variable with three levels: low, some, and high levels of peer influence (comparison group). This variable was constructed based on the average score of the Provision of Social Relations Subject scale measured in wave 1 and 3 (Turner et al., 1983; Earls et al., 2006d, 2007).

2.3. Analytic Approach

Hierarchical Generalized Linear Model with a Logit link was used to estimate the growth curve models for the trajectories of alcohol use (Raudenbush et al., 2004; Raudenbush and Bryk, 2002). More specifically, a three level model was used. The first level had the outcome measurement that varies across time age is the time indicator used in the modeling. The second level was composed of time invariant student measurements and the third level consisted of school covariates at wave 1. The school level was based on the attended school at wave 1.

Six models were fitted in order to estimate the effect of school characteristics on adolescent alcohol use. Model 1 fitted the unconditional model incorporating the time structure in the alcohol trajectory. This trajectory was modeled as a linear function of time (the quadratic term was not significant). Model 2 included all student covariates except parental support. Model 3 included parental support. Model 4 incorporated school characteristics modeling both the intercept and the trajectories. Additional models were tested for the cross-level interaction between school characteristics and parental support on both the intercept (starting point) and the slop (trajectories). From all the tested combinations only two models presented significant interactions. These models are model 6 that estimated how school sector moderated the effect of parental support on the intercept (base line) and model 7 that estimated how school levels of poverty, only for public schools, moderated the effect of parental support on the intercept (baseline).

3. Results

3.1. Alcohol trajectories and student level

A simple characterization of the alcohol trajectories is depicted in Figure 1. Adolescent alcohol use increase at age 12 and increments as age increases. Tables 1 and 2 show descriptive statistics for the adolescents and schools’ characteristics by alcohol consumption across the three waves. By the end of wave 1, 14% of the sample had consumed alcohol in the past year. This percentage increased as time passes (i.e. students got older). Twenty five percent of students had consumed alcohol in the past year in wave 2, and by the end of wave 3, the percent of adolescents who consumed alcohol increased to nearly 45%. Students who consumed alcohol belonged to families with weak or low family support.

Figure 1.

Figure 1

Open in a new tab

Average probability of alcohol consumption by age. The y-axis represent probabilities and x-axis represent time (age) (n = 2205).

Table 1.

Parental support and school characteristics by adolescent alcohol use by waves. (a) Significant differences across groups (p < 0.05) within waves for parental support, school size, school sector and percentage of low-income students attending public schools. (b) Standard deviations are in parenthesis, rounded with no decimals.

Wave 1 Wave 2 Wave 3 Total
%No %Yes %No %Yes %No %Yes N
Alcohol consumption 86 14 75 25 56 44 2205
Parental Support
Low family support 72 28 59 41 44 56 271
Some family support 86 14 74 26 54 46 1085
Strong family support 91 9 81 19 65 35 849
School characteristics
School size 897 (611) 1400 (890) 865 (598) 1274 (845) 830 (564) 1136 (795) 2205
Average % low income 80 (18) 73 (19) 80 (18) 74 (19) 81 (17) 75 (19) 1849
Private 83 17 71 29 50 50 356
Public 87 13 76 24 58 42 1849
N 1898 307 1450 483 996 774 2205
Open in a new tab

Table 2.

Individual covariates by alcohol use across all waves. (a) Significant differences across groups (p < 0.05) for age, ethnicity, peer influences, religiosity, and sensation seeker. (b) Significant differences across groups (p < 0.05) for age, ethnicity, peer influences, religiosity, caregiver education, family income, sensation seeker, school size and percent low income. (c) Significant differences across groups (p < 0.05) for age, ethnicity, peer influences, religiosity, and family income. (d) Standard deviations are displayed in parenthesis. (n = 2205).

Wave 1 Wave 2 Wave 3 Total
%No %Yes %No %Yes %No %Yes N
Alcohol consumption 86 14 75 25 56 44 2205
Age 11.54 (2.29) 14.64 (1.25) 13.29 (2.25) 16.18 (1.76) 15.54 (2.23) 17.94 (2.13)
Gender
Female 85 15 75 25 56 44 1109
Male 87 13 75 25 57 43 1096
Ethnicity
Hispanic 86 14 75 25 56 44 1056
Black 89 11 79 21 63 37 821
White 80 20 67 33 43 57 328
Parental Occupation 41.98 (13.97) 42.38 (14.23) 41.99 (13.76) 42.54 (14.03) 42.24 (13.48) 42.69 (13.72)
Parental Education
Less than high school 85 15 279 93 55 45 419
Some high school 85 15 73 27 59 41 432
High school 86 14 76 24 56 44 306
Some tertiary/no bach. 88 12 78 22 57 43 790
Bachelor degreeor more 84 16 69 31 52 48 258
Family Income
Less than 10000 88 12 78 22 62 38 470
Between 10000–19999 88 12 78 22 60 40 443
More than 20000 85 15 73 27 53 47 1292
Sensation seeker 2.72 (0.54) 2.81 (0.53) 2.72 (0.01) 2.8 (0.02) 2.73 (0.49) 2.74 (0.48)
Peer influences
Low 90 10 79 21 66 34 560
Some 86 14 76 24 56 44 1051
Strong 83 17 68 32 47 53 594
Religiosity 61.04 (5.27) 59.74 (5.42) 61.27 (5.01) 59.64 (5.18) 61.56 (4.9) 60.1 (5.01)
N 1898 307 1450 483 996 774 2205
Open in a new tab

The estimations for the growth curve model are presented in Tables 3 and 4. Table 3 displays the results for models with only student time invariant characteristics. Model 1 is the unconditional model with a linear structure for time assuming random intercepts at school levels. This means that the average baseline for the alcohol trajectories were not the same among schools. In the case of the trajectories the model assumed no random effects meaning that the increment of alcohol use was assumed constant across students and schools. The intercept in model 1 was −1.19 (ln(0.304)); thus, for a fourteen-year old, the estimated probability of alcohol consumption was 0.234. The average rate of deviation from the individual mean age was estimated to be 1.404 (log odds ratio), meaning that there is a strong positive association between age and the consumption of alcohol. As expected, the older the students get the higher their probability of alcohol use.

Table 3.

Growth curve model for alcohol use trajectories in Chicago adolescents: Unconditional model and individual level covariates.

Model 1 Model 2 Model 3
Adolescent Level
Intercept 0.304*** 0.468*** 0.383***
Female 0.873* 0.827**
Ethnicity
Hispanic 0.776* 0.771
Black 0.623*** 0.589***
Caregiver Occupation 1.001 1.001
Caregiver Education
Less than high school 1.05 0.976
Some high school 1.048 0.989
High school 0.938 0.887
Some tertiary/non bach. 0.856 0.864
Family Income
Between 10000–19999 0.99 0.981
More than 20000 1.186 1.169
Sensation seeker 1.299*** 1.289***
Peer influences
Low peer influence 0.704*** 0.589***
Some peer influence 0.881*** 0.795***
Religiosity 0.968*** 0.972***
Family support
Low parental support 2.653***
Some parental support 1.62 ***
Age slope 1.404*** 1.418*** 1.455***
Family support
Low family support 0.885*
Some family support 1.002
Random effects for intercept
Between students 0.66 *** 0.62 *** 0.6 ***
Between schools 1.159*** 1.12 *** 1.01 ***
Open in a new tab

Significant difference across groups: p < .10,

*

p < .05,

**

p < .01,

***

p < .001. (n = 2205).

Table 4.

Growth curve model for alcohol use adjusted by school characteristics and family support in Chicago adolescents.

Model 4 Model 5 Model 6
Intercept 0.408*** 0.34 *** 0.396***
Family support
Low family support 2.665*** 4.87 *** 2.766***
Some family support 1.611*** 2.061*** 1.617***
School characteristics
Percent low income 0.835** 0.836** 0.888
School size 1.653*** 1.651*** 1.643***
Public 0.572** 0.737 0.597***
Cross level interactions
Low family support x Public 0.495*
Some family support x Public 0.736
Low fam. support x prcnt low income 0.762*
Some fam. support x prcnt low income 0.927
Age slope 1.441*** 1.507*** 1.503***
Family support
Low parental support 0.896* 0.903* 0.911
Some parental support 1.013 1.017 1.019
School characteristics
Percent low income 0.996
School size 0.967** 0.972** 0.973**
Public 1.067
Random effects for intercept
Between students 0.64 ** 0.63 ** 0.63 **
Between schools 0.65 *** 0.65 *** 0.64 ***
Open in a new tab

Significant difference across groups: p < .10,

*

p < .05,

**

p < .01,

***

p < .001. (n = 2205). For the case of models with parental support the analysis are only for public schools (n = 1849).

The effect of parental support on both the starting point and the slope of the trajectories of alcohol use after controlling for all individual covariates are displayed in Model 3. This results show that adolescents who perceived their parents to be less supportive, low (2.653) and some (1.62) parental support, were more likely to have higher levels of alcohol consumption at the starting point. However, only adolescents who reported low levels of parental support showed a significant negative coeffcientin the slope of the trajectories (ln(0.88)=−.01228). This means that students with lower levels of parental support might have smaller rates of change in their alcohol trajectories. Although the difference is small still suggests that students with high or medium parental support were less likely to consume in the baseline but later have a steeper rate of consumption.

3.2. School level

Table 4 presents the models that adjust by school size, school sector (public versus private), and school levels of poverty (in public schools). This table also presents the cross-level interactions between parental support and school characteristics. Model 4 shows the effects of the three school characteristics on both the intercept and the slope. These three characteristics had a significant coeffcienton the starting point. Students attending private schools, larger schools, and school with lower concentrations of poverty (in the case of public schools) were more likely to have a greater levels of alcohol use at the base line. Per each increment in a standard deviation of school size the initial starting point increases by a factor of 1.65 (log odds). Holding all other factors constant, students attending private schools have a probability of 0.28 of consuming alcohol at the starting point while the probability for students attending public is 0.18. Similar to school size public–school students attending schools with lower levels of poverty increased their chances to consume alcohol at the starting point by .17 per standard deviation in the scale of poverty.

Interestingly, the rate of change in the trajectories of alcohol only depended on the school size. The rate of change was reduced by a factor of 0.033 per each standard deviation increment in the school size.

3.3. Parental support and school characteristics

Key findings reveling the interactions between family support and school sector, and family support and school levels of poverty are presented in Models 5 and 6. These interactions were significant only at the intercept (base line). Given the complexity of the models, findings are easier to interpret showing how the predicted probabilities vary across time depending on family support and these two school characteristics. Thus, Figures 1 and 2 depict these adjusted probabilities, holding all covariates fixed to their mean.

Figure 2.

Figure 2

Open in a new tab

Alcohol use adjusted probability by parental support and school sector. The y-axis represent probabilities and x-axis represent time (age) (n = 2205 nested in 558 schools).

It is interesting to notice that the effect of parental support on the adolescents baseline alcohol use was moderated by the school sector (see Figure 2). On average, students with lower levels of parental support showed a higher starting point; however students with lower levels of parental support who were attending private schools had the highest starting point, while students attending public schools with higher levels of parental support would have the lowest starting point. On average, students with the same level of parental support, had higher probabilities of consuming alcohol across time if they attended private schools.

A second key finding is depicted in Figure 2, which shows the probabilities of alcohol depending on the moderated effect of parental support by public–school poverty levels. The effect of parental support is more protective in the case of students who attend schools with higher levels of poverty. In other words, students attending schools with lower levels of poverty and who had lower levels of parental support were more likely to use alcohol at the base line and across time than student who have same lower levels of parental support but attend schools with higher levels of poverty. It is interesting to note that for higher levels of parental support the differences across the levels of poverty is not as large as the differences in the levels of poverty for student with low parental support. This might imply that parental support has a stronger protective effect in public schools with higher levels of poverty. The implications of key findings are discussed in the last section of this paper.

4. Discussion

The study findings indicate that parental support might be a protective factor for both the baseline and the rate of change of adolescents alcohol trajectories. These findings also suggest that school size, sector, and levels of poverty were related to the baseline, but only school size was related to the average rate of change. In addition, the type of school that students attend might increase or decrease the protective effect of parental support at the beginning of the trajectories. In general, for same levels of parental support, students attending private schools report more consumption across time compared to students attending public schools. Similar results were found in the case of school levels of poverty, where students with similar levels of parental support were more likely to consume alcohol if attending schools with lower levels of poverty.

A partial explanation of these cross level–interactions, could be based on findings showing a positive relationship between alcohol use and SES -family income and parental education- (Goodman and Huang, 2002; Reading, 2002; Blum et al., 2000). The positive relationship between alcohol use and SES suggest that adolescents with more economic resources could be more likely to engage in alcohol use. Similarly, adolescents attending schools with lower levels of poverty might have more financial resources and therefore more purchasing power to acquire alcohol. In other words, the possibility to acquire alcohol might increase if an adolescent attends a school with clusters of students with more economic resources.

It also could be the case that public schools with higher levels of poverty provide the adolescent with peers who might not think about consuming alcohol instead consume other substances. However, this interpretation is speculative and more research is needed to understand the complexity and meaning of school effects on alcohol use.

Similarly, on average private schools are composed by students with more economic resources; thus an adolescent attending a private school might have higher probabilities of engaging in alcohol use. However, the effect of school sector could also reflect some school structural and cultural characteristics. Unfortunately, school data are insufficient to test this interpretation.

There is another explanation –based on peers effects– that can account for the findings related to school sector. Schools are social environments that can provide similar peers for the adolescents (Epstein, 1989); peers can exert influences on adolescents’ behaviors increasing the peer–influences susceptibility to consume alcohol (Schulenberg et al., 1999; Dielman et al., 1987) or providing more opportunities to use alcohol (Gottfredson and Hirschi, 1990). Thus private school can provide mind-liked and more homogeneous peers with more resources and opportunities to consume alcohol decreasing the protective effect of parental support (Steinberg et al., 1994).

As far as the effect of school size, it is plausible that larger schools could be less cohesive environments where students might find: less connectedness with school (McNeely et al., 2002), less participation and engagement (Silinsa and Mulfordb, 2004; Lee and Smith, 1995), and less attachment with school (Crosnoe et al., 2004); thus larger schools might be environments where adolescent might be more likely to consume alcohol. In other words, adolescents attending larger schools are in environments where adolescent may not find as much protection such as school attachment, cohesive environments and opportunities to engage in school activities.

This study findings must be considered within the context of several limitations, some of these are: (i) The lack of more school level information which limits the understanding of the findings related to the school effects on alcohol use. (ii) Other measures of school poverty based on the student body composition (e.g. family income and home resources) could expand the modeling of school poverty to private school. (iii) Alcohol use was modeled using dummies that account for whether the adolescent consumed alcohol in the past year. Despite the advantage of modeling a parsimonious outcome, the estimation of growth curves might be different for other definitions of alcohol use and alcohol abuse or dependence. The specification of the linear relationship between time and alcohol use could not be the most accurate, unfortunately non-linear specifications were not significant mainly because of the focus on school context where the consumption of alcohol might be mostly linear and monotonic.

Notwithstanding these limitations, the use of a large longitudinal dataset and the availability of school data allowed for a unique examination of how parental support interacts with school characteristics in shaping alcohol use over time. The work in this paper resonates with the works of Perra et al. (2012) and West et al. (2004) supporting the public health importance of considering the school context and experiences in schools as factors that expand our understanding of adolescent substance use and other health issues.

As the growing literature on school effects in non–academic areas has been claiming, further research is needed to expand our knowledge of the ways by which schools interact with what students bring from home, and how all these might influence youth development.

Figure 3.

Figure 3

Open in a new tab

Alcohol use adjusted probability by parental support and school poverty levels for students attending only public schools. LPS:Low parental support, SPS:Some parental support, HPS:High parental support. LLP:Lower levels of poverty, MLP:Medium levels of poverty, HLP:High levels of Poverty. The y-axis represent probabilities and x-axis represent time (age) (n = 1849 nested in 385 schools).

Acknowledgments

Role of Funding Source

This investigation was supported by the National Institutes of Health under Ruth L. Kirschstein National Research Service Award T32 DA007267. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.

Footnotes

Contributors

The only author of this manuscript was responsible for all steps in the preparation of the manuscript from the conceptualization of the problem, data analysis, writing and others. Data are public available from ICPSR http://www.icpsr.umich.edu/icpsrweb/PHDCN/ and the Department of Education Statistics http://nces.ed.gov/ccd/.

Conflict of Interest

The author declares having no conflicts of interest.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Allen M. Comparing the influence of parents and peers on the choice to use drugs –a meta-analytic summary of the literature. Criminal Justice and Behavior. 2003;30 (2):163–186. [Google Scholar]
  2. Bachman JG, Wallace JM, O’Malley PM, Johnston LD, Kurth CL, Neighbors HW. Racial/ethnic differences in smoking, drinking, and illicit drug use among american high school seniors, 1976–89. American Journal of Public Health. 1991;81 (3):372–377. doi: 10.2105/ajph.81.3.372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Blum RW, Beuhring T, Shew ML, Bearinger LH, Sieving RE, Resnick MD. The effects of race/ethnicity, income, and family structure on adolescent risk behaviors. American Journal of Public Health. 2000;90 (12):18791884. doi: 10.2105/ajph.90.12.1879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Buss A, Plomin R. Temperament: Early Developing Personality Traits. Erlbaum; 1984. [Google Scholar]
  5. Caldwell CH, Sellers RM, Bernat DH, Zimmerman MA. Racial identity, parental support, and alcohol use in a sample of academically at-risk african american high school students. American Journal of Community Psychology. 2004;34 (1):71–82. doi: 10.1023/b:ajcp.0000040147.69287.f7. [DOI] [PubMed] [Google Scholar]
  6. Cleveland HH, Wiebe RP. The moderation of adolescenttopeer similarity in tobacco and alcohol use by school levels of substance use. Child Development. 2003;74:279–291. doi: 10.1111/1467-8624.00535. [DOI] [PubMed] [Google Scholar]
  7. Crosnoe R, Johnson M, Elder GHJ. School size and the interpersonal side of education: An examination of race/ethnicity and organizational context. Social Science Quarterly. 2004;85 (5):1259–1274. [Google Scholar]
  8. Dielman T, Campanelli PC, Shope JT, Butchart AT. Susceptibility to peer pressure, self-esteem, and health locus of control as correlates of adolescent substance abuse. Health Education & Behavior. 1987;14 (2):207–221. doi: 10.1177/109019818701400207. [DOI] [PubMed] [Google Scholar]
  9. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2005. Project on human development in chicago neighborhoods (phdcn): Master file, wave 2, 1997–2000. [DOI] [Google Scholar]
  10. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2006a. Project on human development in chicago neighborhoods (phdcn): Emotionality, activity, sociability, and impulsivity temperament survey, wave 1 1994 1995. [DOI] [Google Scholar]
  11. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2006b. Project on human development in chicago neighborhoods (phdcn): Family environment scale, wave 1, 1994–1995. [DOI] [Google Scholar]
  12. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2006c. Project on human development in chicago neighborhoods (phdcn): Master file, wave 1, 1994–1997. http://dx.doi.org/10.3886/ICPSR13580.v3. [Google Scholar]
  13. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI, : 2006d. Project on human development in chicago neighborhoods (phdcn): Provision of social relations (subject), wave 1, 1994–1995. http://dx.doi.org/10.3886/ICPSR13598.v1. [Google Scholar]
  14. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2006e. Project on human development in chicago neighborhoods (phdcn): Substance use, wave 1, 1994–1997. http://dx.doi.org/10.3886/ICPSR13602.v1. [Google Scholar]
  15. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI, : 2006f. Project on human development in chicago neighborhoods (phdcn): Substance use, wave 2, 1997–2000. http://dx.doi.org/10.3886/ICPSR13659.v1. [Google Scholar]
  16. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2006g. Project on human development in chicago neighborhoods (phdcn): Substance use, wave 3, 2000–2002. [DOI] [Google Scholar]
  17. Earls FJ, Brooks-Gunn J, Raudenbush SW, Sampson RJ. Tech rep. Inter-university Consortium for Political and Social Research (ICPSR) [distributor], ICPSR; Ann Arbor, MI: 2007. Project on human development in chicago neighborhoods (phdcn): Provision of social relations (subject and young adult), wave 3, 2000–2002. [DOI] [Google Scholar]
  18. Earls FJ, Visher CA. Tech Rep NCJ163603. 9. Department of Justice, Office of Justice Programs, National Institute of Justice; Washington, D.C: 1997. Project on human development in chicago neighborhoods: A research update [nij research in brief] [Google Scholar]
  19. Epstein JL. The selection of friends: Changes across the grades and in different school environments. 4. John Wiley & Sons; 1989. pp. 158–187. [Google Scholar]
  20. Fletcher A, Bonell C, Hargreaves J. School effects on young people’s drug use: A systematic review of intervention and observational studies. Journal of Adolescent Health. 2008;42 (3):209–220. doi: 10.1016/j.jadohealth.2007.09.020. [DOI] [PubMed] [Google Scholar]
  21. Fothergill KE, Ensminger ME. Childhood and adolescent antecedents of drug and alcohol problems: A longitudinal study. Drug and Alcohol Dependence. 2006;82 (1):61– 76. doi: 10.1016/j.drugalcdep.2005.08.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Foxcroft DR, Lowe G. Adolescent drinking behaviour and family socialization factors: a meta-analysis. Journal of Adolescence. 1991;14 (3):255–273. doi: 10.1016/0140-1971(91)90020-r. [DOI] [PubMed] [Google Scholar]
  23. Foxcroft DR, Lowe G. Adolescent drinking, smoking and other substance use involvement: links with perceived family life. Journal of Adolescence. 1995;18 (2):159–177. [Google Scholar]
  24. Goodman E, Huang B. Socioeconomic status, depressive symptoms, and adolescent substance use. Archives of Pediatrics & Adolescent Medicine. 2002;156 (5):448–453. doi: 10.1001/archpedi.156.5.448. [DOI] [PubMed] [Google Scholar]
  25. Gottfredson M, Hirschi T. A general theory of crime. Stanford University Press; 1990. [Google Scholar]
  26. Guilamo-Ramos V, Jaccard J, Turrisi R, Johansson M. Parental and school correlates of binge drinking among middle school students. American Journal of Public Health. 2005;95 (5):894–899. doi: 10.2105/AJPH.2003.018952. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hoffmann JP. Extracurricular activities, athletic participation, and adolescent alcohol use: Gender-differentiated and school-contextual effects. Journal of health and social behavior. 2006;47 (3):275–290. doi: 10.1177/002214650604700306. [DOI] [PubMed] [Google Scholar]
  28. Kima J, McCarthy WJ. School-level contextual influences on smoking and drinking among asian and pacific islander adolescents. Drug and Alcohol Dependence. 2006;84:56–68. doi: 10.1016/j.drugalcdep.2005.12.004. [DOI] [PubMed] [Google Scholar]
  29. Kumar R, O’Malley PM, Johnston LD, Schulenberg JE, Bachman JG. Effects of school-level norms on student substance use. Prevention Science. 2002;3 (2):105–124. doi: 10.1023/a:1015431300471. [DOI] [PubMed] [Google Scholar]
  30. Lee V, Smith JB. Effects of high school restructuring and size on early gains in achievement and engagement. Sociology of Education. 1995;68:241–270. [Google Scholar]
  31. Liberman A. Tech Rep NCJ 217397. Department of Justice, Office of Justice Programs, National Institute of Justice; Washington, D.C: 2007. Adolescents, neighborhoods, and violence: recent findings from the project on human development in chicago neighborhoods [nij research in brief] [Google Scholar]
  32. Maccoby E, Martin J. Socializing in the context of family: parent-child interaction. 4. Wiley & Sons; 1983. pp. 1–101. [Google Scholar]
  33. McNeely CA, Nonnemaker JM, Blum RW. Promoting school connectedness: Evidence from the national longitudinal study of adolescent health. Journal of School Health. 2002;72 (4):138–146. doi: 10.1111/j.1746-1561.2002.tb06533.x. [DOI] [PubMed] [Google Scholar]
  34. Meece JL, Eccles JS. Handbook of research on schools, schooling and human development. Routledge: 2010. [Google Scholar]
  35. Moos R, Moos B. Family Environment Scale Manual: Development, Applications, Research. 3. Consulting Psychologist Press; 1994. [Google Scholar]
  36. National Center for Education Statistics. Common core of data (ccd) 1998–2007 Http://nces.ed.gov/ccd/
  37. Perra O, Fletcher A, Bonell C, Higgins K, McCrystal P. School-related predictors of smoking, drinking and drug use: Evidence from the belfast youth development study. Journal of Adolescence. 2012;35 (2):315–324. doi: 10.1016/j.adolescence.2011.08.009. [DOI] [PubMed] [Google Scholar]
  38. Piko B. Perceived social support from parents and peers: Which is the stronger predictor of adolescent substance use? Substance Use & Misuse. 2000;35 (4):617–630. doi: 10.3109/10826080009147475. [DOI] [PubMed] [Google Scholar]
  39. Raghunathan T, Solenberger P, Van Hoewyk J. Tech rep, Survey Methodology Program. Survey Research Center, Institute for Social Research-University of Michigan; Ann Arbor, MI: 2007. Iveware: imputation and variance estimation software guide. [Google Scholar]
  40. Raudenbush SW, Bryk A, Cheong YF, Congdon R, du Toit M. Hierarchical linear and nonlinear modeling. Scientific Software Int; 2004. [Google Scholar]
  41. Raudenbush SW, Bryk AS. Hierarchical linear models: applications and data analysis methods. Sage Publications; 2002. [Google Scholar]
  42. Reading R. Socioeconomic status, depressive symptoms and adolescent substance use. Child: Care, Health and Development. 2002;28 (5):434–434. [Google Scholar]
  43. Sampson RJ, Morenoff JD, Raudenbush S. Social anatomy of racial and ethnic disparities in violence. American Journal of Public Health. 2005;95 (2):224–232. doi: 10.2105/AJPH.2004.037705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Schulenberg J, Maggs JL, Dielman TE, Leech SL, Kloska DD, Shope JT, Laetz VB. On peer influences to get drunk: A panel study of young adolescents. Merrill–Palmer Quarterly. 1999;45 (1):108–142. [Google Scholar]
  45. Silinsa H, Mulfordb B. Schools as learning organisations – effects on teacher leadership and student outcomes. School Effectiveness and School Improvement: An International Journal of Research, Policy and Practice. 2004;15 (3):443–466. [Google Scholar]
  46. Steinberg L, Fletcher A, Darling N. Parental monitoring and peer influences on adolescent substance use. Pediatrics. 1994;93 (6):1060–1064. [PubMed] [Google Scholar]
  47. Stice E, Barrera M, Chassin L. Relation of parental support and control to adolescents’ externalizing symptomatology and substance use: A longitudinal examination of curvilinear effects. Journal of Abnormal Child Psychology. 1993;21 (6):609–629. doi: 10.1007/BF00916446. [DOI] [PubMed] [Google Scholar]
  48. Turner R, Frankel B, Levin D. Social support: Conceptualization, measurement, and implications for mental health. 4. JAI Press; 1983. pp. 67–111. [Google Scholar]
  49. Valois RF, Thatcher WG, Drane JW, Reininger BM. Comparison of selected health risk behaviors between adolescents in public and private high schools in south carolina. Journal of School Health. 1997;67 (10):434–440. doi: 10.1111/j.1746-1561.1997.tb01292.x. [DOI] [PubMed] [Google Scholar]
  50. West P, HSAL School effects on pupils’ health behaviours: evidence in support of the health promoting school. Research papers in Education. 2004;19 (3):261–291. [Google Scholar]

RESOURCES