Content uploaded by Elena Nasaescu
Author content
All content in this area was uploaded by Elena Nasaescu on Feb 12, 2023
Content may be subject to copyright.
International Journal of Drug Policy 112 (2023) 103946
Contents lists available at ScienceDirect
International Journal of Drug Policy
journal homepage: www.elsevier.com/locate/drugpo
Research Paper
A longitudinal study of protective factors against substance use in early
adolescence. An ecological approach
Joaquín Rodríguez-Ruiz
∗
, Izabela Zych , Vicente J. Llorent , Inmaculada Marín-López ,
Raquel Espejo-Siles, Elena Nasaescu
University of Cordoba, Spain
Keywords:
Substance use
Adolescence
Ecological approach
Longitudinal study
Background: Adolescent substance use has been widely related to dierent individual, school, family and commu-
nity factors. Yet, the number of studies with all these variables together in a model from an ecological perspective
is still low, and they rarely used a longitudinal design. The aim of this study was to explore, from an ecological
perspective, the prospective impact of dierent individual, school, family and neighbourhood factors on adoles-
cent substance use.
Methods: This was a longitudinal study with a one-year follow up. There were 881 participants (M
age
= 12.57;
48.1% females) at wave 1, of which 686 (M
age
= 13.51; 51.8% females) were followed-up at wave 2. Validated
questionnaires were used for data collection.
Results: Regression analyses showed that higher substance use was predicted by high family socio-economic
status cross-sectionally and longitudinally, and it was related to low neighbourhood socio-economic status cross-
sectionally only. Participants who disliked school and had a poor academic performance were more likely to use
substances, both cross-sectionally and longitudinally.
Conclusions: The inclusion of families in substance use prevention programmes could be a key component in these
interventions. Moreover, promotion of a positive school climate could protect adolescents from using substances.
Research projects usually address adolescent substance use tak-
ing into account personal ( Rodríguez-Ruiz et al., 2021 ), interpersonal
( Foster & Spencer, 2013 ) or contextual ( Kipping et al., 2015 ) risk or
protective factors, but rarely approach substance use from a holis-
tic and more complex perspective. An Ecological Theory proposed by
Bronfenbrenner (1979) includes dierent social and interpersonal en-
vironments that inuence behaviours and human development. These
levels include individual, school, family and neighbourhood domains,
which are likely to impact adolescent substance use, but their inuence
still need to be thoroughly described.
Individual factors
Individual factors such as sex, age or academic performance have
been related to adolescent substance use. Some studies showed higher
prevalence rates of substance use in boys in comparison to girls
( Halladay et al., 2020 ; Lee et al., 2021 ), but sex dierences have de-
creased in the past decade ( Kraus et al., 2018 ). Spanish adolescent fe-
males reported even higher levels of alcohol and tobacco use than males
∗ Corresponding author at: Department of Psychology, University of Cordoba, Avda. San Alberto Magno s/n, 14004, Cordoba, Spain.
E-mail address: m42roruj@uco.es (J. Rodríguez-Ruiz) .
(
Moreno et al., 2020 ). Age is another important individual factor that
has been broadly studied in relation to substance use. The STUDES Re-
port ( Ministerio de Sanidad, 2021 ) showed that 14 years of age was
the mean onset of the most popular substances (alcohol and tobacco)
used. Moreover, recent research has shown that substance use is higher
among older students ( Zych et al., 2020 ). Thus, most studies show that
substance use is more common in males and older adolescents, but nd-
ings are inconsistent.
Regarding academic performance, previous studies found that illicit
drug use was signicantly related to school dropout ( Brière et al., 2014 )
and alcohol users were less likely to continue their studies later in life
( Arria et al., 2020 ). Low academic performance also increases the like-
lihood of alcohol and illicit drug use according to data reported by Nor-
wegian ( Heradstveit et al., 2017 ) and US students ( Meda et al., 2017 ).
School-based relationships and connectedness
Social Bond Theory ( Hirschi, 1969 ) states that social control and
cohesion prevent adolescents from delinquent behaviour and increase
their likelihood to behave according to social standards. In line with
https://doi.org/10.1016/j.drugpo.2022.103946
0955-3959/© 2022 Published by Elsevier B.V.
J. Rodríguez-Ruiz, I. Zych, V.J. Llorent et al. International Journal of Drug Policy 112 (2023) 103946
this theory, some studies found a relationship between dierent school
variables and adolescent substance use. It was found that a posi-
tive school climate delays substance use onset ( Daily et al., 2020 ).
King et al. (2020) reported higher odds of cannabis use among African
American male students who disliked going to school. Low school con-
nectedness has been associated with opioid use in Canada ( Syed et al.,
2021 ) and marijuana use in the USA ( Mulla et al., 2020 ). Therefore,
most studies in the eld show that a positive school climate can be pro-
tective against substance use.
Relationships with teachers and classmates also have an impact
on adolescent substance use. Wenzel et al. (2009) included the
promotion of desirable student-teacher relationships in a drug pre-
vention programme in Germany and obtained desirable outcomes.
Daily et al. (2020) found that bonding with teachers was a strong lon-
gitudinal protective factor against substance use. The relation between
bonding with classmates and substance use has barely been reported by
scientic literature. School friendships ( Forster et al., 2015 ) and peer
support ( Rodzlan et al., 2021 ) have been identied as protective factors
against substance use, but little is known about the possible protective
role of bonding with classmates.
Family and neighbourhood status
Family socio-economic status has also been related to adolescent sub-
stance use. Gerra et al. (2020) identied low socio-economic status as
a risk factor for episodic and frequent substance use among 16-year-old
students from 28 European countries. Low family socio-economic status
also predicted substance use in a sample of Czech high school students
in a structurally disadvantaged region ( Petruzelka et al., 2020 ).
Neighbourhood is another context of inuence in adolescent be-
haviour including substance use. Tucker et al. (2013) found that high
rates of neighbourhood unemployment, as an indicator of low socio-
economic status, increased the odds of marijuana use and binge drink-
ing. Lee et al. (2018) conducted a longitudinal project according to
which neighbourhood stability in childhood decreased the likelihood of
alcohol and cannabis use later in adulthood. The results of the studies
described above indicate a link between low socio-economic status and
adolescent substance use. However, new longitudinal studies are needed
to test the impact of the family and neighbourhood socio-economic sta-
tus on substance use and intoxication combined with other possible risk
and protective factors.
Ecological perspective
Despite the considerable amount of scientic data relating adoles-
cent substance use to individual, school, family and neighbourhood fac-
tors, these relationships still need to be analysed together to approach
substance use from an ecological perspective. There are only several
studies that used the ecological perspective to explain substance use.
Among them, a cross-sectional study with gang-involved US adoles-
cents in Grades 8, 10 and 12 found less family rules, more access to
drugs in the neighbourhood and higher acceptance of substance use by
friends and family among participants who use substances more fre-
quently ( Bishop et al., 2020 ). Another cross-sectional study conducted
by Connell et al. (2010) with a sample of US adolescents identied indi-
vidual (being male, higher academic performance and low antisocial be-
haviour), family (parental monitoring and parental disapproval of con-
sumption) and community (availability of substances) protective factors
against substance use. A longitudinal study by Shih et al. (2017) showed
that neighbourhood disorganization at age 16 was the strongest predic-
tor of alcohol, tobacco and other substances use one year later, followed
by relationships with peers. Thus, research studies focused on adolescent
substance use from an ecological perspective are usually conducted with
specic populations, rarely use a longitudinal design and, when they do,
the number of factors measured is low.
The current study
There is a fruitful body of research about substance use and dierent
risk and protective factors, but most of these studies focused on specic
populations or did not explore the use of dierent substances. Moreover,
it is still necessary to approach substance use from an ecological per-
spective, analysing unique relations of dierent risk and protective fac-
tors with the use of dierent substances, which would provide a better
understanding of this phenomenon. Some previous studies approached
substance use from an ecological perspective, albeit chronological rela-
tions could not be established in most of them because they used cross-
sectional designs. Most of the past results cannot be generalized given
that participants had a particular prole such as gang-involved youth
( Bishop et al., 2020 ), non-metropolitan students ( Connell et al., 2010 )
or girls involved in justice system ( Staton et al., 2020 ). Thus, the objec-
tive of the current study was to explore, from an ecological perspective,
the cross-sectional and prospective impact of dierent factors -including
individual, school, family and neighbourhood- on the use of dierent
licit and illicit substances and intoxication in early adolescence.
Based on the scientic literature, we hypothesised that: i. rates of
substance use would be higher among boys, older students and partici-
pants with poorer academic performance; ii. low levels of liking school,
bonding with teachers and bonding with classmates would be related to
more substance use; iii. Low neighbourhood and family socio-economic
status would be risk factors for substance use.
Method
Participants
The original sample included 905 participants, but students with
more than 33% of missing data were excluded. In the nal sample, there
were 881 participants (48.1% females, 51.9% males) at wave 1 (W1) se-
lected by convenience from schools in the province of Cordoba (Spain).
Participants were enrolled in Grade 1 and Grade 2 of Secondary Edu-
cation, with a mean age of 12.57 years ( SD = 0.80). Regarding socio-
economic status, 94% self-identied their families as neither rich nor
poor, 5% declared their families were rich or very rich, and 1% poor or
very poor. Moreover, 88.3% considered their neighbourhood as neither
rich nor poor, 8.3% rich or very rich and 3.4 % poor or very poor.
At wave 2 (W2), 686 participants (51.8% females, 48.2% males)
were followed-up. At this wave they were enrolled in Grade 2 and Grade
3 of Secondary Education ( M
age
= 13.51; SD = 0.72). Thus, 78% of the
participants were followed-up one year later.
Instruments
Individual characteristics were measured by asking the participants
about their current age (continuous variable), school year (1 = Grade 1,
2 = Grade 2, 3 = Grade 3) and sex (0 = female, 1 = male).
To explore socio-economic status, participants were asked, “When
compared to other Spanish families, I consider my family as ”and “When
compared to other Spanish neighbourhoods, I consider my neighbour-
hood as ”. These items were answered on a ve-point Likert scale includ-
ing: 5 = “very rich ”, 4 = “rich ”, 3 = “neither rich nor poor ”, 2 = “poor ”,
1 = “very poor ”.
Substance use was measured using a subscale ( Ω = .94, 𝛼= .93 at
W1; Ω = .92, 𝛼= .92 at W2) of the Self-Reported Antisocial Behavior
Questionnaire (SRA; Loeber et al., 1989 ) with some additional items. The
original scale had seven items that measured beer, wine, strong alcohol
(whisky, rum, vodka, gin), tobacco, cannabis and other illicit drug use.
Another item to measure cocaine use was added. This questionnaire was
answered on a four-point scale (1 = never; 2 = yes, once; 3 = yes, twice;
4 = yes, more times) and focused on substance use during the past school
year. Beer and wine were grouped together as “soft alcohol ”and cocaine
and other illicit drugs were grouped together as “other illicit drugs ”.
2
J. Rodríguez-Ruiz, I. Zych, V.J. Llorent et al. International Journal of Drug Policy 112 (2023) 103946
A brief scale ( 𝛼= .94, Ω = .94 at W1; Ω = .95, 𝛼= .95 at W2) in-
cluding three items created ad hoc was used to measure substance intoxi-
cation . Items included: “Have you ever got drunk with alcohol? ”, “Have
you ever been heavily aected by any drug (excluding alcohol)? ”, and
“Have you ever drunk a lot and quickly to get drunk? ”. Items were an-
swered on a scale (1 = never; 2 = yes, once; 3 = yes, twice; 4 = yes,
more times).
A three-factor questionnaire ( 𝛼= .90; Ω = .90) was administrated to
measure school climate factors . The instrument was designed by the z-
proso project team ( Ribeaud & Eisner, 2010 ) and used in recent studies
(e.g. Zych et al., 2021 ), showing adequate psychometric properties. The
three dimensions were liking school ( 𝛼= .82; Ω = .82), with items such
as “I like going to school ”; bonding with teachers ( 𝛼= .82; Ω = .82),
with items such as “My teachers are fair with me ”; and bonding with
classmates ( 𝛼= .83; Ω = .84) with items such as “I get along with other
kids ”. Items were answered using a scale including 1 = never, 2 = almost
never, 3 = sometimes, 4 = usually, and 5 = always .
Different aspects of academic performance were measured with two
questions: “How many times were you expelled from school in the last
year? ”with an open answer, and “What grade do you usually achieve? ”.
This last question was answered according to four options in the Spanish
education system: 1 = “fail ”, 2 = “pass ”, 3 = “very good ”, and 4 = “out-
standing ”.
Design and procedure
This was a prospective longitudinal study with two waves of data
collection over two consecutive school years. School board directors
were contacted to ask for their collaboration in the research and, if
they agreed, parents were asked for collaboration and parental consent
forms were collected. Data were collected in October and November
2020 (W1) and October and November 2021 (W2). Participants lled
in the questionnaires in schools under the supervision of the research
team. In seven out of sixteen schools, data were collected online at W1
because they did not have enough space to guarantee social distancing
during the COVID-19 pandemic. In these cases, students were super-
vised by their teachers, who previously received instructions from the
research team. An anonymous code was used to match participants at
W1 and W2. The current study was approved by the Ethics Commit-
tee of the University of Blinded for Peer Review. The study followed all
the national and international ethical standards including Declaration
of Helsinki and data protection laws.
Data analyses
Descriptive analyses were conducted to describe socio-demographic
characteristics of the sample. Linear and ordinal regression analyses
were run to explore unique predictors of substance use. Linear regres-
sion analyses were used for alcohol use, illicit substance use and in-
toxication, while ordinal regression analyses were used for tobacco and
cannabis use, given that these two substances were measured with a sin-
gle item. The dependent variables were all substances and intoxication
at wave 1; whereas independent variables were wave 1 age, sex, family
SES, neighbourhood SES, number of school exclusions due to miscon-
duct, grades, liking school, bonding with teachers and bonding with
classmates. Other regression analyses were run with the same indepen-
dent variables at wave 1 and all substances and intoxication at wave 2
as dependent variables in order to test longitudinal predictors. Regres-
sion analyses at wave 2 included previous substance use as predictor
of substance use one year later. Previous substance use was a dichoto-
mous variable (0 = no past use, 1 = use at least once of at least one
substance). All these analyses were run using SPSS version 25 software.
Instrument´s reliability was tested by calculating polychoric alpha and
Mcdonald´s omega for each scale using FACTOR software ( Lorenzo-Seva
& Ferrando, 2006 ).
Results
Cross-sectional factors related to substance use
Tables 1 and 2 shows dimensions related to substance use and intox-
ication at wave 1. Soft alcohol use was related to family socio-economic
status ( 𝛽= .16, p = .000), neighbourhood socio-economic status ( 𝛽= -
.09, p = .013), liking school ( 𝛽= -.13, p = .004) and bonding with teach-
ers ( 𝛽= -.20, p < .001). Strong alcohol use was linked to family socio-
economic status ( 𝛽= .15, p < .001), school exclusions ( 𝛽= .17, p < .001),
grades ( 𝛽= -.16, p < .001), liking school ( 𝛽= -.12, p = .006) and bond-
ing with teachers ( 𝛽= -.09, p = .042). Neighbourhood socio-economic
status (Est = -1.188, p = .016), school exclusions (Est = .505, p = .048),
grades (Est = -1.340, p < .001) and liking school (Est = -.610, p = .003)
were associated with tobacco use. The likelihood of illicit drugs use was
higher among participants reporting low levels of neighbourhood socio-
economic status ( 𝛽= -.08, p = .031), bonding with teachers ( 𝛽= -.10,
p = .034) and bonding with friends ( 𝛽= -.11, p = .009). Intoxication was
related to neighbourhood socio-economic status ( 𝛽= -.08, p = .020),
school exclusions ( 𝛽= .09, p = .012), grades ( 𝛽= -.10, p = .009) and
bonding with friends ( 𝛽= -.11, p = .009).
Longitudinal predictors of substance use
As can be seen in Tables 3 and 4 higher family socio-economic sta-
tus predicted all substances use one year later: soft alcohol ( 𝛽= .09,
p = .028), strong alcohol ( 𝛽= .10, p = .007), tobacco (Est = 1.506,
p = .002), cannabis (Est = 1.822, p = .007), other illicit drugs ( 𝛽= .12,
p = .004). School exclusions were longitudinally related to strong al-
cohol ( 𝛽= .08, p = .029) and intoxication ( 𝛽= .09, p = .027). Low
grades predicted soft alcohol ( 𝛽= -.09, p = .020), strong alcohol ( 𝛽= -
.13, p = .001), tobacco (Est = -.830, p < .001), cannabis (Est = -.704,
p = .036) and intoxication ( 𝛽= -.15, p < .001). Soft alcohol ( 𝛽= -.14,
p = .006), strong alcohol ( 𝛽= -.13, p = .006) and intoxication ( 𝛽= -
.13, p = .009) were predicted negatively predicted by liking school the
previous year.
Table 1
Cross-sectional predictors of alcohol use, illicit drug use and intoxication.
Soft alcohol Strong alcohol Other illicit drugs Intoxication
Beta t p Beta t P Beta t p Beta t p
Age .10 2.62 .009 .10 2.88 .004 -.01 -0.29 .772 .09 2.18 .029
Sex .07 2.04 .042 -.03 -0.79 .431 -.01 -0.05 .959 -.03 -0.83 .410
Socio-economic status (family) .16 4.61 .000 .15 4.41 < .001 .02 0.58 .561 .065 1.79 .074
Socio-economic status (neighbourhood) -.09 -2.52 .012 -.04 -1.16 .246 -.08 -2.17 .031 -.08 -2.32 .020
School exclusions .02 0.44 .663 .17 4.80 < .001 .03 0.65 .514 .09 2.52 .012
Grade -.06 -1.65 .099 -.16 -4.24 < .001 -.04 -1.04 .300 -.10 -2.64 .009
Liking school -.13 -2.93 .004 -.12 -2.76 .006 .01 0.06 .953 -.05 -1.02 .308
Bonding with teachers -.20 -4.17 < .001 -.09 -2.03 .042 -.10 -2.12 .034 -.01 -0.12 .908
Bonding with friends .06 1.55 .122 -.01 -0.12 .902 -.11 -2.63 .009 -.11 -2.63 .009
3
J. Rodríguez-Ruiz, I. Zych, V.J. Llorent et al. International Journal of Drug Policy 112 (2023) 103946
Table 2
Cross-sectional predictors of tobacco and cannabis use.
Tobacco Cannabis
Est. SE p Est. SE p
Age 0.254 0.193 .187 0.476 0.385 .217
Sex -0.249 0.322 .439 0.366 0.691 .596
Socio-economic status (family) 0.539 0.624 .388 0.206 1.007 .838
Socio-economic status (neighbourhood) -1.188 0.493 .016 -1.270 0.878 .148
School exclusions 0.505 0.255 .048 0.111 0.562 .843
Grade -1.340 0.247 < .001 -0.360 0.489 .462
Liking school -0.610 0.204 .003 -0.672 0.478 .160
Bonding with teachers -0.140 0.207 .497 -0.300 0.390 .442
Bonding with friends 0.287 0.202 .157 -0.538 0.370 .146
Table 3
Prospective predictors of alcohol use, illicit drug use and intoxication one year later.
Soft alcohol Strong alcohol Other illicit drugs Intoxication
Beta t p Beta t p Beta t p Beta t p
Age .05 1.28 .202 .03 0.78 .434 -.01 -0.06 .951 .04 0.97 .334
Sex (male) -.03 -0.87 386 -.13 -3.60 .000 -.03 -0.71 .476 -.16 -4.24 .000
Socio-economic status (family) .09 2.20 .028 .10 2.73 .007 .12 2.92 .004 .07 1.76 .079
Socio-economic status (neighbourhood) .03 0.73 .463 .03 0.71 .48 -.03 -0.82 .413 -.04 -1.13 .260
School exclusions -.01 -0.02 .984 .08 2.20 .029 .05 1.17 .242 .09 2.22 .027
Grade -.09 -2.33 .020 -.13 -3.34 .001 -.07 -1.57 .117 -.15 -3.72 .000
Liking school -.14 -2.74 .006 -.13 -2.73 .006 -.09 -1.64 .102 -.13 -2.62 .009
Bonding with teachers .03 0.47 .637 .01 0.26 .79 .01 0.15 .879 .05 0.89 .374
Bonding with friends .04 0.85 .397 .04 0.82 .42 -.03 -0.66 .511 -.03 -0.69 .491
Previous substance use .35 8.51 .000 .34 8.45 .000 .04 0.97 .334 .24 5.83 .000
Table 4
Prospective predictors of tobacco and cannabis use one year later.
Tobacco Cannabis
Est. SE p Est. SE p
Age 0.069 0.181 .702 0.024 0.310 .939
Sex -0.925 0.295 .002 -0.741 0.495 .135
Socio-economic status (family) 1.506 0.491 .002 1.822 0.679 .007
Socio-economic status (neighbourhood) -0.319 0.449 .477 0.014 0.734 .985
School exclusions 0.189 0.338 .576 0.687 0.408 .092
Grade -0.830 0.195 < .001 -0.704 0.336 .036
Liking school -0.297 0.184 .105 -0.559 0.307 .069
Bonding with teachers 0.139 0.192 .470 0.437 0.299 .143
Bonding with friends -0.178 0.177 .314 -0.063 0.292 .828
Previous substance use 1.261 0.299 < .001 1.620 0.524 .002
Discussion
Adolescent substance use is a global health concern ( Hall et al.,
2016 ). Although this phenomenon and its possible protective factors
have been widely studied in cross-sectional projects, it is still necessary
to conduct new research in this eld using a holistic approach. For that
reason, the current study, based on Bronfenbrenner´s Ecological The-
ory (1979) , aimed to explore a model focused on dierent factors lon-
gitudinally linked with adolescent substance use, including individual,
school, family and neighbourghood domains.
According to our rst hypothesis, we expected boys, older students
and participants with low academic performance to score high on sub-
stance use. Although soft alcohol use (beer and wine) was more preva-
lent among boys at wave 1, girls were more likely to report intox-
ication and strong alcohol and tobacco use at wave 2. This is con-
trary to previous research where boys reported higher level of sub-
stance use ( Halladay et al., 2020 ; Lee et al., 2021 ), although Moreno
et al. (2016) found that licit substance use was more prevalent among
female students. Yet, our ndings are in agreement with the current
trend to reduce sex dierences in substance use ( Kraus et al., 2018 ),
related to an increase in social choices made by females ( Rahav et al.,
2006 ).
In line with Zych et al. (2020) , intoxication and alcohol use are more
likely as age increases. A plausible explanation is that there is a crucial
developmental change from Grade 1 to Grade 2, in which the vast major-
ity of students acquire the proper characteristics of adolescence. Congru-
ent with previous research ( Meda et al., 2017 ; Heradstveit et al., 2017 ),
low academic performance predicted more intoxication and tobacco and
alcohol use, both cross-sectionally and one year later. In addition, school
exclusions were cross-sectionally related to more intoxication and licit
substance. Thus, adolescents with poor school performance should be a
target population in drug use prevention strategies.
Secondly, we hypothesised that substance use would be more preva-
lent among students with lower scores in liking school, bonding with
teachers and bonding with peers. Consistent with previous literature
( Daily et al., 2020 ; King et al., 2020 ), liking school and bonding with
teachers have been identied as protective factors against substances
use, especially licit substances, which are the most commonly used sub-
stances at this age. Perhaps the positive feelings towards teachers as
attachment gures or the perception of schools as useful prevent ado-
lescents from getting involved in antisocial behaviours including alcohol
and tobacco use. Nonetheless, there was no evidence of a relationship
between bonding with teachers and substance use one year later. Teach-
ers change from one school year to another and students can establish
4
J. Rodríguez-Ruiz, I. Zych, V.J. Llorent et al. International Journal of Drug Policy 112 (2023) 103946
dierent types of relationships with dierent teachers. In general, our
ndings support the idea of promoting desirable student-teacher rela-
tionships as an eective component of substance use prevention pro-
grammes ( Wenzel et al., 2009 ).
According to the previous studies, higher levels of school friendship
and peer support were found to be related to lower levels of substance
use ( Forster et al., 2015 ; Rodzlan et al., 2021 ), but there is a gap in
knowledge regarding the relation between substance use and bonding
with classmates. Our results showed that participants who scored higher
in bonding with classmates were less likely to report intoxication and
illicit substance use. This is congruent with previous ndings that linked
substance use and unhealthy relationships with classmates that involved
problems such as bullying ( Gaete el al., 2017 ; Pengpid & Peltzer, 2019 )
or cyberbullying ( Choi et al., 2019 ; Sharp et al., 2019 ). However, as hap-
pened with teachers, bonding with classmates did not have a prospective
impact on substance use, which could be caused by possible changes in
the relationships among classmates over time. These results support So-
cial Bond Theory ( Hirschi, 1969 ), given that liking school can be related
to more involvement and bonding with teachers and classmates can in-
crease attachment. It entails a protective factor against substance use as
it decreases the likelihood of being exposed to delinquent behaviour.
Based on scientic literature ( Gerra et al., 2020 ; Petruzelka et al.,
2020 ), according to our third hypothesis, we expected substance use to
be higher among students living in families with lower socio-economic
status. Nevertheless, in our sample, participants who reported high fam-
ily socio-economic status tended to use more alcohol cross-sectionally
and reported more intoxication, tobacco and cannabis use one year later.
The availability of money to purchase alcohol and other substances
could explain the relation between high family socio-economic status
and substance use. The discrepancy between our ndings and previ-
ous research could be due to dierences in the studied substances or in
the target population. Gerra et al. (2020) only studied the use of illicit
drugs, while our results link high family socio-economic status with a
wide range of substances and intoxication. Pretruzelka et al. (2020) fo-
cused on a sample from a structurally disadvantaged region, whereas
our sample involves general population.
In relation to the fourth hypothesis and in line with scientic liter-
ature ( Lee et al., 2018 ; Tucker et al., 2013 ), low neighbourhood socio-
economic status was a predictor of soft alcohol and illicit substance use,
as well as intoxication at wave 1. Some studies found that low income in
some neighbourhoods is related to more illegal activities including drug
dealing ( Chang et al., 2016 ). In consequence, there is a broader avail-
ability and easier access to substances in these contexts. Moreover, the
lack of leisure activities has been identied as a risk factor to drug use
( Levy, 2008 ), which can be more common in disadvantaged neighbour-
hoods. Drug prevention programmes in these contexts should include
components such as fostering employment or increasing the range of
leisure activities as possible elements of success in prevention strate-
gies. Nevertheless, the eect of neighbourhood socio-economic status
on substance use was nonsignicant at the 1-year follow-up. The im-
pact of the neighbourhood on adolescence could not be as stable as the
impact of families. Family is a closer context and its impact more stable
over time, while adolescents can do dierent activities or establish di-
verse relationships outside their neighbourhoods as they grow up. From
a socio-ecological perspective, it supports the idea that closer contexts
have a stronger and more stable inuence on people´s behaviour. Ac-
cording to our outcomes, individual factors (academic performance or
school exclusions) as well as family SES (close context) remain stable or
even become more important over time. However, the impact of distant
domains, like neighbourhood SES or bonding with teachers, disappeared
from wave 1 to wave 2.
The biggest strength of the current study is that it provides a wide
range of ecological factors associated with the use of dierent substances
in early adolescence. Given that we used a longitudinal design, chrono-
logical relations among variables can be established. Also, a large and
diverse sample was studied. Therefore, the results are probably gener-
alizable to the population and similar contexts. However, some limita-
tions should be taken into account. First, although signicant associa-
tions were found between substance use and many of the factors, the
eect size was weak in some cases. Second, socio-economic status was
measured considering the subjective perception of participants, simi-
larly to previous studies ( Kim & Han, 2020 ). These results come from a
Spanish sample, so they may not be generalisable to other cultures or
nationalities. Future research could study these variables further from
a longitudinal perspective including later adolescence or even adult-
hood in order to obtain a more complex and comprehensive perspec-
tive. Other school variables could also be related to substance use, as
well as new forms of online problem behaviours such as buying drugs
online ( Oksanen et al., 2021 ). Problem behaviours tend to form pat-
terns ( Nasaescu et al., 2020 ), and they should be studied from a holistic
perspective. In addition, studies considering each single illicit substance
independently from an ecological perspective could be useful in the fu-
ture.
Even with some limitations, the current study has implications for
policy and practice. Substance use prevention programmes should in-
clude families, given that they are the most consistent context impact-
ing adolescent substance use over time. Our study found evidence on
the importance of school context including bonding to teachers and
classmates as protective factors against substance use. Previous studies
found that school climate policy documents are not always well designed
( Llorent et al., 2021 ). Based on our ndings, educational administrations
and schools should improve the promotion of a positive school climate,
in which students perceive the utility of school and the support of teach-
ers, also promoting bonding to classmates. Moreover, these results have
important implications for research, as they showed a dierential im-
pact of neighbourhood and family socio-economic status on substance
use. Future studies could further explore these associations using more
accurate measures for socio-economic status, such as family income or
unemployment rates in the neighbourhood.
Declarations of interest
None.
CRediT authorship contribution statement
Joaquín Rodríguez-Ruiz: Writing – original draft, Conceptualiza-
tion, Methodology, Investigation, Formal analysis, Writing –review
& editing, Visualization. Izabela Zych: Methodology, Investigation,
Formal analysis, Writing –review & editing, Project administration,
Funding acquisition, Visualization, Resources. Vicente J. Llorent: Con-
ceptualization, Supervision, Writing –review & editing, Project ad-
ministration, Resources. Inmaculada Marín-López: Conceptualization,
Methodology, Investigation, Writing –review & editing, Visualization.
Raquel Espejo-Siles: Methodology, Writing –review & editing, Visu-
alization. Elena Nasaescu: Methodology, Writing –review & editing,
Visualization.
Acknowledgments
The current study was funded by a research grant for the project
“School bullying as a determinant of substance use: a longitudinal study
of risk and protective factors ”, granted by the Spanish Ministry of
Health, Consumer Aairs and Social Welfare within the National Plan
against Drugs 2019 (reference 2019/016). The rst author received a
grant for the university faculty training from the Spanish Ministry of
Science, Innovation and Universities (FPU19/02907).
References
Arria, A. M., Allen, H. K., Caldeira, K. M., Vincent, K. B., & O’Grady, K. E
(2020). Excessive drinking and drug use during college: Prospective associations
with graduate school plans and attendance. Journal of American College Health .
10.1080/07448481.2018.1535494 .
5
J. Rodríguez-Ruiz, I. Zych, V.J. Llorent et al. International Journal of Drug Policy 112 (2023) 103946
Bishop, A. S., Fleming, C. M., & Nurius, P. S. (2020). Substance use proles among gang-
involved youth: Social ecology implications for service approaches. Children and Youth
Services Review, 119 , Article 105600.
10.1016/j.childyouth.2020.105600 .
Brière, F. N., Fallu, J.-S., Morizot, J., & Janosz, M. (2014). Adolescent illicit
drug use and subsequent academic and psychosocial adjustment: An exam-
ination of socially-mediated pathways. Drug Alcohol Dependence, 135 , 45–51.
10.1016/j.drugalcdep.2013.10.029 .
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and
design . Harvard University Press
.
Chang, S. W., Coulson, N. E., & Wang, P. (2016). Optimal drug policy in low-income neigh-
borhoods. Journal of public economic theory, 18 (5), 726–751.
10.1111/jpet.12147 .
Choi, K. S., Earl, K., Lee, J. R., & Cho, S. (2019). Diagnosis of cyber and non-physical
bullying victimization: A lifestyles and routine activities theory approach to con-
structing eective preventative measures. Computers in Human Behavior, 92 , 11–19.
10.1016/j.chb.2018.10.014 .
Connell, C. M., Gilreath, T. D., Aklin, W. M., & Brex, R. A. (2010). Social-
ecological inuences on patterns of substance use among non-metropolitan high
school students. American Journal of Community Psychology, 45 (1-2), 36–48.
10.1007/s10464-009-9289-x .
Daily, S. M., Mann, M. J., Lilly, C. L., Bias, T. K., Smith, M. L., & Kristjansson, A. L. (2020).
School climate as a universal intervention to prevent substance use initiation in early
adolescence: A longitudinal study. Health Education and Behavior, 47 (3), 402–411.
10.1177/1090198120914250 .
Forster, M., Grigsby, T. J., Bunyan, A., Unger, J. B., & Valente, T. W. (2015). The protective
role of school friendship ties for substance use and aggressive behaviors among middle
school students. Journal of School Health, 85 (2), 82–89.
10.1111/josh.12230 .
Foster, K., & Spencer, D. (2013). It’s just a social thing ”: Drug use, friendship and border-
work among marginalized young people. International Journal of Drug Policy, 24 (3),
223–230.
10.1016/j.drugpo.2012.12.005 .
Gaete, J., Tornero, B., Valenzuela, D., Rojas-Barahona, C. A., Salmivalli, C., Valen-
zuela, E., & Araya, R. (2017). Substance use among adolescents involved in bullying:
A cross-sectional multilevel study. Frontiers in Psychology, 8 (JUN), 1–14.
10.3389/fp-
syg.2017.01056
.
Gerra, G., Benedetti, E., Resce, G., Potente, R., Cutilli, A., & Molinaro, S. (2020). Socio-
economic status, parental education, school connectedness and individual socio-
cultural resources in vulnerability for drug use among students. International Journal
of Environmental Research and Public Health, 17 , 1–17.
10.3390/ijerph17041306 .
Hall, W. D., Patton, G., Stockings, E., Weier, M., Lynskey, M., Morley, K. I., & Degen-
hardt, L. (2016). Why young people’s substance use matters for global health. The
Lancet. Psychiatry, 3 (3), 265–279.
10.1016/S2215-0366(16)00013-4 .
Halladay, J., Woock, R., El-Khechen, H., Munn, C., MacKillop, J., Amlung, M., Ogrod-
nik, M., Favotto, L., Aryal, K., Noori, A., Kien, M., & Georgiades, K. (2020). Patterns
of substance use among adolescents: A systematic review. Drug and Alcohol Depen-
dence, 216 , Article 108222.
10.1016/j.drugalcdep.2020.108222 .
Heradstveit, O., Skogen, J. C., Hetland, J., & Hysing, M. (2017). Alcohol and illicit drug
use are important factors for school-related problems among adolescents. Frontiers in
Psychology, 8 , 1–11.
10.3389/fpsyg.2017.01023 .
Hirschi, T. (1969). Causes of delinquency . University of California Press .
Kim, H. R., & Han, M. A. (2020). Associations between problematic smartphone use, un-
healthy behaviors, and mental health status in Korean Adolescents: Based on data from
the 13th Korea Youth Risk Behavior Survey (2017). Psychiatry Investigation, 17 (12),
1216–1225.
10.30773/pi.2020.0007 .
King, K. A., Fuqua, S. H., Vidourek, R. A., Merianos, A. L., & Yockey, R. A. (2020). Does
marijuana use among African American adolescent males dier based on school fac-
tors? Journal of Ethnicity in Substance Abuse , 1–11.
10.1080/15332640.2020.1824840 .
Kipping, R. R., Smith, M., Heron, J., Hickman, M., & Campbell, R. (2015). Multiple risk
behaviour in adolescence and socio-economic status: Findings from a UK birth cohort.
European Journal of Public Health, 25 , 44–49.
10.1093/eurpub/cku078 .
Kraus, L., Seitz, N. N., Piontek, D., Molinaro, S., Siciliano, V., Guttormsson, U., Arpa, S.,
Monshouwer, K., Leifman, H., Vicente, J., Griths, P., Clancy, L., Feijão, F., Flo-
rescu, S., Lambrecht, P., Nociar, A., Raitasalo, K., Spilka, S., Vyshinskiy, K., & Hi-
bell, B. (2018). ‘Are the times A-Changin’? Trends in adolescent substance use in
Europe. Addiction, 113 (7), 1317–1332.
10.1111/add.14201 .
Lee, J. O., Jones, T. M., Kosterman, R., Cambron, C., Rhew, I. C., Herrenkohl, T. I., &
Hill, K. G. (2018). Childhood neighborhood context and adult substance use problems:
The role of socio-economic status at the age of 30 years. Public Health, 165 , 58–66.
10.1016/j.puhe.2018.09.011 .
Lee, Y. T., Huang, Y. H., Tsai, F. J., Liu, H. C., Sun, F. J., Tsai, Y. J., & Liu, S. I. (2021).
Prevalence and psychosocial risk factors associated with current cigarette smoking
and hazardous alcohol drinking among adolescents in Taiwan. Journal of the Formosan
Medical Association, 120 , 265–274.
10.1016/j.jfma.2020.05.003 .
Levy, M. S. (2008). Listening to our clients: The prevention of relapse. Journal of Psychoac-
tive Drugs, 40 , 167–172.
10.1080/02791072.2008.10400627 .
Llorent, V. J., Farrington, D. P., & Zych, I. (2021). School climate policy and its relations
with social and emotional competencies, bullying and cyberbullying in secondary ed-
ucation. Revista de Psicodidáctica (English ed.), 26 (1), 35–44
.
Loeber, R., Stouthamer-Loeber, M., Van Kammen, W., & Farrington, D. P. (1989). De-
velopment of a new measure of self-reported antisocial behavior for young children:
Prevalence a reliability. In M. W. Klein (Ed.), Cross national research and self-reported
crime and delinquency (pp. 203–225). Springer
.
Lorenzo-Seva, U., & Ferrando, P. J. (2006). FACTOR: A computer program to t the ex-
ploratory factor analysis model. Behavior Research Methods, 38 (1), 88–91
.
Meda, S. A., Gueorguieva, R. V., Pittman, B., Rosen, R. R., Aslanzadeh, F., Tennen, H.,
Leen, S., Hawkins, K., Raskin, S., Wood, R. M., Austad, C. S., Dager, A., Fallahi, C.,
& Pearlson, G. D. (2017). Longitudinal inuence of alcohol and marijuana use on
academic performance in college students. PLoS ONE, 12 (3), 1–16. 10.1371/jour-
nal.pone.0172213
.
Ministerio de, Sanidad (2021). Encuesta sobre uso de drogas en enseñanzas secundarias en
España (ESTUDES), 1994-2021 . Gobierno de España, Ministerio de Sanidad
.
Moreno, C., Ramos, P., Rivera, F., Jiménez-Iglesias, A., García-Moya, I., Sánchez- Queija,
I., Moreno-Maldonado, C., Paniagua, C., Villafuerte-Díaz, A., & Morgan, A. (2016).
Informe técnico de los resultados obtenidos por el Estudio Health Behaviour in School-
aged Children (HBSC) 2014 en España. Ministerio de Sanidad, Servicios Sociales e
Igualdad.
Moreno, C., Ramos, P., Rivera, F., Sánchez-Queija, I., Jiménez-Iglesias, A., García-Moya, I.,
Moreno-Maldonado, C., Paniagua, C., Villafuerte-Díaz, A., Ciria-Barreiro, E., Mor-
gan, A., & Leal-López, E. (2020). La adolescencia en España: salud, bienestar, familia,
vida académica y social. Resultados del Estudio HBSC 2018 . Ministerio de Sanidad
.
Mulla, M. M., Bogen, K. W., & Orchowski, L. M. (2020). The mediating role of school con-
nectedness in the associations between dating and sexual violence victimization and
substance use among high school students. Preventive Medicine, 139 , Article 106197.
10.1016/j.ypmed.2020.106197 .
Nasaescu, E., Zych, I., Ortega-Ruiz, R., Farrington, D. P., & Llorent, V. J. (2020). Longi-
tudinal patterns of antisocial behaviors in early adolescence: A latent class and latent
transition analysis. The European Journal of Psychology Applied to Legal Context, 12 (2)
000–000.
10.5093/ejpalc2020a10 .
Oksanen, A., Miller, B. L., Savolainen, I., Sirola, A., Demant, J., Kaakinen, M., &
Zych, I. (2021). Social media and access to drugs online: A nationwide study in the
United States and Spain among adolescents and young adults. The European Journal
of Psychology Applied to Legal Context, 13 (1), 29–36.
10.5093/ejpalc2021a5 .
Pengpid, S., & Peltzer, K. (2019). Bullying victimization and externalizing and
internalizing symptoms among in-school adolescents from ve ASEAN coun-
tries. Children and Youth Services Review, 106 (August), Article 104473.
10.1016/j.childyouth.2019.104473 .
Petruzelka, B., Vacek, J., Gavurova, B., Kubak, M., Gabrhelik, R., Rogalewicz, V., & Bar-
tak, M. (2020). Interaction of socio-economic status with risky internet use, gambling
and substance use in adolescents from a structurally disadvantaged region in central
Europe. International Journal of Environmental Research and Public Health, 17 , 1–18.
10.3390/ijerph17134803 .
Rahav, G., Wilsnack, R. W., Bloomeld, K., Gmel, G., & Kuntsche, S. (2006). The inu-
ence of societal level factors on men´s and women´s alcohol consumption and alcohol
problems. Alcohol and Alcoholism, 41 , 47–55.
10.1093/alcalc/agl075 .
Ribeaud, D., & Eisner, M. (2010). Risk factors for aggression in pre-adolescence: Risk do-
mains, cumulative risk and gender dierences-results from a prospective longitudinal
study in a multi-ethnic urban sample. European Journal of Criminology, 7 , 460–498.
10.1177/1477370810378116 .
Rodríguez-Ruiz, J., Zych, I., Llorent, V. J., & Marín-López, I. (2021). A longitudinal study
of preadolescent and adolescent substance use: Within-individual patterns and protec-
tive factors. International Journal of Clinical and Health Psychology, 21 , Article 100251.
10.1016/j.ijchp.2021.100251 .
Rodzlan, W. S., Saminathan, T. A., Ab Majid, N. L., Miaw, J. L., Mat, H., Abd, H. A.,
Robert, T. G., Ahmad, A., Lsmail, H., Rashid, R. A., & Mohd, M. F (2021). Poly-
substance use among adolescents in Malaysia: Findings from the National Health
and Morbidity Survey 2017. PLOS One, 16 (1), Article 0245593. 10.1371/jour-
nal.pone.0245593
.
Sharp, A., Young, M. S., & Moore, K. A. (2019). Relationship between substance use,
bullying, and other delinquent behaviors among high school students: A secondary
analysis of the Florida Youth Substance Abuse Survey. Journal of Behavioral Health
Services and Research, 46 (4), 570–585.
10.1007/s11414-019-09649-5 .
Shih, R. A., Parast, L., Pedersen, E. R., Troxel, W. M., Tucker, J. S., Miles, J.,
Kraus, L., & D’Amico, E. J (2017). Individual, peer, and family factor modication
of neighborhood-level eects on adolescent alcohol, cigarette, e-cigarette, and mar-
ijuana use. Drug and alcohol dependence, 180 , 76–85.
10.1016/j.drugalcdep.2017.
07.014
.
Staton, M., Dickson, M. F., Tillson, M., Leukefeld, C., Webster, J. M., Knudsen, H. K., &
Oser, C. B. (2020). A preliminary examination of substance use risk among metropoli-
tan and non-metropolitan girls involved in the juvenile justice system. Journal of Child
& Adolescent Substance Abuse, 29 (1), 46–57. 10.1080/1067828x.2020.1827471 .
Syed, S., Sampasa-Kanyinga, H., Hamilton, H. A., Kingsbury, M., & Colman, I. (2021).
Low school belongingness and non-prescription opioid use among stu-
dents in Ontario, Canada. Canadian Journal of Public Health, 112 , 449–455.
10.17269/s41997-020-00460-w .
Tucker, J. S., Pollard, M. S., de la Haye, K., Kennedy, D. P., & Green, H. D. (2013). Neigh-
borhood characteristics and the initiation of marijuana use and binge drinking. Drug
and Alcohol Dependence, 128 , 83–89.
10.1016/j.drugalcdep.2012.08.006 .
Wenzel, V., Weichold, K., & Silbereisen, R. K. (2009). The life skills program IPSY: Pos-
itive inuences on school bonding and prevention of substance misuse. Journal of
Adolescence, 32 (6), 1391–1401.
10.1016/j.adolescence.2009.05.008 .
Zych, I., Farrington, D. P., Llorent, V. J., Ribeaud, D., & Eisner, M. P. (2021). Childhood risk
and protective factors as predictors of adolescent bullying roles. International Journal
of Bullying Prevention, 3 , 138–146.
10.1007/s42380-020-00068-1 .
Zych, I., Rodríguez-Ruiz, J., Marín-López, I., & Llorent, V. J. (2020). Longitudinal stability
and change in adolescent substance use: A latent transition analysis. Children and
Youth Services Review, 112 , 1–7.
10.1016/j.childyouth.2020.104933 .
6
