Skip to main content
. 2023 Dec 21;14:1306439. doi: 10.3389/fpsyt.2023.1306439

Table 3.

Aim and major findings related to peer influence and selection effects on adolescent’s cannabis use.

Data Source Authors (year) Objective Synthesis of results: Major findings related to the social network analyses
Add Health Wang et al. (2018) (37) To examine the co-evolution of adolescent friendship network ties and whether there was interdependence in usage of cigarettes, alcohol, and cannabis Peer influence effects on cannabis use:
School 1:

  • In-degree (own popularity) (β = 0.03, p > 0.05)

  • Cannabis use peer influence (β = 1.43, p < 0.01)

  • Number of friends who smoked (β = 0.02, p > 0.05)

  • Number of friends who drank (β = −0.04, p > 0.05)

School 2:

  • In-degree (own popularity) (β = 0.02, p > 0.05)

  • Cannabis use peer influence (β = 1.32, p < 0.001)

  • Number of friends who smoked (β = 0.03, p > 0.05)

  • Number of friends who drank (β = −0.04, p > 0.05)

Peer selection effects on cannabis use:
School 1:

  • Similarity cannabis use:β = 0.27, p < 0.001.

School 2:

  • Similarity cannabis use on school 2:β = 0.22, p < 0.01.
Schaefer (2018) (38) The aim was in the systematic network selection processes that lead adolescents into friendships with substance-using peers Peer influence effects on cannabis use:

  • Friend’s cannabis use (average similarity): M1(β = 2.122, p < 0.001); M2(β = 2.029, p < 0.01); M3(β = 1.887, p < 0.05); M4(β = 2.338,p < 0.001); M5(β = 2.184, p < 0.001)

Peer selection effects on cannabis use:

  • Cannabis ego: M1 (β = −0.24, p > 0.05); M2 (β = −0.48, p > 0.05); M3 (β = −0.048, p > 0.05); M4 (β = −0.029, p > 0.05); M5 (β = −0.107, p > 0.05).

  • Cannabis alter: M1 (β = 0.20, p > 0.05); M2 (β = 0.163, p > 0.05); M3 (β = 0.207, p > 0.05); M4 (β = 0.143, p > 0.05); M5 (β = 0.12, p > 0.05).

  • Cannabis similarity: M1 (β = 0.395, p < 0.01); M2 (β = 0.268, p < 0.05); M3 (β = 0.321, p < 0.05); M4 (β = 0.395, p < 0.001); M5 (β = 0.267, p > 0.05).

  • Ego’s cannabis use x R alter: M1 (β = −0.108, p > 0.05); M2 (β = −0.512, p < 0.05); M3 (β = −0.132, p > 0.05); M4 (β = −0.337, p < 0.1); M5 (β = −0.291, p > 0.05).

  • R ego x alter’s cannabis use: M1 (β = 0.199, p > 0.05); M2 (β = −0.269, p > 0.05); M3 (β = 0.046, p > 0.05); M4 (β = 0.384, p < 0.1); M5 (β = −0.567, p < 0.1).
De la Haye et al.
(2015) (39)		 The current study tests whether the observed tendency for adolescents to select friends with similar histories of marijuana use (42) is explained by friends’ selection on other risk factors associated with substance use Peer influence effects on cannabis use:
School 1:

  • Friends’ lifetime cannabis use: M1: (PE = 0.52, p < 0.01); M2: (PE = 0.52, p < 0.01); M3: (PE = 0.50, p < 0.01)

School 2:

  • Friends’ lifetime cannabis use: M1: (PE = 0.24, p > 0.05); M2: (PE = 0.24, p > 0.05); M3: (PE = 0.10, p > 0.05)

Peer selection effects on cannabis use:
School 1:

  • Any history of cannabis use ego: M1 (PE = –0.16, p > 0.05); M2 (PE = –0.05, p > 0.05); M3 (PE = 0.00, p > 0.05)

  • Any history of cannabis use alter: M1 (PE = –0.10, p > 0.05); M2 (PE = –0.10, p > 0.05); M3 (PE = 0.01, p > 0.05)

  • Same history of cannabis use: M1 (PE = 0.27, p < 0.01); M2(PE = 0.20, p < 0.05); M3 (PE = 0.18, p < 0.05)

  • Current cannabis use ego: M2 (PE = –0.12, p > 0.05); M3 (PE = –0.08, p > 0.05)

  • Current cannabis use alter: M2 (PE = 0.06, p > 0.05); M3 (PE = 0.11, p > 0.05)

  • Same current cannabis use: M2 (PE = 0.17, p < 0.01); M3 (PE = 0.16, p < 0.01)

School 2:

  • Any history of cannabis use ego: M1 (PE = –0.27, p < 0.01); M2 (PE = –0.23, p < 0.05); M3 (PE = –0.23, p > 0.05)

  • Any history of cannabis use alter: M1 (PE = 0.14, p > 0.05); M2 (PE = 0.15, p > 0.05) M3 (PE = 0.12, p > 0.05)

  • Same history of cannabis use: M1 (PE = 0.32, p < 0.01); M2 (PE = 0.33, p < 0.01); M3 (PE = 0.30, p < 0.01)

  • Current cannabis use ego: M2 (PE = –0.06, p > 0.05)

  • Current cannabis use alter: M2 (PE = –0.06, p > 0.05); M3 (PE = 0.12, p > 0.05)

  • Same current cannabis use: M2 (PE = –0.02, p > 0.05)
Vogel et al. (2015) (40) To examine the moderating influence of school connectedness, school drug culture, and global network density on the association between peer network status and cannabis use. Peer influence effects on cannabis use:

  • Own popularity: M1 (β = 0.04, OR = 1.04, CI95% = 1.02, 1.07, p < 0.05), M2 (β = 0.04, OR = 1.04, CI95% = 1.01, 1.07, p < 0.05); M3 (β = 0.04, OR = 1.04, CI95% = 1.02, 1.07, p < 0.05)

  • Peer substance use (alcohol and tobacco): M1 (β = 0.64, OR = 1.90, CI95% = 1.70, 2.13, p < 0.001), M2 (β = 0.64, OR = 1.90, CI95% = 1.70, 2.12, p < 0.001); M3 (β = 0.64, OR = 1.90, CI95% = 1.70, 2.12, p < 0.001)

  • Network centrality: M1 (β = −0.10, OR = 0.90, CI95% = 0.78, 1.04, p > 0.05), M2 (β = −0.09, OR = 0.91, CI95% = 0.79, 1.05, p > 0.05); M3 (β = −0.10, OR = 0.91, CI95% = 0.79, 1.04, p > 0.05)

  • Own popularity x density: M3 (β = <0.00, OR = 0.99, CI95% = 0.97, 1.02, p > 0.05)

  • Own popularity x school connectedness: M3 (β = −0.05, OR = 0.95, CI95% = 0.91, 0.98, p < 0.05)

  • · Own popularity x school drug use (alcohol & tobacco): M3 (β = <0.00, OR = 0.99, CI95% = 0.99, 1.01, p > 0.05)
Tucker et al. (2014) (41) To examine whether structural features of friendships moderate friends’ influence on adolescent cannabis use over time. Peer Influence effects on cannabis use:
School 1:

  • Friends’ cannabis use: M1: not significant*; M2: not significant*; M3: PE = 0.85, p = 0.069.

  • Friends’ cannabis use (influence) × friendship reciprocity M1:PE = 1.14, p = 0.028

  • Friends’ cannabis use (influence) × friend popularity M2: PE = 0.12, p = 0.189

  • Friends’ cannabis use (influence) × popularity difference M3: PE = −0.02, p = 0.500

School 2:

  • Friends’ cannabis use: M1: not significant*; M2: not significant*; M3: PE = 0.53, p = 0.109

  • Friends’ cannabis use (influence) × friendship reciprocity M1: PE = 0.51, p = 0.254

  • Friends’ cannabis use (influence) × friend popularity M2:PE = 0.15, p = 0.041

  • Friends’ cannabis use (influence) × popularity difference M3: PE = 0.01, p = 0.709

Note: reciprocity, friend popularity and popularity difference were not measured individually: they were measured in interactions.
Peer Selection effects on cannabis use:
School 1

  • Ego’s cannabis use M1: PE = −0.01, p = 0.860; M2: PE = −0.01, p = 0.905; M3: PE = −0.02, p = 0.909

  • Alter’s cannabis use M1: PE = −0.19, p = 0.652; M2: PE = −0.27, p = 0.684; M3: PE = −0.26, p = 0.873

  • Squared alter cannabis use M1: PE = 0.14, p = 0.314; M2: PE = 0.16, p = 0.444; M3: PE = 0.16, p = 0.764

  • Similar/same cannabis use M1:PE = 1.53, p = 0.003; M2:PE = 1.49,p = 0.000; M3:PE = 1.49,p = 0.004

School 2

  • Ego’s cannabis use PE = 0.10, p = 0.102; M2: PE = 0.09, p = 0.103; M3: PE = 0.09, p = 0.049

  • Alter’s cannabis usePE = 0.49,p = 0.040; M2: PE = 0.37, p = 0.064; M3: PE = 0.39, p = 0.071

  • Squared alter cannabis use PE = 0.41, p = 0.195; M2: PE = −0.09, p = 0.267; M3: PE = −0.10, p = 0.282

  • Similar/same cannabis usePE = 1.03,p = 0.000; M2:PE = 1.04,p = 0.000; M3:PE = 1.02,p = 0.000
De la Haye et al. (2013) (42) (A) To determine the extent to which friendship networks influence cannabis use (influence effects) and cannabis use influences friendship selection (selection effects). (B) to assess if a multiplicative model of risk explains differences in cannabis-based selection and influence. Peer Influence effects on cannabis use:
On lifetime (M1)
School 1:

  • Friends’ cannabis use lifetime: not significant*

  • Friends’ cannabis use last month:PE = 1.31,p = 0.001

School 2

  • Friends’ cannabis use lifetime: not significant*

  • Friends’ cannabis use last month: PE = 0.61, p = 0.116

On last month (M2)
School 1

  • Friend’s cannabis use: PE = 0.63, p = 0.126

  • School 2

  • Friend’s cannabis use: PE = 0.51, p = 0.125

Peer Selection effects on cannabis use:
On lifetime (M1)
School 1

  • Ego’s cannabis use:PE = −0.15,p = 0.049

  • Alter’s cannabis use: PE = −0.10, p = 0.099

  • Same cannabis use: PE = 0.27,p = 000

School 2

  • Ego’s cannabis use (lifetime):PE = −0.14,p = 0.032

  • Alter’s cannabis use (lifetime): PE = 0.11, p = 0.104

  • Same cannabis use (lifetime):PE = 0.43,p = 0.000

On last month (M2)
School 1

  • Ego’s cannabis use: PE = −0.02, p = 0.822

  • Alter’s cannabis use: PE = −0.25, p = 0.558

  • Squared alter cannabis use: PE = 0.16, p = 0.268

  • Same cannabis use:PE = 1.49,p = 0.000

School 2

  • Ego’s cannabis use: PE = 0.10, p = 0.069

  • Alter’s cannabis use: PE = 0.39, p = 0.238

  • Squared alter cannabis use (past month): PE = −0.09, p = 0.459

  • Same cannabis use:PE = 1.02,p = 000
Osgood et al. (2014) (43) To examine the association of substance use with the types of positions adolescents hold in cohesive peer groups within the friendship networks of their schools’ grade-cohort. Peer Influence effects on cannabis use:

  • Indegree (own popularity): (Coef = 0.013, p > 0.05).

  • Friendship group position (core member as the reference group):

Peripheral member (Coef = 0.112, p > 0.05).
Isolate (Coef = 0.142, p > 0.05).
Liaison (Coef = 0.312, p < 0.01).
Non-member (Coef = 0.155, p > 0.05).

  • Friend’s cannabis use:

Friends’ mean use (Coef = 3.763, p < 0.01).
Friends’ mean use × total simple number of friends (Coef = 2.408, p < 0.01).

  • Out-degree: (Coef = −0.125, p < 0.01).

  • Reach: (Coef = 0.008, p < 0.01).
Context of Adolescent Substance Use Study Ennett et al. (2006) (44) To examine the peer context of adolescent substance use, social network analysis was used to measure three domains of attributes of peer networks: social embeddedness, social status, and social proximity to substance users. Peer influence effects on cannabis use:
Social embeddedness:

  • Reciprocity: ages 11 (OR = 0.55, p > 0.05); ages 13: (OR = 0.75, p > 0.05); ages 15: (OR = 1.01, p > 0.05).

  • Neighbourhood density: ages 11 (OR = 0.49, p > 0.05); ages 13: (OR = 0.24, p > 0.05); ages 15: (OR = 0.12, p < 0.001).

  • Out nominations (out-degree): ages 11 (OR = 1.28, p > 0.05); ages 13: (OR = 1.26, p < 0.05); ages 15: (OR = 1.25, p < 0.01).

Social position (group member is the reference group):
Isolate: ages 11 (OR = 1.11, p > 0.05); ages 13: (OR = 0.96, p > 0.05); ages 15: (OR = 0.83, p > 0.05).
Bridge: ages 11 (OR = 1.00, p > 0.05); ages 13: (OR = 0.98, p > 0.05); ages 15: (OR = 0.96, p > 0.05).
Social Status:

  • Normed indegree: ages 11 (OR = 1.66, p < 0.05); ages 13: (OR = 1.35, p < 0.001); ages 15: (OR = 1.09, p > 0.05)

  • Reach centrality: ages 11 (OR = 1.05, p < 0.05); ages 13: (OR = 1.04,p < 0.001); ages 15: (OR = 1.02, p < 0.05)

  • Betweenness centrality: ages 11 (OR = 1.07, p > 0.05); ages 13: (OR = 1.06, p > 0.05); ages 15: (OR = 1.06, p > 0.05)

  • Bonacich power centrality: ages 11 (OR = 1.00, p > 0.05); ages 13: (OR = 1.01, p > 0.05); ages 15: (OR = 1.03,p < 0.01).

Social proximity to cannabis users:

  • Best friend cannabis user: ages 11 (OR = 23.34, p < 0.001); ages 13: (OR = 7.82, p < 0.001); ages 15: (OR = 2.62, p < 0.01).

  • No. neighbourhood cannabis users: ages 11 (OR = 7.39, p < 0.001); ages 13: (OR = 3.77, p < 0.001); ages 15: (OR = 1.93, p < 0.001)

  • Distance to cannabis user: ages 11 (OR = 0.25, p < 0.001); ages 13: (OR = 0.41,p < 0.001); ages 15: (OR = 0.66, p < 0.001).

* Authors did not provide the exact value. R, risk factors of 4 attributes: family connectedness, school belonging, grade point average (GPA), religiosity and self-control. PE, parameter estimate; OR, odds ratio. See last column from Table 2 for understanding which models are M1, M2, M3 in each article. Significant of bold values.