Skip to main content

Table 2 Association between age of first cannabis use and later life outcomes

From: Too young for Cannabis? Choice of minimum legal age for legalized non-medical Cannabis in Canada

 (1)(2)(3)(4)
Outcome:Education (CTUMS)Current cigarette smoking (CTUMS+CTADS)Self-reported general health (CTADS)Self-reported mental health (CTADS)
Panel A: AIPW estimates    
Age of first use 18 years0.275***−0.094***0.077**0.044
 (0.026)(0.009)(0.037)(0.034)
 z = 10.5, p = 0.00z = −10.3, p = 0.00z = 2.1, p = 0.04z = 1.3, p = 0.20
Age of first use 19–20 years0.394***−0.121***0.057**0.086***
 (0.020)(0.010)(0.025)(0.023)
 z = 19.9, p = 0.00z = − 11.8, p = 0.00z = 2.3, p = 0.02z = 3.7, p = 0.00
Age of first use 21–24 years0.590***−0.126***0.110***0.068**
 (0.040)(0.011)(0.030)(0.030)
 z = 14.8, p = 0.00z = −12.0, p = 0.00z = 3.6, p = 0.00z = 2.2, p = 0.03
Panel B: MMWS estimates    
Age of first use 18 years0.283***−0.094***0.0700.034
 (0.030)(0.008)(0.041)(0.037)
 t = 9.36, p = 0.00t = 11.47, p = 0.00t = 1.72, p = 0.12t = 0.92, p = 0.38
Age of first use 19–20 years0.414***−0.123***0.067**0.089**
 (0.023)(0.010)(0.023)(0.029)
 t = 17.92, p = 0.00t = 12.46, p = 0.00t = 2.90, p = 0.02t = 3.11, p = 0.01
Age of first use 21–24 years0.630***−0.141***0.093**0.029
 (0.043)(0.009)(0.037)(0.038)
 t = 14.55, p = 0.00t = 15.33, p = 0.00t = 2.53, p = 0.03t = 0.77, p = 0.46
R squared0.0140.0130.0040.002
N35,90442,61065986593
  1. Note: Data are from CTUMS 2004–2012 and CTADS 2013 and 2015 as indicated in table. Panel A shows the AIPW estimates and Panel B shows the MMWS estimates. Estimates are the difference in outcome for the specific age group of first cannabis use relative to the reference category of age of first use < 18 years. Linear regression models were estimated for all outcomes (including ordered outcomes i.e., education, general health and mental health as the Stata command ‘teffects aipw’ for AIPW does not allow use of ordered logit or ordered probit outcome models.) All models include province and year fixed effects, as well as controls for: respondent’s age, household size, sex, place of residence (urban/rural), marital status and use of tobacco products other than cigarettes). Models in column 1 also control for language spoken at home. All models except column 2 control for respondent’s smoking status. Sample includes respondents with current age 21–65 years who initiated cannabis before age 25. Only respondents who used cannabis more than once in lifetime are included. Standard errors (in parentheses) are clustered at province level. Significance levels are: *** p < 0.01, ** p < 0.05, * p < 0.1