This paper investigates the relationship between cannabis use and thinking about suicide, or suicidal ideation. Our empirical analysis is based on a 30 year longitudinal study of a cohort of children born in 1977 in Christchurch, New Zealand. In this cohort, suicidal ideation is more prevalent among regular cannabis users than it is among individuals who have never used cannabis regularly. In order to determine the extent to which this correlation represents a causal relationship, we use a bivariate hazard framework to model the uptake of regular cannabis use and the onset of suicidal ideation. In addition to reverse causality, our framework accounts for unobserved common confounders by allowing the unobserved heterogeneity terms in these transitions to be correlated. We find that, after controlling for a rich set of individual and family characteristics, the unobserved heterogeneity in the two transitions is significantly correlated. Once this is accounted for, there remains a significant direct effect of regular cannabis use on suicidal ideation of males.
Jenny Williams (University of Melbourne)
Date & time:
11 Oct 2010 15:00 pm - 11 Oct 2010 16:30 pm
External seminars home