During the C-71 hearing there was some discussion regarding how effective the questionnaire is regarding suicide prevention. New laws were passed in the 1990s that involved asking questions of firearms owner applicants regarding their psychiatric history. According to a Department of Justice report the current questions regarding past suicide history were added to the application form in January 1993. These were phased in slowly and auditing of the implementation showed that new applications for firearms in 1993 and 1994 were much smaller than previous years.
There are multiple factors that contribute to suicide, unemployment. alcohol use, etc. are some of those. New drugs to treat depression were also implemented in the 1990s.
In order to control for other factors that contribute to suicide a Difference in Differences (DID) statistical model was created to use non firearm suicide as a control. A DID model will demonstrate any differences in trends of firearms and non firearms suicide after laws are implemented and whether these trends are different from each other. It is potentially possible that if firearms ownership decreases it will lower firearm suicide as that method will be less available, however it is also possible that people who commit suicide will choose another method making up for the difference and negating any benefit from limiting suicide by firearm. This is called the substitution effect.
An intervention year of 1993 was chosen as indicated above.
As seen and reported in the data below, there was no statistically significant difference in trends after the intervention year. In other words there was no added benefit overall in suicide reduction after implementation of the questionnaire on application forms.
The DID model is as follows:
ln(suicide) = intercept + B1(year) + B2(firearm_non_firearm_suicide) + B3(intervention) + B4(year)( firearm_non_firearm_suicide) + B5(firearm_non_firearm_suicide)(intervention) + B6(year)(intervention) + B7(year)( firearm_non_firearm_suicide)(intervention)
(year) is 1981 to 2015 (firearm_non_firearm_suicide) is a variable to indicate if that instance of suicide was by firearm or non firearm where 0 is suicide by non-firearm and 1 is by firearm (intervention) is a variable to indicate if that instance of suicide occurred before or after the year where the law changed
B6 is the change in trend in non firearm suicide post intervention B7 is the change in trend in firearm suicide post intervention
A significant result is a value less than 0.05 as standard in scientific studies, though a trend towards using more stringent significant value of 0.005 is becoming popular.