Table 3. Robustness check of the effect of the restriction on carrying guns on gun-related mortality, Colombia, 2008–2014.
Time period | No. of observations | Effect of policy, β coefficient (SE) | P |
---|---|---|---|
2011–2013 (gun-related mortality) |
323 |
−0.39 (0.10) |
< 0.01 |
2008–2011 (gun-related mortality) |
432 |
0.28 (0.58) |
0.64 |
2012–2014 (gun-related mortality) |
324 |
−0.13 (0.46) |
0.77 |
2008–2014 (city-specific time trends) | 756 | −2.51 × 10−5 (7.85 × 10−6) | 0.01 |
SE: standard error.
Notes: The outcome is the logarithm of the monthly mortality rate per 100 000 population associated with gun injuries. Gun-related deaths include homicides, suicides and unintentional incidents. All models are a difference-in-differences panel design with city and the month fixed effects for all Colombian cities with a population of more than 500 000 in 2008. All models include calendar month dummies. The β coefficients represent the interaction of a dummy variable accounting for month of the enactment and a dummy variable identifying the treatment city. The coefficients for the periods before and after enactment of the restriction compare the intervention and the control cities. The coefficient for the city-specific time trends model represents an interaction of a dummy variable accounting for the month of the enactment, a dummy variable identifying the treatment city and the time variable accounting for every month in the study period.