Skip to main content
. 2018 Dec 31;13(12):e0208925. doi: 10.1371/journal.pone.0208925

Table 4. Fixed effect regression models for adjusted associations between homicide rates and BFP coverage in the Brazilian municipalities by gender, 2004–2012.

Variable Male analysis model Female analysis model
RR (95% CI) RR (95% CI)
BFP coverage 0–30% 1.000 1.000
BFP coverage >30–70% 0.837 (0.790–0.888) 0.804 (0.723–0.894)
BFP coverage >70% 0.770 (0.726–0.818) 0.730 (0.655–0.813)
% of inhabitants receiving BFP benefit 1.008 (1.007–1.009) 1.013 (1.010–1.016)
Per capita income BR$ (monthly) 0.938 (0.924–0.952) 1.001 (1.000–1.001)
% unemployed people 0.986 (0.981–0.991) 1.021 (1.008–1.034)
Policing rate 1.000 (1.000–1.000) 1.000 (1.000–1.000)
Guns availability 1.000 (1.000–1.001) 1.003 (1.002–1.004)
% of people with low education level 1.040 (1.035–1.044) 0.997 (0.986–1.007)
Urbanization rate 0.995 (0.992–0.997) 1.001 (0.996–1.007)
Time (year) 1.094 (1.085–1.103) 0.998 (0.976–1.020)
Number of observations 46890 33921
Number of municipalities 5210 3769
Number of homicides 338776 35871

Abbreviations: CI = Confidence Interval; RR = Rate Ratio.