Table 3

Multivariate analysis of associations between establishment characteristics and frequency of attempted straw purchases* and undocumented purchases* in the past year and firearm theft in the past 5 years

Straw purchaseUndocumented purchaseTheft
CharacteristicOR*95% CIp ValueOR*95% CIp ValueOR95% CIp Value
Licensee type0.0007<0.00010.001
 Pawnbroker1.911.31 to 2.782.451.68 to 3.562.071.24 to 3.21
 DealerReferentReferentReferent
Sales (n)<0.00010.0090.0006
 ≥50010.075.35 to 18.972.921.57 to 5.424.692.12 to 10.39
 200–4993.111.78 to 5.452.131.20 to 3.772.040.95 to 4.34
 100–1991.771.02 to 3.082.021.13 to 3.592.891.36 to 6.12
 <100ReferentReferentReferent
Sales to women (%)†0.046
 ≥251.190.62 to 2.27
 11–241.250.66 to 2.37
 6–102.251.20 to 4.23
 0–5Referent
Sales to law enforcement (%)†0.099
 10+0.650.38 to 1.13
 5–91.170.66 to 0.06
 ≥1, <50.790.46 to 1.35
 <1Referent
Multiple sales (%)†0.03
 ≥51.821.08 to 3.08
 2–41.711.01 to 2.90
 1–1.91.040.60 to 1.78
 <1Referent
Tactical rifle sales (%)‡0.06
 ≥201.350.80 to 2.28
 6–191.260.75 to 2.12
 2–50.720.45 to 1.17
 0–1Referent
Gun show sales (%)†0.094<0.0001
 ≥251.570.78 to 3.185.943.02 to 11.68
 >0, <250.560.29 to 1.112.151.13 to 4.05
 0ReferentReferent
Denied sales (%)§<0.00010.070.10
 ≥54.202.49 to 7.091.861.07 to 3.211.530.78 to 2.98
 2–43.892.26 to 6.731.971.11 to 3.482.131.07 to 4.25
 11.691.03 to 2.761.831.08 to 3.081.130.58 to 2.20
 <1ReferentReferentReferent
Trace requests (%)¶0.00010.0960.001
 ≥22.301.38 to 3.811.721.02 to 2.903.341.69 to 6.60
 >0.5, <22.641.56 to 4.451.530.89 to 2.621.550.76 to 3.17
 >0, ≤0.51.220.68 to 2.201.060.57 to 1.941.380.63 to 3.03
 0ReferentReferentReferent
  • *Results obtained by ordinal regression. ORs represent the effect of a given characteristic on the odds of reporting attempted straw purchases or undocumented purchases with a greater frequency than was reported by the referent group.

  • †Percentage of overall firearm sales in 2010.

  • ‡Percentage of rifle sales in 2010.

  • §Percentage of sales that were denied after a background check, average over 5 years.

  • ¶Annual number of trace requests, average over 5 years, as percentage of overall firearm sales in 2010.

  • The following variables were deleted from the table because they were not retained in any model: handgun sales, inexpensive handgun sales, internet sales.