32
International Journal of Security and Terrorism • Volume: 4 (1)
3.2.2.1. Logit Coefficients of Command Model
According to the results shown in Table 3, all of the factor variables are positively associated
with the dependent variable of the first (i.e., command) model.
Table 3. Logit Coefficients of Command Model with Dependent Variable Command-Level
Managers’ Use of Crime Analysis
Variable
B
SE
z
P > | z |
Crime rates
-0.002
0.004
-0.57
0.571
Agency size
0.275
0.180
1.52
0.128
Unions
-0.200
0.249
-0.80
0.421
Budget
-7.10e-07
1.39e-06
-0.51
0.610
Hierarchy
0.166
0.190
0.87
0.383
Crime analysis unit
0.636
0.275
2.31
0.021
Statistical
analysis
0.417
0.121
3.43
0.001
Crime analysis
0.797
0.133
5.99
0.000
Intelligence analysis
0.363
0.127
2.86
0.004
Survey analysis
0.426
0.117
3.65
0.000
Patrol strategy analysis
0.641
0.124
5.16
0.000
Displacement analysis
0.177
0.121
1.47
0.142
τ
1
-2.217
0.456
τ
2
1.503
0.448
Note. N = 352. Approximate likelihood-ratio test of parallel regression assumption:
χ2 (12 df) = 15.67, p = .2067.
The only significant control variable is crime analysis unit, which is positively associated
with the dependent variable.
3.2.2.2. Logit Coefficients of Patrol Model
According to the results shown in Table 4, all of the factor variables—except survey analysis
and displacement/diffusion analysis—are positively associated with the dependent variable
of the second model (i.e., patrol model). Among the control variables, crime analysis unit
is the positively associated with the dependent variable, while size is negatively associated
with the dependent variable.
33
Intelligence - Led Policing:
How the Use Of Crime Intelligence Analysis Translates in to the Decision-Making
Table 4. Logit Coefficients of Patrol Model with Dependant Variable Patrol Officers’ Use of
Crime Analysis
Variable
B
SE
z
P > | z |
Crime rates
0.002
0.004
0.46
0.644
Agency size
-0.520
0.173
-3.01
0.003
Unions
0.396
0.248
1.60
0.111
Budget
1.33e-06
1.34e-06
0.99
0.321
Hierarchy
0.111
0.189
0.59
0.557
Crime analysis unit
0.629
0.275
2.29
0.022
Statistical analysis
0.315
0.119
2.65
0.008
Crime analysis
1.060
0.139
7.62
0.000
Intelligence analysis
0.357
0.126
2.82
0.005
Survey analysis
0.059
0.109
0.54
0.592
Patrol strategy analysis
0.273
0.117
2.33
0.020
Displacement analysis
0.079
0.120
0.66
0.508
τ
1
-1.990
0.438
τ
2
1.880
0.437
Note. N = 352. Approximate likelihood-ratio test of parallel regression assumption:
χ2 (12 df) = 7.58, p = .8169.
3.2.2.3. Logit Coefficients of Detective Model
Table 5 shows that three of the factor variables (i.e., statistical analysis, crime analysis, and
intelligence analysis) are positively associated with the dependent variable (i.e., detectives’
use of crime analysis). As in the previous models (i.e., command model and patrol model),
crime analysis unit is positively significant when controlled.
34
International Journal of Security and Terrorism • Volume: 4 (1)
Table 5. Logit Coefficients of Detective Model with Dependent Variable Detectives’ Use of
Crime Analysis
Variable
B
SE
z
P > | z |
Crime rates
-0.003
0.004
-0.65
0.516
Agency size
-0.071
0.175
-0.41
0.683
Unions
0.425
0.252
1.69
0.092
Budget
-2.29e-08
1.36e-06
-0.02
0.987
Hierarchy
0.095
0.020
-0.50
0.618
Crime analysis unit
0.610
0.276
2.21
0.027
Statistical analysis
0.460
0.125
3.68
0.000
Crime analysis
1.184
0.147
8.05
0.000
Intelligence analysis
0.499
0.129
3.85
0.000
Survey analysis
0.211
0.114
1.84
0.065
Patrol strategy analysis
0.210
0.121
1.72
0.085
Displacement analysis
0.0425
0.122
0.35
0.727
τ
1
-2.590
0.460
τ
2
1.350
0.441
Note. N = 352. Approximate likelihood-ratio test of parallel regression assumption:
χ2 (12 df) = 9.50, p = .6598.
4. Discussion and Conclusion
The current study focused on the association of crime analysis functions with the
organizational decision-making process at three ranks: command-level managers,
patrol officers, and detectives. In addition, some internal (organizational) and external
(environmental) determinants are controlled on the organizational decision-making process.
Doing so allowed the researchers to partially test a new policing model, intelligence-led
policing or, more specifically, Ratcliffe’s (2008) 3-i model. As mentioned previously, there
are three components (crime analysis, decision-making, and criminal environment) and
three processes in the 3-i model that all begin with the letter i (interpretation, influence, and
impact). In this model, crime intelligence analysts are assumed to interpret data from the
criminal environment and then influence the decision- makers, who are assumed to make
decisions or policies that impact crime and prolific offenders in the environment (i.e., the
criminal environment). The current study was a partial testing of the 3-i model because only
the perceived relationship of crime intelligence analysis and the decision-making components
are explored, whereas the rest of the model is not studied because of insufficient proper data
about criminal environment and criminals in that environment.
As indicated in earlier sections, although the findings in general seem to support the
hypotheses, the findings indicate only an association based on the odds ratios between the
variables—and not causality.
Almost all of the crime analysis types studied were found to be significant, as
expected, with the relevant level of decision-making within the organization. Further, the