Skip to main content
. 2020 Sep 9;16(3):e1112. doi: 10.1002/cl2.1112

Table 8.

Mean percent change by design (RCTs vs. quasi‐experiments) and construct

95% CI
Construct Design k Mean % change Lower Upper Q Between p (Q Between)
Arrests QE 6 −13.5 −24.0 −1.5 3.889 .049
RCT 7 2.4 −7.9 13.8
Complaints against officer QE 7 −10.8 −35.2 22.9 0.261 .609
RCT 15 −19.3 −34.8 0.0
Dispatched calls for service QE 3 9.3 5.4 13.4 12.974 .000
RCT 3 −0.6 −4.2 3.1
Incident reports QE 1 −19.6 −24.5 −14.4 46.705 .000
RCT 2 0.0 −0.2 0.1
Officer‐initiated calls for service QE 3 34.1 13.3 58.8 11.577 .001
RCT 5 0.1 −0.1 0.2
Assault on officer/officer injuries/resistance QE 3 4.7 −61.5 184.9 0.040 .841
RCT 12 16.3 −5.3 42.8
Response time QE 1 −0.5 −2.4 1.4 0.253 .615
RCT 2 0.3 −2.1 2.7
Field interviews or stop and frisk QE 1 −50.4 −66.2 −27.2 11.679 .001
RCT 3 1.0 −12.0 15.9
Traffic stops or traffic tickets QE 3 −11.3 −45.1 43.2 0.275 .600
RCT 2 10.7 −43.9 118.4
Use of force QE 7 −3.1 −28.5 31.4 0.091 .764
RCT 19 −8.1 −22.8 9.3

Note: k = number of effect sizes; Q Between is the test of the difference between the means with associated significance value p(Q Between).

Abbreviations: QE, quasi‐experimental study; RCT, randomized controlled trial.