Skip to main content
. 2015 Aug 10;112(34):E4661–E4670. doi: 10.1073/pnas.1422009112

Table 4.

Impacts of sticker assignments

Experimental arms Claims Maximum speed (km/h) Average speed (km/h) Reckless event Passenger voice
A. Main effects
 Placebo −0.014 0.908 0.146 0.027 0.049
(0.010) (0.745) (0.792) (0.033) (0.035)
 Any treatment sticker −0.017** −0.979 −1.019** 0.020 0.005
(0.008* (0.606) (0.506) (0.024) (0.027)
 Treatment = placebo, P value 0.29 0.00 0.13 0.80 0.11
B. Individual vs. collective action
 Individual −0.012 −1.257* −0.853 0.024 −0.003
(0.008) (0.649) (0.629) (0.026) (0.029)
 Collective action −0.022*** −0.687 −1.194** 0.017 0.013
(0.008) (0.637) (0.484) (0.027) (0.030)
C. Message types
 Text message only 0.001 −0.692 −1.404*** 0.022 −0.012
(0.010) (0.707) (0.532) (0.032) (0.033)
 Supportive: Voice −0.024** −0.039 −0.984* 0.036 0.018
(0.010) (0.685) (0.509) (0.030) (0.033)
 Event aversion: Crash −0.028*** −2.039*** −0.966 0.012 0.001
(0.010) (0.717) (0.811) (0.030) (0.034)
 Consequence aversion: Injury −0.016* −1.210* −0.755 0.010 0.010
(0.010) (0.734) (0.824) (0.031) (0.034)
 Observations 23,474 4,405 4,405 1,471 1,471

Clustered SEs in parentheses. Significance at *10%, **5%, and ***1% levels.

In the claims regression, entries are coefficients on the reported independent variable interacted with postrecruitment in a difference in difference specification. In the speed and survey regressions, entries reflect simple differences.