Ferguson Effect on self We presented respondents with eight statements regarding the ex-
less enjoyable to have a career in law enforcement,” and “made it more difficult for you to be motivated at work.” A complete list of the items used to measure the effect of negative publicity on respondents is available in Appendix A. Principal components analysis (PCA) with varimax rotation demonstrated that the eight items loaded onto two distinct components – one pertaining to less motivation (λ=4.18, load- ings N0.66) and the other to increased danger on the job (λ=1.06, load- ings N0.60). Each component demonstrated adequate internal consistency (motivation α = 0.87, danger α = 0.71) and, therefore, were combined into separate additive scales. Less motivation ranges from 5 to 25, with higher scores indicating the respondent felt less mo- tivated to do his/her job as a result of negative publicity over the prior 6 months. Increased danger ranges from 3 to 15, with higher scores reflecting a belief on the part of the respondent that law enforcement had become more dangerous as a result of negative publicity following Ferguson and related events. Descriptive statistics for all variables used in the analyses are presented in Table 1.
5.2.2. Ferguson Effect on colleagues It is possible that regardless of whether officers believed negative
publicity stemming from the Ferguson controversy had affected them, they might believe that it had influenced other police officers, including their colleagues. This is an important consideration given that police of- ficers routinely rely on their colleagues for backup. For example, officers may hesitate to stop suspicious persons if they feel their colleagues are reluctant to use force when it may be necessary. To capture this senti- ment, we asked respondents to indicate their agreement (1 = strongly disagree to 5= strongly agree) with statements regarding whether neg- ative publicity surrounding law enforcement in the previous 6 months had: (1) made it more difficult for coworkers to do their job, (2) made it more difficult for coworkers to be motivated at work, (3) caused co- workers to be less proactive on the job than they were in the past, and (4) caused coworkers to be more apprehensive about using force even though it may be necessary. PCA suggested the four items loaded onto a single component (λ=2.67, loadings N0.76) and Cronbach’s alpha in- dicated strong internal consistency (α = 0.83). Accordingly, we com- bined the items into an additive scale ranging from 4 to 20 with higher scores reflecting a belief that negative publicity surrounding law enforcement had negatively affected colleagues.
5.2.3. Ferguson Effect on citizens’ views It is also possible that, in response to the Ferguson-related contro-
versy, officers have come to believe that citizens’ attitudes toward the police have worsened (see Culhane, Boman, & Schweitzer, 2016). If so, this could have important ramifications for the police. For example, it may lead to further immersion into the “us versus them” nature of the police subculture (Chan, 1996; Neiderhoffer, 1967; Waddington, 1999), which could result in less willingness to work with the commu- nity to solve problems (Braga et al., 2001). To measure the extent to which our sample felt this way about US citizens, respondents were asked to indicate their level of agreement (1 = strongly disagree to 5 = strongly agree) with the following statement: “In general, US citi- zens’ views of the police have gotten worse in the last 6 months.” Sim- ilarly, we asked respondents howmuch they agreed that “Over the past 6months, Marie County (pseudonym) residents’ perceptions of law en- forcement have gotten worse.” This item was used as a fifth dependent variable, Local citizens, in order to consider the possibility that respon- dents felt local citizens’ views differed from those of US citizens more broadly.
5.3. Independent variable
5.3.1. Organizational justice The independent variable of the present study was organizational
justice, which we measured with 18 items intended to capture each component of the construct: procedural, distributive, and interactional
Table 1 Descriptive statistics
N M S.D. Min Max
Dependent variables Less motivation 489 12.85 4.79 5 25 Increased danger 488 10.82 2.52 3 15 Affected colleagues 490 11.74 3.39 4 20 Affected US citizens 504 4.05 0.91 1 5 Affected local citizens 503 2.84 0.99 1 5
Independent variable Organizational justice 426 63.87 13.54 18 90
Controls Self-legitimacy 475 20.47 2.89 5 25 Age 438 2.53 1.03 1 4 Male 435 0.76 – 0 1 Minority 430 0.31 – 0 1 Four year degree 438 0.57 – 0 1 Deputya 425 0.69 – 0 1 Patrol 424 0.39 – 0 1 Experience ≥10 years 429 0.60 – 0 1 Military 434 0.39 – 0 1
a “Mid-level supervisor” is the reference category.
16 J. Nix, S.E. Wolfe / Journal of Criminal Justice 47 (2016) 12–20
justice. All items were measured on the same 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Procedural justice was mea- sured by asking respondents the extent to which they agreed with statements such as “Command staff clearly explains the reasons for their decisions” and “My agency’s policies are designed to allow em- ployees to have a voice in agency decisions (e.g., assignment changes, discipline).” Distributive justice was measured with statements includ- ing “Landing a good assignment in my agency is based on whom you know (reverse coded)” and “Command staff treats employees the same regardless of their gender.” Finally, interactional justice was mea- sured via statements like “Command staff treats employees with kind- ness and consideration” and “Generally, command staff treats employees with respect” (a complete list of the items used to measure organizational justice is available in Appendix A). These itemswere con- sistent with those used in prior research (Cohen-Charash & Spector, 2001; Colquitt et al., 2001) and PCA indicated that they loaded onto a single component (λ = 10.75, factor loadings N0.65). The items also demonstrated strong internal consistency (α= 0.96) and, accordingly, were combined into an additive scale with higher scores reflecting greater perceived organizational justice.
5.4. Controls
In an effort to generate unbiased estimates of the effect of organiza- tional justice on each of the outcome variables, we controlled for several individual officer characteristics. To aid inmaintaining respondents’ an- onymity, we measured age categorically (1 = 21 to 30, 2 = 31 to 40, 3=41 to 50, and 4=51 or older). Gender (1=male), race (1=minor- ity), and education (1= four-year degree or higher)were dummy coded, as were rank (1 = deputy), experience (1 = N10 years), and military background (1= yes). Consistent with the agency’s demographic char- acteristics, about 76% of respondents weremale and 31%were a minor- ity. Finally, we controlled for respondents’ level of self-legitimacy because prior studies have demonstrated its association with a number of desirable officer attitudes and behaviors (Bradford & Quinton, 2014; Tankebe & Meško, 2015; Wolfe & Nix, 2016a, 2016b). Respondents were asked to indicate their level of agreement with the following five statements: “I have confidence in the authority vested inmeas a lawen- forcement officer,” “As a lawenforcement officer, I believe I occupy a po- sition of special importance in society,” “I believe people should always do what I tell them as long as my orders are lawful,” “I am confident I have enough authority to do my job well,” and “I believe law enforce- ment is capable of providing security for all citizens of this county” (Tankebe, 2014). PCA indicated that the items loaded onto a single
component (λ=2.32; loadings N0.56) and Cronbach’s alpha suggested adequate internal consistency (α = 0.71). As such, the items were summed into a scale ranging from 5 to 25, with higher scores indicating a greater sense of self-legitimacy.
5.5. Analytic strategy
The analysis involved the estimation of a series of multivariate re- gression equations that examined the role of organizational justice on officers’ sensitivity to the Ferguson Effect as indicated by five different outcomevariables. Specifically, Models 1 through 3 in Table 2 used ordi- nary least squares regression (OLS) to assess whether organizational justice was associated with less motivation, increased danger, or a greater sense that public scrutiny had negatively impacted respondents’ colleagues in the months following the incident in Ferguson, net of sta- tistical controls. OLS was used because these dependent variables ap- proximated normality (Tabachnick & Fidell, 2013).6 Conversely, ordered logistic regression was used to estimate the relationship be- tween organizational justice and respondents’ perceptions of whether US citizens’ or local residents’ views of the police had gotten worse in the wake of Ferguson in Models 4 and 5, respectively.7
6. Results
Models 1 and 2 in Table 2 were concerned with the extent to which respondents felt they had been directly affected by negative publicity stemming from the Ferguson controversy. Model 1 presents the results of an OLS model that regressed less motivation onto organizational jus- tice and nine control variables. For starters, about 29% of the variation in less motivation was accounted for by the model. Most importantly, deputies who perceived greater organizational justice on the part of their agency and its command staff were significantly less likely to re- port experiencing reduced motivation due to negative publicity sur- rounding law enforcement in the six months following Michael Brown’s death in Ferguson (b=−0.18, p b 0.01). In other words, dep- uties who felt their supervisors treated them fairly and with respect were less likely to experience reduced motivation from the Ferguson Effect.
In Model 2, the increased danger scale was regressed onto organi- zational justice along with each of our control variables. The model accounted for a moderate amount of variation in this measure of the Ferguson Effect (R2 = 0.15). Organizational justice (b = −0.07, p b 0.01) was significantly and negatively associated with a belief among respondents that law enforcement has become more
