Lay evaluations of police and civilian use of force: Action severity scales

Abstract

Portions of this work have been previously presented at the Society for Personality and SocialPsychology annual conferences (2017, 2018, 2019) and the associated pre-conferences on Justice andMorality and Law and Psychology. Additionally, separate analyses of these data are reported in a companion manuscript (Celestin & Kruschke, 2019). Although the two manuscripts are based on the same data, the analyses are (a) mathematically different and (b) focus on very different aspects of the models.The latent severity scales reported here remain essentially unchanged when using more elaborate models, so we present the most parsimonious model here in order to maximize the clarity of the latent scale derivation and the implications of those results.

Lay evaluations of police and civilian use of force: Moral and physical magnitude ratings of officer and civilian actions

“…The severity of the officer and civilian actions were simultaneously estimated, yielding latent moral and physical severity scales for the entire set of actions. These scale estimates are described in detail in a separate report (Celestin & Kruschke, 2018), and the focus of this report is the regression coefficients and corresponding trends in the ratings.…”

“…Whereas pulling away has physical force comparable to handcuffing, pulling away by itself may seem not to merit handcuffing. Celestin and Kruschke (2018) discuss the normativity of actions in terms of the estimated scale values of the actions.…”

“…This separation was made a priori because we assumed that the scale value for the physical force of an action might be quite different than the scale value for the moral severity of an action. Estimated scale values are discussed at length in Celestin and Kruschke (2018).…”

“…In Figure 3 there is a separate line for each civilian action, and the intercept of each line is given by Equation 1 as (β 0,q + β 3 · ) + (β 1 + β 4 · ) · γ where γ is the latent severity of the civilian action (for discussion of the values of γ see Celestin & Kruschke, 2018). Intuitively, as a civilian action becomes more severe, any specific officer reaction should be less punishable, suggesting that β 1 should be negative.…”

“…The "rug plot" along the horizontal axis indicates the estimated scale values of the 22 actions (i.e., modal posterior χ from Eq. 1, reported in Celestin & Kruschke, 2018), with their SOLOF categories color coded. The panels show a separate line for each civilian action.…”