In the past few decades, aggression researchers have emphasized the role of cognitive structures and processes in understanding aggressive behaviors. Understanding how people think can deepen our understanding of when and why people behave aggressively. As a consequence of all this research, aggression researchers, like any in-group that looks for useful labels, have created a vocabulary for describing various phenomena. It is commonplace to read an academic aggression article and see terminology such as hostile expectation bias or hostile attribution bias.
As buzzy as these labels sound, sometimes they are used imprecisely.
Take, for instance, a class of phenomena that aggression researchers have lumped together under the umbrella hostile appraisals—a sort of label under which we can place other labels. So far, there have been three supposedly distinct phenomena that have been placed under this umbrella: hostile attribution bias, hostile perception bias, and hostile expectation bias. In the future, I wouldn’t be surprised to see others tucked under this umbrella too.
Here is a good description of hostile appraisals taken from Bushman (2016). I’ll use this passage just to provide concrete examples, but this type of wording is commonplace.
This passage (and others like it) sound pretty straightforward. But there is a subtle conflation between describing the hostile biases as an empirical effect and as a process through which the effect operates.
Let’s break that down.
Are hostile appraisals empirical effects? It seems so. Hostile expectation bias, for example, is described as “the tendency to assume that people will react to potential conflicts with aggression.” Thus, hostile expectations are an observation. The meta-analysis this passage is taken from examined whether exposure to violent media affected hostile expectations. Thus, the empirical effect is that violent media is causing a systematic change in an outcome variable that is labeled “hostile expectations.” The cognitive process through which violent media causes a change in hostile expectations is important, but is conceptually distinct from this basic violent media-hostile expectation (input-output) relationship.
Are hostile appraisals cognitive processes? It also seems so. These hostile biases are claimed to influence the appraisal and decision process. That these biases are considered a cognitive process is abundantly clear when you look at other passages from this manuscript such as, “[a]ccording to the General Aggression Model, internal states can influence appraisal and decision processes. This meta-analysis focuses on hostile appraisals.” When you look at the accompanying diagram, a common-sense reading is that these hostile appraisals are something that occurs in the box labeled “present internal state” because that would be the point at which they can affect the box labeled “appraisal and decision processes.”
So what? This conflation of hostile-appraisals-as-an-effect and hostile-appraisals-as-a-process can lead to statements such as, “hostile expectation biases (an effect) are caused by hostile expectation biases (a cognitive process).” This obviously is a non-sensical statement.
For this reason, we need to maintain conceptual distinctions between the empirical effect and the process through which that effect operates. If we describe hostile appraisals as an empirical effect, which I think is the right course of action, then we need to describe the process through which that effect operates using different terminology. Ideally, the process is described in clear, mechanistic, and testable terms. You cannot merely say there is a “hostile bias” happening in the cognitive process—namin’ ain’t explainin’.
This conceptual distinction is important beyond precise wording. If we reserve the label hostile appraisals exclusively to talk about empirical effects, then it is possible that we can discover several cognitive processes through which that effect might operate. For example, we might find that sometimes a hostile expectation occurs through process A and sometimes through process B. Further, we might find that several different phenomena that we lump together under the hostile appraisal umbrella might actually have a common process. For example, we might discover that hostile expectations and hostile attributions work through the same cognitive process and we give them different labels merely because we use different outcome measures. In this case, suppose violent media initiates cognitive process A. If we measure the expression of cognitive process A using one outcome, we might label the effect a “hostile attribution,” and if we measure the expression of the same cognitive process using another outcome, we might label the effect a “hostile expectation.” These possible discoveries get lost if we do not respect the conceptual distinction between the to-be-explained effect and the process through which those effects operate.