Received: 15 January 2019 | Revised: 19 February 2019 | Accepted: 25 February 2019 DOI: 10.1002/ab.21829 COMMENTARY The competitive reaction time task: The development and scientific utility of a flexible laboratory aggression paradigm Wayne A. Warburton 1 | Brad J. Bushman 2 1 Department of Psychology, Macquarie University, Sydney, New South Wales, Australia 2 Department of Communication and Psychology, The Ohio State University, Columbus, Ohio Correspondence Wayne A. Warburton, Department of Psychology, Macquarie University, NSW 2019, Australia. Email: wayne.warburton@mq.edu.au Abstract Laboratory measures play an important role in the study of aggression because they allow researchers to make causal inferences. However, these measures have also been criticized. In particular, the competitive reaction time task (CRTT) has been criticized for allowing aggression to be operationalized in multiple ways, leaving it susceptible to “p‐hacking.” This article describes the development of the CRTT and the ways in which its paradigm flexibility and analytic flexibility allows it to test a wide range of hypotheses and research questions. This flexibility gives the CRTT significant scientific utility, but as with any research paradigm, comes with the obligation that it has to be used with integrity. Although safeguards exist and there is little evidence of misuse, study preregistration can increase confidence in CRTT findings. The importance of findings such as those of Hyatt et al. (in press), which provide further evidence for the validity of the CRTT, are also noted. KEYWORDS aggression, analytic flexibility, competitive reaction‐time task, laboratory measures, noise blast 1 | INTRODUCTION For theoretical and practical reasons, it is critical for researchers to investigate why people sometimes behave aggressively. Aggression is conceptualized as any behavior intended to harm another person who does not want to be harmed (Anderson & Bushman, 2002; Baron & Richardson, 1994). In the field of human aggression, as in other fields, research evidence is seen as convincing if it is found across a range of methodologies, each of which corrects for the weakness of another (Warburton, 2014, 2015; Warburton & Anderson, 2015, 2018). In aggression research, laboratory experiments typically measure relatively mild forms of aggression but have the powerful strength of being able to confer causality to an effect through designs that experimentally control for the influence of other factors. Thus, they play a crucial role in the investigation of aggressive behavior and are especially valuable when they provide causal data related to effects that have already been found in cross‐sectional, observa- tional, longitudinal, and brain imaging studies. Although there have been several laboratory measures of aggressive behavior used in past decades (Anderson & Bushman, 1997; Ritter & Eslea, 2005), this number has been somewhat limited by the ethical restrictions that must necessarily guide the creation of such measures (i.e., the aggressive impulse must be short‐lived and relatively mild so that participants are not permanently harmed). The measure must also reliably and validly measure aggression and should generalize to behavior outside the laboratory. In addition, to reduce socially desirable responding, participants usually need to be naïve to the laboratory aggression measure that is being used in their study. As a consequence, when a laboratory aggression measure becomes well known (e.g., through media coverage or being taught in undergraduate courses), data from that paradigm may not be free of social desirability biases and the inclusion of suspicion checks becomes crucial. Together these factors have ensured that labora- tory aggression paradigms are often complex, few in number, and sometimes controversial until they are methodologically refined to the point of well‐established validity (and occasionally even then). Aggressive Behavior. 2019;1–8. wileyonlinelibrary.com/journal/ab © 2019 Wiley Periodicals, Inc. | 1