Weapons as Aggression-Eliciting Stimuli


Berkowitz and Lepage (1967) designed a study to test the hypothesis that individuals who are in a state of anger are more likely to act out their aggression if cues associated with violence and aggression are present.  The sample consisted of 100 male students from the University of Wisconsin who were all enrolled in an introductory level psychology course.    

This study used an experimental research method because it manipulated the independent variable and presumably involved random assignment (although this was not stated in the text). There were two main independent variables.  The first one was the subject’s level of anger and this was determined by whether the subject was shocked once or seven times.  The second independent variable was the kind of cue present near the shock button when it was the subject’s turn to evaluate the confederate.  For one group there was no object, in the control group there was a neutral object (a badminton racquet), and for the last group there was a gun that was supposedly part of a different study.  This last group was further separated into 2 subgroups with some being told that the gun belonged to the confederate while others were told that it was left behind by someone else.  These independent variables were then combined to see how they affected the dependent variable, which was the level of aggression the subject displayed.  The dependent variable was measured by how many shocks the subject delivered to the confederate.

The procedure ran as follows: volunteers were told that they were participating in a study to test the physiological effects of stress.  To do this the subject and the other participant (who was actually a confederate) were both given a social problem and they had to think up ways to solve it.  After they completed this task (in separate rooms) their problem solving ideas were then exchanged so they could evaluate each other.  The evaluation was done by pushing a button that was supposed to shock the person in the other room (although they still could not see each other); 1 shock represented the best rating while a lesser evaluation was communicated through a higher number of shocks.  The confederate was the first to do the evaluation.  The number of shocks given to the actual volunteer was already determined as 1 or 7 though (depending on the random assignment) and was not based on a real rating.  After this came the volunteer’s turn to do the same evaluation, but the number of shocks was not predetermined.  Next to the shock button was one of the previously stated objects, and the gun was the only cue hypothesized to elicit increased aggression.

The results of this study confirmed the hypothesis.  Those participants who were more angered (given 7 shocks) and were cued by the violent object (a gun) and told that it belonged to the person they were rating, outwardly expressed their aggression the most by giving the confederate a higher number of shocks.  The next highest number of shocks was by the group in the presence of a gun, but had been told the gun was left behind by someone else.  Those who did not see any objects gave on average one less shock and the least number of shocks were given by those in the presence of the badminton racquets.  On the other hand, when the volunteer was not as angered (only shocked once by the confederate), outward expression of aggression was relatively low and stable regardless of what type of cue was present.  The researchers used these results to theorize that a person who is already aroused and is then cued by a violent object is more likely to have an impulsive reaction to act more aggressively.


Overall this study was well designed in order to test the given hypothesis that weapons are aggression-eliciting stimuli.  The method of using different objects to induce a given response is very similar to the proven phenomenon of priming.  Priming is where certain information is more attended to when related cues are presented.  Therefore the results of Berkowitz and Lepage (1967) make sense because weapons are connected to aggression, which increases the person’s awareness of his or her aggressive feelings, and consequently makes the outward act of aggression more likely. 

Based on the results, chances are high that these men would always act in this way when in a similar situation, so this study can be considered reliable (that is, it is repeatable).  Validity is not as strong, though.  Validity refers to whether the study is measuring what it purports to measure. When the participant was already aroused (given 7 shocks) there was a significant difference in the amount of retaliation depending on which cue was present.  However, this retaliation did not depend on the cues if the participant was not as initially aroused (only given 1 shock).  So how can they be measuring the impact of a priming mechanism like the gun in the room if they need participants to already be aroused? I am not sure they are measuring their variables correctly. That being said, it did show that although the cues do have an effect on aggressive behavior, initial aggression level plays a much larger role in the causal relationship.  The ethicalness of this study is also questionable.  Receiving and delivering shocks could potentially cause physical pain and also have a negative effect on one’s emotional well-being.  Nonetheless, most participants probably did not suffer any serious consequences.  Also, due to the nature of this specific research question it does not seem like there is another way to measure aggression that would be anymore ethical.


One major methodological problem that should have been addressed is the sample that was obtained.  The sample used in this experiment is not a good representation of humans in general, because it only involved college-aged men.  It is possible that women or people of different ages may respond differently to the cues.  Women are often thought of as less violent, so their reaction to a negative stimulus might cause them to deliver fewer shocks.  A weapon makes the seriousness of the situation salient and may cause some people to think rationally about their behavior in the near future.  Clearly this proposal requires actual testing before making further assumptions, but it does show the need for a more diverse sample of participants. 

Along the same line as the previous issue, a follow-up study could more carefully look at the relationship between peoples’ attitudes towards guns (or other weapons) and their corresponding level of aggressive behavior when given the chance to retaliate.  This would be more of a quasi-experiment because in order to test the independent variable of attitudes towards weapons the groups could not be randomly assigned.  Three existing groups would be used; those who support weapons, those who are against them, and those who feel neutral (the control group).  The hypothesis would predict that if prior arousal level was high, participants who support weapons would show increased aggression when cued by the gun, but the group of participants with negative attitudes towards guns would not be as aggressive. If the subjects did not receive prior arousal (if they were only shocked once by their “evaluator”), then neither group would be significantly affected by the cues.

Even if initial aggression is a greater cause in inducing violent behavior than the existence of weapon-related cues, this study has serious implications for social policies related to gun control.  It is apparent from the results that if someone is angry and is near a gun, then that person will likely act more aggressively than he or she typically would.  Since the guns in the experiment were not loaded and the situation was controlled, the heightened aggression was not transferred over to actually using the guns.  In a private home though, arguments occurring with a gun nearby might make it more likely that a gun will be used.  Knowing that the mere presence of a weapon increases violence should urge lawmakers to consider adopting stricter gun laws.

Brief summary

Berkowitz and Lepage (1967) conducted a study in which they hypothesized that priming people with an aggressive object (a gun) would lead them to act aggressively. The authors gave electrical shocks (from 1 to 7 of them) to 100 male undergraduates. They told them that one of their peers had delivered the shock. The participant could then retaliate, but they did so in the presence of either a gun or a tennis racket (which was supposedly left in the researcher room from a different study). Participants given the highest number of shocks (7) gave higher retaliation shocks to the peer, but this was more likely when they were in the presence of a gun (compared to a tennis racket). The authors concluded that the guns increased aggressive responses from male participants who were highly aroused.


Berkowitz, L., & Lepage, A. (1967). Weapons as aggression-eliciting stimuli. Journal of Personality and Social Psychology, 7, 202-207. doi: 10.1037/h0025008