The action movie genre would hardly be complete without one of its most reliable narrative devices: the barroom brawl. Alcohol coats everything like gun powder, one fight breaks out, and an entire room soon descends into drunken fisticuffs. Glasses break, drunkards wield chairs like mad lion tamers, and a body invariably goes flying out the window. It’s all in the name of entertainment, but the relationship between alcohol and aggression is not a Hollywood invention. The U.S. Bureau of Justice reports that, when it comes to violent assault, 2/3 of relationship-partner perpetrators are inebriated, as are 1/3 of stranger-assailants.

One theory for this high correlation is the attention-allocation model, first posited in 1990. This model works on the assumption that alcohol narrows the beam of attention so that drinkers are only able to focus on the most obvious stimuli in any stressful situation. This model could precipitate a fight if the drinker ignores all subtle conciliatory cues in favor of large, aggressive signs. If someone steps on your toe, the pain itself is far easier to process than the many small hand and eye movements that signify that an apology is genuine. And such shortsightedness could easily lead one to throw a punch. 

Researchers tested this model by setting up an experiment to measure aggression from inebriated individuals who were under equal stress but varying levels of distraction. Subjects competed in a game with an imaginary opponent. The stakes were high because a loss in the game put one at the mercy of whatever level of electrical shock the opponent chose to administer. Inebriated participants predictably chose to give harsher shocks, on average, than sober participants.

While drinking did increase aggression in the control case, intoxicated subjects distracted by a moderately difficult memory task were actually less aggressive than the sober group. This alone does not prove that attention deficit is the key to alcohol aggression because the distraction could merely prevent the expression of a still-simmering aggression.

Researchers also found no difference in levels of aggression if the distraction was very small or very large, and this variable outcome may undermine the study’s hypothesis. If the attention model is correct in assuming that aggressive and conciliatory cues require different levels of attention, then it would be logical for a moderate amount of distraction to create the most aggression. Higher levels should (under the model) make it impossible to focus on even aggressive cues and so correct the processing imbalance. The results of this study are certainly not conclusive enough to prove the validity of the attention-allocation model. But one good piece of advice can be taken from the story: the next time you feel a bar brawl coming on, start a (mildly distracting) sing along.