According to Daniel Fessler and Colin Holbrook the central role of conflict in human history has led to us becoming expert at making these judgments. Rather than considering each factor in turn, our representation of the odds of a winning a fight is summarised efficiently in a sense of the physical dominance of our would-be opponent.
The researchers tested this idea on the streets of Santa Monica California. They approached 149 men who were either on their own or in a group of 2 to 7 friends. Each participant was taken to one side and shown a picture of a turbaned, bearded terrorist pointing a gun. The photo was cropped so the aggressor's physical size was hidden. The participants were asked to estimate his physical size and muscularity (the terrorist and rating scales are shown above).
The key finding was that participants with one or more friends tended to estimate that the terrorist was shorter by around one and a half inches, and less muscular (having more than one friend in tow didn't exaggerate this effect). In contrast, participants who were alone or smaller stature tended to guess that the terrorist was more physically formidable.
A problem with this first study is that there may be something different about men in groups compared with men who are on their own, and perhaps it's this inherent difference that explains their diverging judgments about the terrorist (e.g. maybe more confident men are more likely to hang out in groups). To get around this, Fessler and Holbrook headed for a public boardwalk by the ocean and approached only men who were in a group. Half of them were tested near their buddies, the others were tested about 100 yards away behind a tent barrier. Once again, the men tested with their friends nearby tended to estimate that the terrorist was less physically formidable, as compared with men tested on their own.
"These findings indicate that the immediate presence of allies is an important factor in men's estimations of the formidability of potential opponents," the researchers said - a result that they suggested could be relevant for "violence prevention, policing and military science". There are some obvious study limitations. The terrorist's physical proportions were kept deliberately ambiguous, which means we don't know how the presence of allies would affect the perception of an aggressor's size if that information was more readily available. Also, would the study replicate with women?