According to science, people are mostly good.

A recent study gives us a fascinating look into how humans “naturally” behave.  Scientists collected data from the online game Pardus, which has over 300,000 players, and crunched it through a series of analysis techniques for complex systems.  The game is one where the players are free to behave in any way they choose, with no built in advantages for good or bad.  The bottom line?  Humans are mostly good, most of the time, and they generally behave well even if there are no laws preventing bad behavior.

According to the study:

 The analysis of binary timeseries of players (good-bad) shows that the behavior of almost all players is ‘good’ almost all the time. Negative actions are balanced to a large extent by good ones. Players with a high fraction of negative actions tend to have a significantly shorter life. This may be due to two reasons: First because they are hunted down by others and give up playing, second because they are unable to maintain a social life and quit the game because of loneliness or frustration. We interpret these findings as empirical evidence for self organization towards reciprocal, good conduct within a human society. Note that the game allows bad behavior in the same way as good behavior but the extent of punishment of bad behavior is freely decided by the players. (LINK)

Another interesting observation was that negative behavior is more repetitive than reciprocal.  In other words, those who do bad often repeat bad behaviors.  Scientists also found that people who have recently been wronged are 10 times more likely to do wrong themselves compared to people who have been treated fairly.  This is a powerful truth that has far-reaching implications. There are, apparently, bad apples, and they can spoil the bunch.  (Hint for Christians:  Here’s your empirical justification for rewarding good behavior and punishing bad behavior without reference to a deity.)

There is also an implicit warning in this finding:  In the complex real world, there is often little difference between being wronged and feeling as if you’ve been wronged.  We would expect that people who perceive themselves as victims of wrongdoing would also be more likely to act wrongly.  This is a chilling observation when we consider that the Republican Party and FOX News specialize in villainizing an entire class of people (liberals) and convincing its viewers that they are being wronged.  It has created a culture where many, many people believe there is a left wing conspiracy to take their money, their freedom, and their rights.  It has taught us that anger is the correct response to political opposition.  If this study is correct, they are causing great societal harm, and greatly increasing the probability that individuals will do evil.

Of course, this study plunges another dagger into Christian theology, which asserts that humans are inherently evil, and need the moral foundation of an external deity.  Rather, it appears, humans are exactly as evolutionary theory predicts — they are good most of the time, naturally.  Habitual evil-doers are rare, and the group sees to it that they are punished.  Evil begets evil, and good begets good.  It’s very logical, very natural, and very predictable.  There was no external pressure to be good in this game, either from a god or game coding, and people behaved well most of the time.

This study is very important for several reasons.  First, it has the advantage of complete data, which is almost unheard of in social science.  Studying humans has always involved some degree of educated guesswork because we simply cannot know everything about any individual.  In this case, every player’s actions were logged for the entire duration of their time online.  The scientists literally knew everything thateverybody did.

Another difficulty in social science is the ethics of manipulating people’s private lives.  We simply cannot put subjects in dangerous or life-altering circumstances just to see what happens.  In a game, however, we have much more latitude for experimentation.  Pardus was chosen because it contains 8 core activities that define every human society:  Communication, trade, making friends, making enemies, ending friendships, reconciling with enemies, attacking, and punishing.  Furthermore, there are no restrictions in the game itself on whether or not people may act badly.  Each player may be as good or bad as he wants.  In many ways, it is the perfect analog for the real-life experiment we couldn’t ethically perform — putting a lot of people on an island with no rules to see what happened.

Finally, this study eliminated one of the most consternating problems with studying real humans:  Actions can very rarely be seen as “all good” or “all bad.”  The nuance of complex human relationships makes binary analysis virtually impossible.  In the game environment, the mechanics are very simple.  Attacking people is bad.  Being nice to people is good.  Trading fairly is good.  Cheating is bad.  Simple.  Binary.  Crystal clear.

Continuing research in this area has exciting implications for social engineering, especially in societies that are going “down the wrong path.”  While most people in game were good most of the time, the scientists found that there’s truth to the platitude that evil begets evil.  Pockets of bad behavior tended to be self-reinforcing, and people tended to act badly if they had been the target of a bad action.  By understanding the nature of these pockets of bad behavior, we can learn how to identify them before they grow, and intervene to turn them in a more prosocial direction.