Saturday, March 30, 2013

In group out group biases



·         Babies (presocialization) have it with food preferences. Funny study – puppets that preferred the same food they liked (compared to another food) got helped or harmed by second puppet.  They liked the helper, not the harmer.  But the opposite happened when the first puppet preferred the other food better. They liked the harmer.  Studies show they notice/prefer same language to other language.  They notice gender but no research on prefer.  They don’t even notice race (yeah!).
·         Adults have it for race, gender, age, obesity – all the stereotypes.
·         Add UX Tribal Sides paper

There is an unconscious bias that can significantly impact how people interact with other people and how they interact with even modestly anthropomorphized UIs.  It is the in-group/out-group bias. 

·         When the other person/UI shares some attribute with us, we have a natural, unconscious affinity for them. 
·         When the other person/UI differs from us on this attribute, we have a natural, unconscious bias against them.  At the extremes this leads to racism, sexism, and other explicit, conscious, and even institutionalized and systemic prejudice, but that isn’t what I am talking about here.  I am talking about these subtle influences that we don’t even realize is impacting our thinking, but can impact our decisions and our behavior under the radar. 

At the most basic level, the in-group/out-group bias is adaptive because when we share attributes with someone, we are more likely to share basic values and preferences with them and will be better able to predict their behavior and interact with them more effectively.  So it makes evolutionary sense for us to gravitate towards entities who resemble us in ways that are somehow related to values, preferences, and behaviors.   And if they don’t share our values or preferences, watch out! But since we don’t really know what attributes about a person are relevant, this becomes a more general phenomenon – applying to any attribute that is salient.  You like the same music that I do – you must be OK. 

Then if you go out another degree of freedom, do you care how that person is treated?  If they are similar to you, and you instinctively like them, maybe you will also like a third or fourth person who treats that similar person nicely.  And perhaps you would instinctively like someone who treats the similar person poorly – maybe schadenfreude and maybe “the enemy of my enemy is my friend.”   Or a combination.

So this brings me to a recent study by Kiley Hamlin and colleagues.  She looked at infants (9-month and 14-month) so the in-group/out-group biases that they show would have emerged before culture and socialization had a chance to influence what they found salient and considered relevant.  She wanted to see what bias we were born with.  She figured they would like similar and dislike dissimilar people, or in this case rabbit-looking puppets.  But what would they think of some dog-looking puppets that either stole a ball the rabbit puppet was playing with (the mean dog puppet), or helped them retrieve it and gave it back (the nice dog puppet). 

She started the study by determining which of two foods the infant preferred.  About half preferred a cracker and half preferred a green bean.  She then showed them a rabbit -looking puppet.   In one condition, the rabbit -looking puppet preferred the same food as the infant.  In the second condition, the rabbit -looking puppet preferred the other food.  This is not a huge difference, and certainly not something that should drive an instinctive in-group/out-group bias.  But what else do infants know about?

The dog puppet was then either nice to the rabbit puppet (giving the ball back) or mean (running away with the ball).    When the rabbit puppet had preferred the same food as the infant, the infant liked the nice dog puppet more than they liked the mean dog puppet, even though the nice or mean behavior was towards the rabbit - puppet, not them.  

But the results were reversed when the rabbit puppet preferred the other food.  In this case, the infant liked the mean dog puppet better.  If you take away the food preference part of the study, you would imagine that babies would like nice dog puppets and dislike or at least like less the mean dog puppets.  But for this “out-group” rabbit puppet (what a loser – he prefers THAT food !!), they already show a prejudice – liking it when a dog puppet steals hill ball. 

The effect was larger for the 14-month old infants, so it is something that culture teaches us.  But if it is ALL cultural, then somehow it starts earlier than 9-months old.