Psychology Behind Direct Benefit Effects and Information Cascades
Recently in lecture, I have found information cascades and direct-benefit effects to be the most interesting subject. In particular, I have been interested in the psychology of them, because I doubt that in many of the scenarios discussed in class (like the marble scenario) people would actually find themselves calculating the exact chance of a red or blue marble turning up given the previous number and type of responses, they would estimate.
Luckily, a good perspective from which to look at the psychology of decision making based on the choices of people who came before you came up in a problem set recently. Students were asked to consider two invitee lists to an event. A list where guests can see who has already responded and how they responded, and a list where it is impossible to see who is coming and who is not. Keeping with the theme, I searched for articles discussing Evite, an online invitation service which can show invitees who is coming and who is not. The article I came across, by the New York Times, centered around a topic I had not thought of, but which was nevertheless very relevant information cascades. It had to do with signaling, in this case, in what manner invitees accepted or rejected the invites they received. The article made me reflect on how different signaling relating to the same final decision (going to the party or not) changed future users’ decisions.
On Evite, when rejecting or accepting an invitation, a message can be written. In the article, a user who is rejecting an event writes “What’s going on on the 11th? This is the third invitation for a party I had to decline for the same day. I will be in Florida…” I realized immediately that because it is possible to make a statement justifying your attendance status, not making a statement justifying your attendance status is also de facto a statement. What is the effect of a ‘no’ response with no message versus a ‘no’ response with a message saying ‘I will be at my mother’s funeral that day’? I believe that when considering attending an event or not, a user will be more likely to go if many friends are going, and less likely if less friends come. Furthermore, I think that if few friends are going, but they all have good excuses as to why they are not going, people will themselves feel more compelled to go if they do not have an equally good excuse. On the other hand, if friends are not attending events and post excuses like “I can’t come, im feeding my dog at that time”, or even no excuses at all, users will feel less pressure to come up with a good excuse themselves, and are more likely to not attend for an illegitimate reason. Obviously, these questions and conclusions delve beyond the mathematical and probabilistic models which information cascades and direct-benefit effects fit into, however I think their consideration might result in more accurate predictions. Could seeing friends not attending but for very good reasons (their mother died) shame someone into an attending an event they don’t want to go to but do not have a good excuse for? I would love to explore this further, however it seems to lie beyond the scope of this class.
Article mentioned above: https://www.nytimes.com/2006/12/07/fashion/07evite.html