Surfing a Web of Trust: Reputation and Reciprocity on CouchSurfing.comPresentation Transcript
Surfing a web of trust: Reputation and Reciprocity on CouchSurfing.com Debra Lauterbach, Hung Truong, Tanuj Shah, Lada Adamic University of Michigan School of Information
Results of predictive modeling
Discussion & future work
“ Worldwide network for making connections between travelers and the local communities they visit”
Whose couch would you stay on?
Online->offline transactions require even more trust
It’s not always easy to decide who is trustworthy
This is why having a reputation system is key
The importance of trust
Bialski & Batorski (2006) examined which factors contribute to higher trust between CouchSurfing friends
Molz (2007) examined the sociological meaning of reciprocity in the context of hospitality exchanges
Does reciprocity exist, and to what degree?
Who is doing the vouching?
Who is being vouched for?
Can we predict which connections are vouched?
Our research questions
Date they joined CouchSurfing
Number of profile views
1,500,000+ friendship connections
Friendship degree (scale of 1-7)
Whether they met in person (Y/N)
How they met (through CS, or offline)
# days traveled together
# days hosted
# days surfed
Whether they vouched (Y/N)
Rating of their overall experience (-1 to 1 scale)
Date friendship connection was made
Most users have few friends (62% have none)
A small number of users conduct the majority of surfing/hosting experiences
Patterns of activity
Direct reciprocity: 12-18% of stays are directly reciprocated
Reciprocity A B
Generalized reciprocity: large strongly connected component (1/3 of active users)
Vouching means you believe that friend to be trustworthy
You can only vouch for others if you have at least 3 vouches yourself
Vouching forms a small “web of trust” in the network
6.8% of users have been vouched at least once
1.8% can vouch for others
Reputation System - Vouching “ Respecting the significance of vouching is essential to the integrity of the network... It is very important that you ONLY vouch for people that you have met in person and know well enough to believe that he or she is trustworthy.”
High rate of vouching
95% of users with > 10 friends have been vouched
15.07 vouches given out on average
25% of friendships that can be vouched are
High rate of reciprocity
74.6% of vouches are reciprocated
Is the vouching system being used as intended?
Tight web of trust….or vouching too freely?
Mutual trust….or social pressure to reciprocate?
Why do users vouch others?
A high number of vouches are between “CouchSurfing friends”
Who are users vouching? Friendship degree: 1= Haven’t met yet 2= Acquaintance 3= CouchSurfing friend 4= Friend 5= Good friend 6= Close friend 7= Best friend
Can we use the available friendship connection information to predict if a connection is vouched?
Logistic regression model (10-fold cross-validation)
71% accuracy in predicting whether a random edge is vouched
Most predictive attributes were friendship degree, rating of experience, how they met
Two-step indirect measure for propagating vouches:
Predicting vouches - global measures A B C D ? Indirect vouch score for A->D: = 1/n(B) + 1/n(C)
Results from logistic regression for each variable alone:
Global measures are poor predictors of whether an edge is vouched