the social name-letter effect on online social networks

20
The Social Name-Letter Effect on Online Social Networks Gabriel Magno Ingmar Weber Farshad Kooti This work was done while the first two authors were interns at QCRI

Upload: gabrielmagno1

Post on 03-Aug-2015

159 views

Category:

Social Media


0 download

TRANSCRIPT

The Social Name-Letter Effect on Online Social Networks

Gabriel Magno

Ingmar WeberFarshad Kooti

This work was done while the first two authors were interns at QCRI

The Name-Letter Effect (NLE)

The Name-Letter Effect

Sarah

SamStarbucks

San FranciscoSales

Where the SNE Has Been Observed

● Nuttin (1985): letter-picking– across languages [Hoorens 1990]

● Observed in decision making situations– Cities and jobs [Pelham 2002]

– Brands [Brendl 2005]

– Donations [Bekkers 2010]

– Marriage [Jones 2004, Simonsohn 2011]

– Favorite foods & animals [Hodson 2005]

● Challenging the existence of NLE– Biases– Methodology

N

A

S

Arthur

Motivation● Widely cited effect in psychology,

but has been questioned by some researchers

● Understanding factors affecting social link formation is important

● Successful market example: "share a coke" campaign– People love their name!

NLE on Online Social Networks

Datasets

● Twitter: 2009– 52 Mi users– 1.9 Bi edges

● Google+: 2012– 61 Mi users– 1 Bi edges

Methodology

AUsers X*

Users Y*

Brand X* Brand Y*

C

B

D

(A+C )∗(A+B)

A+B+C+DExpected

value for A

Observed value > Expected value → NLE

A # users X* following brand X*→B # users X* following brand Y*→C # users Y* following brand X*→D # users Y* following brand Y*→

Brand Preference

Statistically significant NLE

Other NLE analysis● Social Link: names with same 1st letter

– Statistically significant, but negative NLE

● Location: 10 US cities, 7 letters, 21 pairs– 8 pairs, 6 positive

● Occupation: 5 jobs, 5 letters, 10 pairs– 1 pair, negative

● Hobbies: 8 sports, 7 letters, 21 pairs– 4 pairs, 1 positive

Overall, no consistent Name-Letter Effect

James John

The Same-Name Effect on OSNs

Name Preference

● Correcting for country (US) and gender● Top 5 popular names

● Similar results for Google+ and females

Users significantly preferred to follow other users with the same name

Ethnicities

● Last name as proxy for race● 5 races (US), top 500 “exclusive” last names

Strong and consistent SNE for all the 5 races

Other SNE analysis

● Language: Brazil, Egypt, Germany– Effect sizes between 6% and 101%

● Social Tie Strength: common friends– People are more affected by SNE when they

are establishing a weak link

● Number of friends:– Users with fewer friends are more likely to

follow other users with the same name

Conclusion

Concluding Remarks● Investigated the existence of the NLE in the context

of Twitter and Google+

– Findings question, at least, the generality of the NLE

● Investigated the existence of SNE

– Robust effect, even when accounting for gender, age, race, and location

Remember that a person's name is, to that person, the sweetest and most important sound in any language.

Dale Carnegie, in the book “How to Win Friends and

Influence People”

Thank You!

http://www.dcc.ufmg.br/~magno

@GabrielMagno

Backup Slides

Social Link Preference

Statistically significant, but negative NLE