Role of Weak Ties - ccb

The Strength of Weak Ties:
A Network Theory Revisited
Mark Granovetter
Weak Ties
Weak Ties
Ties between Ego and an acquaintance
Comprise a low-density network
Role of Weak Ties
A crucial bridge between the two clusters of close friends
- Individuals with few weak ties will be deprived of information
and confined to the provincial news and views of close friends.
Social systems lacking in weak ties will be fragmented & incoherent
- new ideas spread slowly and subgroups separated by race, geography,….
The Impact of Weak Ties on Individuals
Bridging Weak Ties
Weak ties linking different groups
Lead to complex role sets and the need for cognitive flexibility.
Weak Ties & Opportunity for Mobility
Better access to job information through weak ties
For well-educated groups weak-ties are most likely to be used
but less-educated groups rely on strong-ties for job search.
(Ericksen and Yancey, 1980)
The weak ties are effective only so far as the weak ties connect the respo
ndent to a high-status individual. (Lin et al., 1981)
Weak ties are more efficient at reaching high-status individuals.
Only bridging weak ties are of special value.
Strength of Strong Ties
Strong Ties
Greater motivation to be of assistance & easily accessible
Relative frequency of strong ties are greater for low class.
- low-status individuals are numerous as compared to high-class,
 easier to pick friends similar to them
Strong Ties & Economic Insecurity
Under strong pressure, individuals depend on strong ties.
e.g., unemployment, new Ph.D.s (Murray et al., 1981)
Economic Insecurity and lack of social service
- Reciprocity network in black community (Stack, 1974)
: share with one another because of urgency of their needs
 provides a minimal level of economic security
Weak Ties in the Spread of Ideas
Weak ties & cultural diffusion (Fine and Kleinman, 1979)
Children’s culture varying by regions spread via weak ties.
 homogenous subcultures at the end point of diffusion process.
Weak ties & scientific innovations
Scientific field has a center & periphery defined by weak ties to
the center and to other scientific groups
Marginals can better afford to innovate and the innovations are adopted
by the center.
Local bridges and weak ties promote the regular flow of novel
information in differentiated structures. (Friedkin, 1980)
Most of the influence is carried through strong ties. (Weimann, 1980)
 Weak ties: marcointegration/ strong ties: microintegration
Weak Ties and Social Organization
Weak ties & peer socialization
Weak ties can create diverse cultures without becoming similar.
Weak ties can reduce student alienation and increase social solidarity.
- Racial integration can be achieved by producing enough weak contacts.
e.g., rearranging classroom structure (Karweit et al., 1979)
Weak ties & organization integration
Weak-tie mode of organizational integration (highly differentiated
system of specialized staff relationship) leads to high morale and
good service (Blau, 1980)
The group recruited based on weak ties were able to implement many of
aims while that based on strong ties was not successful. (Steinberg, 1980)
Weak intercorporate ties can create cooperation and coordination
while intense ties tend to produce competitive and exclusive relationship.
Empirical validations
How can we define weak ties & bridging weak ties in an online social net
work setting?
# of in and out degrees to other nodes
Using self reported information about the relationship
To what extent do weak ties play a role in product diffusion?
Possibly depends on the product category and the extent of newness
Are weak ties in one setting still weak ties in a different setting?
The Role of Hubs
in the Adoption Process
Goldenberg, Han, Lehmann, and Hong
Influence is a combination of personal and social factors
Personality: “who one is”
Competence: “what one knows”
Strategic social location: “who one knows”
The role of social hubs in the adoption process
Hubs: individuals with a large number of social ties
≠ Opinion Leaders / innovators
Even if a hub is not an innovator, a hub is more likely to adopt early
due to the greater exposure.
H1: Social hubs are more likely to adopt at the early stages
Hubs will adopt first due to their greater exposure to an innovation
even though they are not innovators.
H2a: When hubs adopt, the overall adoption process speeds up.
H2b: Innovator hub adoption has a larger correlation with speed of
adoption than follower hub adoption.
More connections will be activated once hubs adopt  adoption rate ↑
Since innovator hubs adopt earlier than follower hubs, they have
more time to influence the network.
H3: The higher the relative out-degree of a hub,
the greater impact it has on adoption.
in-degree: the number of people who convey information to hubs
 related to when a hub adopts.
out-degree: the number of people to whom hubs convey information
 the influence of hubs on subsequent adoption
H4a: Hubs adoption increases the eventual size of a market.
H4b: Follower hubs have a stronger relation to market size than
innovative hubs.
Adoption by hubs  exposure of an innovation to the market ↑
 market size ↑
Homophily effect
: follower hubs are more similar to most of the population in terms of
innovativeness  more influence on the main market.
H5: Hub adoption at an early stage can be used to predict product
If hubs do not adopt a product soon after its introduction, this may
impede adoption by those are connected to the hubs.
Data & Measures
Data from a Korean social network site, Cyworld
“Scrapping”: adopting items such as pictures or video clips from other
people’s blogs people visit.
Collected item #, time of scrap, creator ID, # of in & out degrees of nodes
Out-degree: # of other nodes ever visited by the hub
In-degree: # of other nodes that have visited the hub
Hubs: people with in and out degrees larger than
three standard deviations above the mean
Innovativeness: the average # of people who adopted an item
in the neighborhood is in the 16% of all hubs.
Empirical Results
Adoption timing
Hub adoption decline over time while nonhub adoption increases over time.
 Social hubs are more likely to adopt at the early stages. (H1 supported)
Hubs adopt early even if they are not innovative. (See Table 2)
Speed of adoption process
Hub adoption had stronger impacts on the adoption speed than typical adoption.
(H2a supported.)
An Innovative hub’ effect on the adoption time more than twice that of
a follower hub. (H2b supported)
Mean time adoption of a node linked to innovative hubs: when 55% adopted
vs. that of follower hubs: when 69% adopted
Hubs that have higher out degrees than in degrees seem to be more effective.
in speeding the adoption process. (H3 supported.)
Empirical Results
Market size
The number of hub adoptions is a good predictor of eventual market size.
(H4a supported)
Follower hubs have around seven times impact on the market size than
innovative hubs (H4b supported.)
Predicting Product Success
Test whether hub adoption can predict product success
(top 30 popular items vs. 30 moderately successful items)
Even when using hubs that adopted at 5% adoption level, 80% hit.
Using only small sample of hubs (280 hubs) adopted at 5% adoption rate, 70% hit.
Contributions & Limitations
Empirical validation of the role of social hubs in the product diffusion
process using a network data in a natural setting
Linking individual level network characteristics to aggregate diffusion.
Can be used for predict product success
Generalizability to other settings
- Costs of “Scrapping” is almost close to 0.
Is diffusion only a function of exposures to an innovation?
Is “scrapping earlier than others” related to innovativeness?
- “Scrapping” can be viewed as somewhat lazy, passive activity.
Incorporating other characteristics of hubs
Whether a hub is bridging diverse nodes or not can affect the adoption
Predicting product success using social hubs
For some products (e.g., luxury, fashion items), adoption by the mass
is not the goal while for some products (e.g., pop music, movies) it is
critical to appeal to the mass market.
How do the roles of social hubs differ in these different settings?
Can we still predict the product success for luxury goods using
the adoption of social hubs?
The end
Thank You!
17 / 10

similar documents