Using personal network composition and structure to explain ethnic identity

1
Using personal network composition and structure
to explain ethnic identity

• José Luis Molina
• Autonomous University of Barcelona, Spain
• Miranda Lubbers
• Autonomous University of Barcelona, Spain
• Chris McCarty
• University of Florida, USA

National Science Foundation - BCS-0417429
2
Ethnicity, identity, identification

• The literature about identity in Social Sciences
  and Humanities is enormous (Psychoanalisis,
  simbolic interactionism, anthropological
  perspectives on ethnicity ...).
• We develop our research in the theoretical
  background that focus both in social structure
  and context of interaction for explaning
  behaviour and cognition.

3
Ethnicity, identity, identification (ii)

• We use the concept of identification (Brubaker,
  2000), instead of identity, because it allows to
  define self-identifications, how ego identifies
  alters and how alters identify ego.

4
Research question

• Do personal networks structure and composition of
  migrants in Spain allow us to explain variance
  in ethnic identifications?

5
Data

• Snowball sampling (2005-2006) 294 immigrants in
  Barcelona from four migrant groups (for the
  Spanish part of the project)
• 78 Senegambians 70 Moroccans 81 Argentineans
  65 Dominicans.
• Personal interviews were held software Egonet
  was used to collect data about
• 1. Characteristics of the respondent
• 2. List of 45 alters (personal network
  delineation)
• 3. Characteristics of each of the alters (network
  composition)
• 4. Whether each pair of alters was related or not
  (network structure)

6
Data (ii)

• We used two open questions in each structured
  interview for eliciting identity data
• Which word or phrase best describes your ethnic
  identity?
• What other word or phrase best describes your
  ethnic identity?

7
Codification

• All responses were codified in a single
  dimension from ethnic-exclusive identification to
  generic-non ethnic identification.
• We codified both responses for each informants in
  the following way

8
Codification (ii)
9
Codification (iii)
10
Types of personal networks

• With the same data we performed a cluster
  analysis on personal network characteristics and
  we found 5 types of personal networks
• The scarce network N 54
• The dense family network N 28
• The multiple subgroups network N 73
• The two worlds connected network N 75
• The embedded network N 50

11
Description of profiles
12
Profile 1. Scarce network
Color country of origin (white foreign, black
Spain) Size country of living (large
Spain, small other country)
13
Description of profiles
14
Profile 2. Dense family network
Color country of origin (white foreign, black
Spain) Size country of living (large
Spain, small other country)
15
Description of profiles
16
Profile 3 Multiple subgroups network
Color country of origin (white foreign, black
Spain) Size country of living (large
Spain, small other country)
17
Description of profiles
18
Profile 4 Two worlds connected
Color country of origin (white foreign, black
Spain) Size country of living (large
Spain, small other country)
19
Description of profiles
20
Profile 5 Embedded network
Color country of origin (white foreing, black
Spain) Size country of living (large
Spain, small other country)
21
Personal Networks identifications
?2 41.3, df 20, p lt .01
22
Network characteristics that are responsible for
this outcome

• Density
• People with denser networks more often identify
  ethnically exclusive
• Number of homogeneous subgroups within the
  network
• People with higher number of subgroups more often
  identify transnationally
• Percentage of Spanish alters in the network
• People who have more relations with Spanish
  alters identify more often transnationally or
  generically

23
Do personal networks allow us to explain
additional variance in identifications?

• Multinomial logistic regression to predict ethnic
  identifications from
• Network profiles
• Control variables Years of residence, country of
  origin, gender, age

24
Results multinomial logistic regression

• Years of residence and culture affect ethnic
  identification
• Transnational identifications were less observed
  among more recent migrants
• Senegambians had most often ethnically exclusive
  identifications Moroccans generic
  identifications Argentineans and Dominicans
  ethnically plural identifications

25
Results multinomial logistic regression

• Age and gender do not affect ethnic
  identification
• Network profile still has a significant effect on
  the ethnic identifications when controlled for
  these background characteristics

26
Discussion

• Looking at personal networks we can understand
  social and cultural phenomena better because they
  are the result of both (macro) structural
  constrains and (micro) individual choices.
• People with dense, homogenous personal networks
  identify themselves according with dominant and
  exclusive categories.
• People with more heterogeneous networks are more
  free to use plural identifications and adapt
  themselves to the multiple contexts of
  interaction in which they are embedded.

27
Thank you

• The paper can be obtained via joseluis.molina_at_uab
  .es