Introduction To Ego Network Analysis

Introduction to Ego Network AnalysisIntroduction toEgo Network AnalysisRich DeJordyDan HalginBoston Collegeand theWinston Center for Leadership & Ethics 2008 Halgin & DeJordyAcademy of Management PDWPage 1

Introduction to Ego Network AnalysisGoals for Today1. Introduce the network perspective–––How is ego-centric analysis different from socio-centricanalysis?When and why ego network analysis?What theories are ego-centric?2. Research design and data collection3. Data analysis4. Review and demo of software tools–Egonet, E-Net 2008 Halgin & DeJordyAcademy of Management PDWPage 2

Introduction to Ego Network AnalysisWhat is Unique about Social NetworkAnalysis? Phenomenon studied– Distinctive type of data,– It’s about relations & structure How we study it– Distinctive tool– Typical statistical methods may not apply How we understand it– One “network perspective”– Based on multiple theories (Simmel, Blau) 2008 Halgin & DeJordyAcademy of Management PDWPage 3

Introduction to Ego Network AnalysisMainstream Logical Data Structure 2-mode rectangularmatrix in which rows(cases) are entities orobjects and columns(variables) areattributes of the cases Analysis consists ofcorrelating columns– Emphasis on explainingone variable 2008 Halgin & DeJordyIDAgeEducationSalary1234Academy of Management PDWPage 4

Introduction to Ego Network AnalysisNetwork Logical Data 0Bob100-Email - 2008 Halgin & DeJordy Individual characteristics only halfthe story.RELATIONS MATTER! People influence each other, ideas& material flow Values are assigned to pairs ofactors Hypotheses can be phrased interms of correlations betweenrelationsAcademy of Management PDWPage 5

Introduction to Ego Network AnalysisRelational Data & Attribute 15000Relational DataAttribute DataSNA provides the ability to combine relational data withattribute data (e.g., homophily, heterogeneity, etc) 2008 Halgin & DeJordyAcademy of Management PDWPage 6

Introduction to Ego Network AnalysisSocio-centric(Whole/ Complete network)Ego-centric(Ego/Personal network)EGOALTERS Focus on the whole group Focus on individual ego networkso Global structure Patterns of interaction used to explain:o Concentration of powero Flow of information or resourceso Status structureso Structureo Compositiono Shape Cases are individual ego networkso Generalized to other ego networks Cases are complete networkso Generalized to other networks 2008 Halgin & DeJordyAcademy of Management PDWPage 7

Introduction to Ego Network AnalysisEgo Network AnalysisMainstreamSocial SciencedataNetworkAnalysisEgoNetworksperspective Combines the perspective of network analysiswith the data of mainstream social science 2008 Halgin & DeJordyAcademy of Management PDWPage 8

Introduction to Ego Network AnalysisEach Ego Network is Treated as itsOwn WorldGeorgeMollyGidgetOr in more typical language, each ego network is treated asa separate case 2008 Halgin & DeJordyAcademy of Management PDWPage 9

Introduction to Ego Network AnalysisWhy Study Ego Networks?Ego’s network is a source of: Information Social support Access to resources Sense-making Normative pressures Influence etc.All of which can influence Ego’s behavior 2008 Halgin & DeJordyAcademy of Management PDWPage 10

Introduction to Ego Network AnalysisWhen to use Ego Network Analysis If your research question is about phenomena ofor affecting individual entities across differentsettings (networks) use the ego-centric approach– Individual people, organizations, nations, etc. If your research question is about differentpatterns of interaction within defined groups(networks), use the socio-centric approach– E.g., who are the key players in a group? How do ideasdiffuse through a group? 2008 Halgin & DeJordyAcademy of Management PDWPage 11

Introduction to Ego Network AnalysisWhich Theories are Ego-centric? Most theories under the rubric of socialcapital are ego-centric Topological– Structural holes / Brokerage– Embeddedness Compositional– Size– Alter attributes 2008 Halgin & DeJordyAcademy of Management PDWPage 12

Introduction to Ego Network AnalysisSteps to a SNA study1. Identify the population Sampling, gaining access2. Determine the data sources Surveys, interviews, observations, archival3. Collect the data Instrument design 2008 Halgin & DeJordyAcademy of Management PDWPage 13

Introduction to Ego Network AnalysisStep 1. Identify the Population Sampling Criteria– Determined by research question High tech entrepreneurs Alumni of defunct organizations Basketball coaches First time mothers returning to the workforce Baseball Hall of Fame inductees Contingent workers People with invisible stigmatized identities 2008 Halgin & DeJordyAcademy of Management PDWPage 14

Introduction to Ego Network AnalysisStep 1. Identify the Population Gaining Access– Same concerns as other research It depends on the sensitivity of the questions thatyou are asking Length of interview can be daunting– Depends on the number of alters 2008 Halgin & DeJordyAcademy of Management PDWPage 15

Introduction to Ego Network AnalysisStep 2: Determine Data Sources SurveysInterviewsObservationsArchival data 2008 Halgin & DeJordyAcademy of Management PDWPage 16

Introduction to Ego Network AnalysisWhite House Diary Data, Carter PresidencyYear 1 2008 Halgin & DeJordyData courtesy of Michael LinkAcademy of Management PDWYear 4Page 17

Introduction to Ego Network AnalysisStep 3: Collect the Data What data should you collect?– What questions need to be answered? How to format your data collection instrument(e.g., a survey, spreadsheet, database, etc.)? 2008 Halgin & DeJordyAcademy of Management PDWPage 18

Introduction to Ego Network AnalysisWhat Questions to Ask? IT DEPENDS!!!– Ego’s relations to alters form variables. Size of ego’s socialsupport network is to ego network analysis what “attitudetoward gun-control” is to traditional case based research. It is the researcher who defines the relations ofinterest. What’s relevant for the phenomena inquestion?– What influences an employee’s turn-over intention?– What influences one’s likelihood of adoption of a newtechnology? 2008 Halgin & DeJordyAcademy of Management PDWPage 19

Introduction to Ego Network AnalysisHow to ask: Tick or Rate? Record yes/no decisions or quantitative assessment?– Yes/no are cognitively easier to determine (therefore reliable,believable)– Yes/no *much* faster to administer– But yes/no provides no discrimination among levels One quantitative rating can replace a series of binaries– “How often do you see each person?” 1 once a year; 2 once a month; 3 once a week; etc.– Instead of three questions: Who do you see at least once a year? Who do you see at least once a month? Who do you see at least once a week?– However, if categories are too similar it may be difficult to differentiate 2008 Halgin & DeJordyAcademy of Management PDWPage 20

Introduction to Ego Network AnalysisQuestion Wording Issues “Friendship” does not mean the same thing toeveryone– Especially across national cultures Some helpful practices– Use one word label plus two or three sentencedescription, plus have full paragraph detailedexplanation available– Use homogeneous samples 2008 Halgin & DeJordyAcademy of Management PDWPage 21

Introduction to Ego Network AnalysisEthnographic Sandwich Ethnography at front end helps to – Select the right questions to ask– Word the questions appropriately– Create enough trust to get the questionsanswered Ethnography at the back end helps to – Interpret the results– Can sometimes use respondents as collaborators 2008 Halgin & DeJordyAcademy of Management PDWPage 22

Introduction to Ego Network AnalysisInstrument Design: Paper or Plastic? Paper medium– Reliable– Reassuring to respondents– Errors in data entry– Data entry is time-consuming Electronic–––––Span distances, time zonesHarder to loseFewer data handling errorsLower response rateEmailed documents vs. survey instruments 2008 Halgin & DeJordyAcademy of Management PDWPage 23

Introduction to Ego Network AnalysisData Collection in an Ego-centric Study1. Attributes about Ego2. Name generator Obtain a list of alters3. Name interpreter Assess ego’s relationships with generated list of alters?4. Alter Attributes Collect data on the list of alters5. Alter – Alter Relationships Determine whether the listed alters are connected 2008 Halgin & DeJordyAcademy of Management PDWPage 24

Introduction to Ego Network AnalysisAttributes about Ego Typical variables for case based analysis– Age– Gender– Education– Profession– SES– Etc. 2008 Halgin & DeJordyAcademy of Management PDWPage 25

Introduction to Ego Network AnalysisSample Name Generators Questions that will elicit the names of alters– From time to time, most people discuss important personalmatters with other people. Looking back over the last sixmonths who are the people with whom you discussed animportant personal matter? Please just telI me their first namesor initials.– Consider the people with whom you like to spend your freetime. Over the last six months, who are the one or two peopleyou have been with the most often for informal social activitiessuch as going out to lunch, dinner, drinks, films, visiting oneanother’s homes, and so on?(Burt, 1998) 2008 Halgin & DeJordyAcademy of Management PDWPage 26

Introduction to Ego Network AnalysisSample Name Interpreter Questions that deal with ego’s relationshipwith [or perception of] each alter– How close are you with alter ?– How frequently do you interact with alter ?– How long have you known alter ? All of these questions will be asked for eachalter named in the previous section 2008 Halgin & DeJordyAcademy of Management PDWPage 27

Introduction to Ego Network AnalysisSample Alter Attribute Questions As far as you know, what is alter ’ s highestlevel of education? (Adapted from Burt, 1984)– Age, occupation, race, gender, nationality, salary, drug usehabits, etc Some approaches do not distinguish betweenname interpreters and alter attributequestions 2008 Halgin & DeJordyAcademy of Management PDWPage 28

Introduction to Ego Network AnalysisSample Alter-Alter Relationship Questions Think about the relationship between alter1 and alter2 . Would you say that they arestrangers, just friends, or especially close?(Adapted from Burt, 1998) Note: this question is asked for each unique alteralter pair. E.g., if there are 20 alters, there are190 alter-alter relationship questions!– Typically, we only ask one alter-alter relationshipquestion 2008 Halgin & DeJordyAcademy of Management PDWPage 29

Introduction to Ego Network AnalysisWhy Ego-Centric Analysis Asks different questions than whole networkanalysis. In fact, many of the various approaches to“Social Capital” lend themselves particularlyto the analysis of Ego-Centric or Personalnetworks 2008 Halgin & DeJordyAcademy of Management PDWPage 30

Introduction to Ego Network AnalysisKinds of Analyses In Ego-Centric Network analyses we aretypically looking to use network-derivedmeasures as variables in more traditionalcase-based analyses– E.g., instead of just age, education, and family SESto predict earning potential, we might also includeheterogeneity of network or brokerage statistics– Many different kinds of network measures, thesimplest is degree (size) 2008 Halgin & DeJordyAcademy of Management PDWPage 31

Introduction to Ego Network AnalysisData Analysis of Ego Networks1. Size–How many contacts does Ego have?2. Composition–––What types of resources does ego have access to? (e.g., quality )Does ego interact with others like him/herself? (e.g., homophily)Are ego’s alters all alike? (e.g., homogeneity?)3. Structure––Does ego connect otherwise unconnected alters? (e.g., brokerage,density, etc)Does ego have ties with non-redundant alters (e.g., effective size,efficiency, constraint) 2008 Halgin & DeJordyAcademy of Management PDWPage 32