Great! Collecting robust quantitative network data is the foundation for any social network analysis, especially when you're planning to integrate it with qualitative insights in a mixed-methods approach.

Here's a comprehensive guide to collecting network data (quantitative ties/connections), broken down into key steps and considerations:

1. Defining Your Network Data Collection Strategy

Before you collect any data, you need to clearly define what you're looking for. This upfront work is crucial for both the validity of your quantitative data and its eventual integration with qualitative findings.

A. Core Questions to Ask Yourself:

• What is your Research Question?


• Example: "How do advice networks among employees influence innovation within the organization?" or "What role do peer support networks play in student success?" This will dictate the nodes and ties.


• What are your Nodes (Actors/Entities)?


• Individuals: People, students, employees, community members.


• Groups/Organizations: Departments, teams, NGOs, companies.


• Other Entities: Websites, documents, ideas (less common for social networks, but possible).


• Crucial for Mixed Methods: Will these nodes be the same individuals/groups you interview or observe qualitatively?


• What are your Ties (Relationships/Connections)?


• This is the most critical definition for network data. Be specific!


• Type of Relation:


• Cognitive/Perceptual: Who do you consider a friend? Who do you trust?


• Affective/Emotional: Friendship, liking, dislike, trust, support.


• Instrumental/Exchange: Advice, collaboration, resource sharing, information flow, task-related interaction.


• Affiliation/Shared Attribute: Co-membership (e.g., in a club), co-authorship, shared location, shared event attendance.


• Communication: Email, phone calls, social media interactions.


• Directionality:


• Directed: A sends email to B (A -> B).


• Undirected: A and B are friends (A <-> B).


• Weight/Strength:


• Binary: Yes/No, a tie exists or not.


• Weighted: How strong is the tie? (e.g., frequency of interaction on a scale of 1-5, amount of shared resources, number of hours collaborated).


• What are your Network Boundaries?


• Whole Network (Sociocentric): You aim to capture all relationships among all members within a predefined group (e.g., all employees in a department, all students in a classroom).


• Pros: Allows for analysis of overall network structure (density, centrality, cliques).


• Cons: Can be difficult to define and enumerate all members; high response burden for large groups.


• Ego Network (Personal Network): You focus on one central individual (the "ego") and their direct connections (the "alters"), and sometimes the connections among those alters. You collect data from many egos.


• Pros: Easier to collect from large, diverse populations; less response burden per individual.


• Cons: Cannot directly analyze whole network properties; provides a "view" from individual perspectives.


• Crucial for Mixed Methods: Ego network collection often aligns well with qualitative interviews where individuals discuss their personal connections.


• What is your Time Frame?


• Cross-sectional: Data collected at a single point in time (a snapshot).


• Longitudinal: Data collected at multiple points in time to observe changes in the network structure.

B. Ethical Considerations (Crucial for SNA!)

Network data is highly sensitive because it reveals connections between people.

• Informed Consent: Clearly explain what data you're collecting, why, and how it will be used. Be explicit about the fact that responses connect individuals.


• Anonymity vs. Confidentiality:


• Anonymity: Not usually possible in whole network studies, as you need to identify nodes to map ties.


• Confidentiality: Promise to protect individual responses and only report aggregated or anonymized findings. This is usually the standard.


• Data Security: Protect all raw data meticulously.


• "Small World" Problem: Be aware that even aggregated data might allow individuals to infer ties if the network is small enough or if specific qualitative examples are shared.

2. Quantitative Network Data Collection Methods

Once your definitions are solid, choose the most appropriate method(s) for collecting your ties.

Method 1: Surveys/Questionnaires (Most Common)

This is the workhorse for collecting network data, especially for whole and ego networks.

A. Whole Network Surveys

• How it Works: All members of a defined group are asked about their relationships with other members of that same group.


• Common Techniques:


• Roster Recall (Sociometric Method):


• Process: Provide each participant with a comprehensive list (roster) of all other members in the network. For each member on the roster, ask a series of questions about their relationship with that specific individual.


• Example Questions (for each person on the roster):


• "How often do you discuss work-related problems with [Name from Roster]?" (1 = Never, 5 = Daily)


• "Have you collaborated on a project with [Name from Roster] in the last 6 months?" (Yes/No)


• "Do you consider [Name from Roster] a close friend?" (Yes/No)


• Pros: High accuracy (no recall bias for names), ensures all possible ties are considered.


• Cons: Very high response burden for large networks (e.g., 50 people = 49 questions per person for a simple binary tie); participants might not know everyone on the roster.


• Tips: Limit network size (ideally < 100 nodes), chunk questions, use online survey tools (Qualtrics, SurveyMonkey) with matrix-style questions.


• Free Recall (Name Generator/Recognizer for Whole Networks):


• Process: Ask participants to list the names of people within the defined network who fit certain criteria.


• Example Questions:


• "Please list the names of up to 5 colleagues in this department you seek advice from regarding work issues."


• "Who are the people in your team you typically socialize with outside of work?"


• Pros: Less response burden than roster recall.


• Cons: Prone to recall bias (people might forget some ties); limits the number of ties reported (e.g., "up to 5").


• Tips: Combine with a "name interpreter" section (see Ego Networks) to get more info on listed alters.

B. Ego Network Surveys

• How it Works: Each participant (the "ego") identifies their connections (the "alters"), and then provides information about those alters and the relationships between those alters.


• Common Techniques:


• Name Generators:


• Process: Open-ended questions asking the ego to generate a list of significant people in their life based on specific criteria.


• Example Questions:


• "From time to time, most people discuss things that are important to them. Looking back over the last six months, who are the people with whom you have discussed matters important to you? Please list first names or initials." (Classic GSS Name Generator)


• "Who are the people who provide you with emotional support?"


• "Who are the people you go to for professional advice?"


• Pros: Captures personally salient ties; relatively quick.


• Cons: Limited number of alters, potential for recall bias.


• Name Interpreters:


• Process: For each alter generated by the ego, ask follow-up questions about that alter's attributes and their relationship with the ego and other alters.


• Example Questions (for each listed alter):


• "What is [Alter's Name]'s gender? Age? Relationship to you (family, friend, colleague)?" (Alter Attributes)


• "How often do you interact with [Alter's Name]?" (Tie Strength)


• "Does [Alter's Name] know [another Alter's Name]?" (Tie between Alters - Crucial for Ego-Network Density)


• Pros: Builds a richer picture of the ego's personal network, allows for calculating ego-network density and composition.


• Cons: Can still be lengthy depending on the number of alters generated.


• Tips: Use software that allows dynamic follow-up questions based on previous answers (e.g., Qualtrics display logic).

Method 2: Archival / Secondary Data

Using existing records can be a powerful and often less intrusive way to collect network data.

• Examples:


• Co-authorship: Publications, patents, research grants. (Nodes = authors/inventors, Ties = co-authorship).


• Co-membership: Boards of directors, clubs, committees. (Nodes = individuals/organizations, Ties = shared membership).


• Communication Logs: Anonymized email data, phone records, messaging app data (with ethical approval!). (Nodes = communicators, Ties = communication frequency/direction).


• Organizational Charts: Reporting structures (Nodes = employees, Ties = hierarchical reporting).


• Meeting Attendance: Shared participation in events. (Nodes = participants, Ties = co-attendance).


• Transaction Data: Financial flows, resource exchanges. (Nodes = entities, Ties = transactions).


• Pros: Unobtrusive, objective (less recall bias), can be longitudinal, large datasets often available.


• Cons: Data might not reflect the actual social ties or their meaning; limited to recorded interactions; ethical challenges with private data (anonymization, access permissions).


• Tips: Carefully define what a "tie" means in the context of your archival data (e.g., "co-authored one paper" vs. "co-authored three papers").

Method 3: Observational Data

Directly observing interactions can provide valuable network data, especially for smaller groups or specific contexts.

• Examples:


• Workplace Interactions: Tracking who talks to whom during meetings, coffee breaks, or project work.


• Classroom Dynamics: Observing student collaboration or peer teaching.


• Community Events: Who interacts with whom at social gatherings.


• Process: Develop a structured observation protocol. Observers record specific interactions between predefined nodes.


• Pros: Captures actual behavior rather than perceptions; rich contextual information.


• Cons: Labor-intensive, time-consuming, prone to observer bias, ethical concerns (privacy, reactivity of participants knowing they're observed), often limited to smaller networks.


• Tips: Use multiple observers to check inter-rater reliability, clearly define types of interactions to be recorded, use tools like interaction matrices.

Method 4: Digital Trace Data / APIs

The digital world is a rich source of network data, particularly from social media and collaboration platforms.

• Examples:


• Social Media: Followers/following on Twitter/Instagram, likes, comments, retweets, mentions, shared posts (Nodes = users, Ties = interactions).


• Collaboration Platforms: Slack channels, GitHub repositories, Google Drive shared documents, internal communication systems (Nodes = users, Ties = messages, shared edits, project contributions).


• Email Data: "To," "From," "CC" fields (Nodes = email addresses, Ties = communication).


• Process: Access data through platform APIs (Application Programming Interfaces), web scraping (with caution regarding terms of service and legality), or direct export from internal systems.


• Pros: Unobtrusive, captures real-time behavior, massive datasets possible, often longitudinal.


• Cons: Major ethical challenges (privacy, data ownership, informed consent for public data); data "noise" (e.g., a "like" doesn't always imply a strong social tie); platform biases (e.g., what interactions a platform privileges); data access limitations.


• Tips: Always prioritize ethical review and consent. Understand the limitations of what the data actually represents. Data cleaning and preprocessing are extensive.

3. Preparing for Mixed Methods Integration

As you collect your quantitative network data, keep your mixed methods design in mind:

• Node Identification: Ensure your quantitative node identifiers (e.g., employee IDs) can be linked back to potential qualitative participants (e.g., names for interviews), while maintaining confidentiality.


• Sampling for Qualitative Phase: Your quantitative analysis can inform your qualitative sampling. For example:


• Identify central actors (high degree, betweenness centrality) for in-depth interviews.


• Select individuals from different network positions (e.g., isolates, members of dense cliques, bridge-builders) to understand diverse experiences.


• Choose specific tie types that are prominent or surprisingly absent for qualitative exploration.


• Defining Tie Meaning: Quantitative data tells you if a tie exists and how strong it is. Qualitative methods will help you understand the meaning and context behind those ties.


• Example: A high frequency of email exchanges (quantitative) might mean collaboration, but qualitative interviews can reveal if it's productive collaboration, conflict resolution, or simply information dissemination.

By systematically working through these steps, you'll be well-prepared to collect robust quantitative network data that forms a strong basis for your mixed-methods analysis. Remember to choose methods that align with your research question, ethical guidelines, and practical constraints.