Propitious Aggregation: Reducing Participant Burden in Ego-centric Network Data Collection Derek Lackaff Department of Radio-Television-Film University of Texas at Austin lackaff@mail.utexas.edu ABSTRACT One of the central challenges of ego-centric or personal social network research is minimizing the quantity of data that is requested from research participants while ensuring high data accuracy and validity. In general, collecting data about increasingly larger ego-centric networks places an increasing burden on respondents. The web-based Propitious Aggregation of Social Networks (PASN, http://pro.pitio.us) survey instrument reduces this burden by leveraging network data already available in the context of social network websites, and by providing an intuitive click-and-drag interface for survey responses. An experiment was conducted (N = 85), and the PASN method was found to produce networks which were significantly larger and more diverse than those produced using standard survey methods, yet required significantly lower time investments from participants. Author Keywords Social networks, social network sites, computer-assisted self interviews, methods, user interfaces. ACM Classification Keywords H.5.2 Information interfaces and presentation: Theory and methods; J.4 Social and behavioral sciences: Sociology. General Terms Design, Experimentation INTRODUCTION Researchers who study ego-centric (or personal) networks attempt to understand how the structure of an individual's social relationships results in meaningful outcomes. Personal networks are comprised of an ego, or focal individual, alters, or social contacts of ego, and the relationships that exist among these actors, which are termed edges. In general, collecting data about increasingly larger ego-centric networks places an increasing burden on respondents. Simply remembering the people in one's network can be surprisingly difficult [3], and providing detailed information about these alters and their relationships can be an overwhelming task to request of respondents [6]. Various survey-based “network generators” have been used in previous research to enable respondents to at least estimate the characteristics of their overall personal networks. This paper presents a method of personal network generation and interpretation that leverages previously-articulated social network data. While research respondents are generally able to recall their more intimate relationships with a reasonable level of accuracy, less intimate relationships are more likely to be forgotten. Findings about informant accuracy have been mixed, and many researchers believe that informant accuracy is a topic of methodological interest [2, 4]. One study found that respondents were unable to recall 3% of best friends, 9% of close friends, and 20% of other friends [3], while another study found that 26% of close friends were forgotten [1]. Participant forgetting also has effects on structural measures of the networks [3]. Researchers generally use instruments called name generators to aid respondents in remembering and describing their networks. Name generators consist of a series of questions that elicit named alters as responses. Traditional name generators first use a general question, e.g. “Who are your closest friends?” then ask for relevant details (relationship, reasons for closeness, communication frequency, etc.) about each name that is provided [5]. Participant burden is a central concern in the design of personal network studies [10]. While a survey or interview may elicit a short list of important alters with just a few minutes of effort, larger networks can involve significantly more time. Data collection can also be repetitive and monotonous for the respondent – after remembering a number of family, friends, and acquaintances, the respondent is generally asked to interpret each of these names in some way relevant to the researcher's project. But perhaps the most intractable burden that is placed on participants is the evaluation of alter-alter relationships in their network: Does alter A know alter B? Does A know C? and so on. The geometric increase of potential alter-alter ties to be evaluated as networks increase in size can quickly become a crushing burden upon participants. A network of Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. CHI 2010, April 10–15, 2010, Atlanta, Georgia, USA. Copyright 2010 ACM 978-1-60558-928-9/10/04...$10.00. CHI 2010: Humans and Sociability April 10–15, 2010, Atlanta, GA, USA 1467