Data from: Exploring formation processes in political discussion networks

Our data comprises thirty personal networks that belong to people affiliated to a local branch of a newly created and rising star Romanian parliamentary political party. At the moment of the data collection process (February, 2019), the political organization was only four-month old. The study participants were randomly selected from a roster of 645 members (the total tally of people affiliated to a local branch situated in a large Romanian urban area). We used a stratified sampling, with three strata based on age: 18 – 25 years-old (group A, "young"), 26 – 55 years-old (group B, "adult"), and more than 56 years old (group C, "senior"). In each stratum, we performed simple random sampling and selected ten respondents. The interviews were administered by phone (computer-assisted telephone interviewing). A participant was eligible to the study if formally affiliated to the local political branch of interest. The identity of each study participant was anonymized to ensure privacy protection. Within the study population, the minimum age was 19 and the maximum, 72. The three clusters (groups A, B, and C) were useful for controlling the age of the interviewees while fitting the statistical models. Stratified sampling was also deployed to avoid any potential sources of bias (e.g., over-representation of some age categories given the skewed distribution of the study population). Given the affiliation to the same political organization, the study participants were deemed equivalent at least on this dimension. For each of the thirty study participants (egos), we collected data about sex (female or male), age and education (with or without higher education studies). We elicited the alters of each ego using the following name generator: Please, mention five people with whom you discuss most often political issues. We applied alter interpreters (questions about alters) referring to sex (female or male), age, and education (with or without higher education studies). Also, we collected information about the alter-alter ties to build the personal networks. We administered a simple question and asked each respondent whether the alters knew each other (Would you say that alter X and alter Y know each other?). Given the form of the question, we elicited undirected alter-alter ties. We profiled the people with whom egos declared to discuss most often political topics. Namely, we measured: a) the perceived level of political agreement between ego and each of the alters (rated from 1–minimum, to 5–maximum), b) whether ego and alters share membership to the same political party, c) the perceived emotional closeness between each pair of alters and between ego and alters – each ego was asked to rate the emotional closeness from 1 (minimum) to 5 (maximum), d) the status of the alters – family or non-family (friends or acquaintances), and e) the ego –alter personal history - the estimated duration in years of each ego-alter tie.

本数据集包含30份个人社交网络数据，均来自罗马尼亚某新晋崛起的议会政党地方支部的在册成员。数据采集于2019年2月，彼时该政治组织仅成立4个月。
本研究的参与者从罗马尼亚某大型城区的该政党地方支部全部645名在册成员中随机抽取。研究采用分层抽样方法，按年龄划分为三层：18-25岁（A组，青年组）、26-55岁（B组，成年组）以及56岁以上（C组，老年组），每层均通过简单随机抽样选取10名受访者。
访谈采用计算机辅助电话访谈（computer-assisted telephone interviewing）方式开展。参与者需满足正式隶属于目标地方支部的条件方可纳入研究，所有受访者身份均已匿名化以保障隐私安全。本研究人群的年龄范围为19岁至72岁。A、B、C三组分层可在拟合统计模型时有效控制受访者年龄变量，同时分层抽样也可规避潜在偏倚来源——例如因研究人群年龄分布偏斜导致的部分年龄组样本占比过高问题。
由于所有参与者均隶属于同一政治组织，至少在政党隶属维度上可认为研究对象具有同质性。针对30名研究主体（ego，自我中心节点），我们采集了其性别（男/女）、年龄以及受教育程度（是否接受过高等教育）相关数据。
我们通过以下名称生成器获取每个自我中心节点的社交关联对象（alter，他者节点）：“请列出你最常讨论政治议题的5位人物。”随后针对每个他者节点，我们采集了其性别、年龄与受教育程度相关信息。此外，为构建完整的个人社交网络，我们还收集了他者节点间的连接关系数据：我们通过简单问题询问受访者“你认为关联对象X和关联对象Y是否互相认识？”，以此获取他者节点间的无向连接关系。
我们对自我中心节点宣称的常讨论政治议题的关联对象进行了多维度刻画，具体测量指标包括：a）自我中心节点与每个他者节点间感知到的政治契合度（1分表示最低，5分表示最高）；b）自我中心节点与他者节点是否隶属于同一政党；c）自我中心节点与各他者节点间、以及任意两个他者节点间的感知情感亲密度——要求受访者按1（最低）至5（最高）的标尺进行评分；d）他者节点的身份属性：亲属或非亲属（朋友或熟人）；e）自我-他者节点交往时长：估算的自我与他者节点建立连接的年数。