Early insight into social network structure predicts climbing the social ladder
Data files
Jun 11, 2025 version files 14.87 MB
-
network_knowledge_task_data.csv
14.86 MB
-
README.md
5.94 KB
Abstract
While occupying an influential position within one’s social network brings many advantages, it is unknown how certain individuals rise in social prominence. Leveraging a longitudinal dataset that tracks an entirely new network of college freshmen (N=187), we test whether ‘climbing the social ladder’ depends on knowing how other people are connected to each other. Those who ultimately come to occupy the most influential positions exhibit early and accurate representations of their network’s general, abstract structure (i.e., who belongs to which communities and cliques). In contrast, detailed, granular representations of specific friendships do not translate into gains in social influence over time. Only once the network stabilizes, do the most influential individuals exhibit the most accurate representations of specific friendships. These findings reveal that those who climb the social ladder first detect their emerging network’s general structure, then fine-tune their knowledge about individual relationships between their peers as network dynamics settle.
https://doi.org/10.5061/dryad.2280gb63t
Description of the data and file structure
These data were collected to study how people's standing in their social networks (i.e., their network centrality) longitudinally depends on their knowledge about the social network's structure (i.e., knowing who is connected to whom, or what communities exist in the network). In this year-long study, we took six measurements of an evolving network of first-year undergraduates by asking every subject to rate their friendship status with every other subject. From these data, we computed two measures of network centrality (eigenvector centrality, which we also refer to as 'influence', and degree centrality, which we also refer to as 'friend count'). We focus on two relevant network measurements, one early timepoint in the Fall semester of the academic year, and one later timepoint in the Spring semester of the academic year. Immediately following these two network measurements, we additionally measured subjects' knowledge about their network by asking them to report (or guess) who is friends with whom among their peers. From these data, we calculated the extent to which their friendship inferences reflect knowledge of the true pairwise friendships reported by their peers (which we refer to as 'micro-level' knowledge) and the broader communities or clusters that comprise the network (which we refer to as 'meso-level' knowledge). We then link micro- and meso-level knowledge to changes in subjects' network centrality over time.
Files and variables
File: network_knowledge_task_data.csv
Description: This file includes trial-by-trial data from the Network Knowledge Task, in which subjects reported (or guessed) whether pairs of their peers were friends. In this task, subjects completed 870 trials, providing two separate ratings of every possible pairwise relationship between 30 of their peers (435 possible relationships). Additionally, this file includes subject-level measures of network centrality (i.e., influence and friend count) and extroversion.
Variables
- semester: Indicates which semester the task data was collected, with 'fall' indicating the first timepoint mid-Fall semester, and 'spring' indicating the second timepoint several months later mid-Spring semester.
- sub_id: Unique identifier for the subject providing task responses.
- from_id: Identifier for the first individual in the pair of peers whose friendship status the subject rated on a given trial. Note that the directionality of from_id and to_id is somewhat arbitrary, as subjects were rating their inferences about a pairwise friendship between the two individuals. Also note that subjects provided two ratings (i.e., two trials) for each distinct pair of peers. Finally, due to concern about the potential identifiability of social network data, note that identifiers in the from_id and to_id columns are consistent within a given subject's task data for a given semester, but are randomized across subjects and semesters. For example, a particular individual might be identified as '1' in the sub_id column, but if another subject provided ratings about the friendship status of sub_id '1' with other peers, then the subject with sub_id '1' might instead be identified with '2' in the from_id and to_id columns associated with another subject's task data.
- to_id: Identifier for the second individual in the pair of peers whose friendship status the subject rated on a given trial. See notes about the from_id variable for more information.
- friend_guess: Binary variable indicating whether the subject responded that the pair of individuals (identified by the from_id and to_id columns) was friends (1) or not friends (0).
- micro_friend: Binary variable indicating whether the pair of individuals (identified by the from_id and to_id columns) actually reported being friends (1) or not being friends (0).
- meso_comm: Binary variable indicating whether the pair of individuals (identified by the from_id and to_id columns) were assigned to the same community by cluster detection algorithms (1), or were not assigned to the same community (0).
- sub_influence_fall: Eigenvector centrality (which we refer to as 'influence'), calculated from the first network measurement in the Fall.
- sub_influence_spring: Eigenvector centrality (which we refer to as 'influence'), calculated from the second network measurement in the Spring.
- sub_friend_count_fall: Degree centrality (which we refer to as 'friend count'), calculated from the first network measurement in the Fall.
- sub_friend_count_spring: Degree centrality (which we refer to as 'friend count'), calculated from the second network measurement in the Spring.
- sub_extroversion: Extroversion, calculated from subjects' responses to the extroversion-related items in the Big Five personality inventory.
- sub_influence_delta: Change over time in eigenvector centrality (which we refer to as 'influence'), which we calculate by subtracting Fall influence from Spring influence.
- sub_friend_count_delta: Change over time in degree centrality (which we refer to as 'friend count'), which we calculate by subtracting Fall friend count from Spring friend count.
Code/software
Included are an R Markdown script and the knit output, which contain the code to calculate subjects' network knowledge and conduct analyses linking their network knowledge to their network centrality over time. Information about packages and versions used for the original analyses can be found in the 'Reproducibility' section at the end of the knit output (.html) document.
Access information
Other publicly accessible locations of the data: https://osf.io/7z3ya/files/osfstorage?view_only=cece86f03fc9481dab58356159c24950