Assessing the bias in samples of large online networks

By Sandra González-Bailón, Ning Wang, Alejandro Rivero, Javier Borge-Holthoefer, Yamir Moreno. 2014

249 citations recorded by the source corpus.

Subfields: Sociology and Political Science, Statistical and Nonlinear Physics, Statistics and Probability, Network Science & Complex Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Sociology and Political Science

In Statistical and Nonlinear Physics

In Statistics and Probability

In Network Science & Complex Systems

Citations in the corpus

Debiasing Community Detection: The Importance of Lowly Connected Nodes (2019, neutral): “network measures and create unrealistic low-degree users [15],”
Sampling Content from Online Social Networks (2015, background): “Additionally, the earlier studies focused only on tweets related to a speciﬁc event (political protests in Spain in mid-2012 [Gonzalez-Bailon et al. 2014], and the Syrian conﬂict during 2011–2012 [Morstatter et al. 2013]), whereas we study various aspects of the sampled tweet streams to derive far…”
Examining a hate speech corpus for hate speech detection and popularity prediction (2018, methodology): “ent in user-generated content that is not under the control of the researcher. Social media data is typically collected by public APIs that may lead to inconsistent results. For example, Gonz´alezBailon et al. (2014) ﬁnd that the Twitter Search API´ yields a smaller dataset than the Stream API when ”
Fairness in Machine Learning: A Survey (2020, background): “…problem, i.e., intrinsic biases in the sample will manifest themselves in any model built on the data [33, 42], inappropriate uses of data leading to (un)conscious bias(es) [41, 42], data veracity and quality [320], data relativity and context shifts [42, 122, 249], and subjectivity filters [41].”
When Are Search Completion Suggestions Problematic? (2020, background): “This also means that our samples are shaped by the limitations to keywords-based data collection [27, 57, 79].”
Testing Propositions Derived from Twitter Studies: Generalization and Replication in Computational Social Science (2015, background): “Most of the Twitter study samples are collected by Twitter’s streaming API but this sampling method is nontransparent to researchers and the procedures might be biased toward users making large amounts of posts [9].”
Sociology in the Era of Big Data: The Ascent of Forensic Social Science (2015, unknown): “For example, González-Bailón et al. (2014) consider the sampling bias introduced by collecting data through publicly available APIs.”
Surpassing the Limit: Keyword Clustering to Improve Twitter Sample Coverage (2015, background): “A number of studies have shown that the method used for sampling can introduce various forms of bias which introduce error into results and remain largly unnoticed [9, 19].”