Fightin' Words: Lexical Feature Selection and Evaluation for Identifying the Content of Political Conflict

By Burt L. Monroe, Michael Colaresi, Kevin M. Quinn. 2008

667 citations recorded by the source corpus.

Subfields: General Social Sciences, Political Science and International Relations, Sociology and Political Science, Natural Language Processing

How researchers see this paper

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

In General Social Sciences

In Political Science and International Relations

In Sociology and Political Science

In Natural Language Processing

Citations in the corpus

Social Biases in NLP Models as Barriers for Persons with Disabilities (2020, methodology): “Using this 4-way balanced dataset, we calculated the log-odds ratio metric (Monroe et al., 2008) for all unigrams and bi-grams (no stopword removal) that measure how over-represented they are in the group of comments that have a disability mention, while controlling for co-occurrences due to chance.”
Demographic Dialectal Variation in Social Media: A Case Study of African-American English (2016, methodology): “This model has broadly similar goals as nonlatent, log-linear generative models of text that condition on document-level covariates (Monroe et al., 2008; Eisenstein et al., 2011a; Taddy, 2013).”
Style Transfer Through Back-Translation (2018, methodology): “Style lexicons are extracted using the log-odds ratio informative Dirichlet prior (Monroe et al., 2008), a method that identifies words that are statistically overrepresented in each of the categories.”
Latent Hatred: A Benchmark for Understanding Implicit Hate Speech (2021, methodology): “Then we computed the log odds ratio with informative Dirichlet prior (Monroe et al., 2008) for each noun, hashtag, and adjective to identify the top 25 words per ideology.”
Towards Emotional Support Dialog Systems (2021, methodology): “Each feature is ranked by the rounded z-scored log odds ratios (Monroe et al., 2008) in the parentheses.”
Native Language Cognate Effects on Second Language Lexical Choice (2018, background): “A common way to identify elements that are statistically over-represented in a particular population, compared to another, is log-odds ratio informative Dirichlet prior (Monroe et al., 2008).”
Something’s Brewing! Early Prediction of Controversy-causing Posts from Discussion Features (2019, methodology): “We utilize Monroe et al.’s (2008) algorithm for comparing language usage in two bodies of text; the method places a Dirichlet prior over n-grams (n=1,2,3) and estimates Z-scores on the difference in rate-usage between controversial and non-controversial posts.”
Predicting Anti-Asian Hateful Users on Twitter during COVID-19 (2021, methodology): “As an exploratory analysis, we identify the representative words of user groups by using the logodds ratio proposed in (Monroe et al., 2008).”
Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models (2023, methodology): “We use the Fightin’ Words method of Monroe et al. (2008) with the informative Dirichlet prior, first computing the weighted log-odds ratios of the words between Ps and corresponding sets of texts that represent each unmarked identity, using the other texts in the dataset as the prior distribution,…”
Detecting Cross-Geographic Biases in Toxicity Modeling on Social Media (2021, methodology): “Since we now have a multi-class setting, one could repeat the (Monroe et al., 2008) method in a one-vs-rest manner.”
Text as Data: The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts (2013, background): “…than a count (Pang, Lee, and Vaithyanathan 2002; Hopkins and King 2010); (2) including some commonly used stopwords such as gendered pronouns (Monroe, Colaresi, and Quinn 2008); (3) a subset of features (either by automated feature selection or a lower dimensional projection) (Hofmann 1999);…”
More Voices Than Ever? Quantifying Media Bias in Networks (2011, background): “Because of such a dependency, computationally identifying bias from media content remains an emerging research topic, and requires insights from other language analysis studies such as sentiment analysis [9] or partisan features in texts [10, 8].”
Topics to Avoid: Demoting Latent Confounds in Text Classification (2019, methodology): “(We refer the reader to (Monroe et al., 2008) for the details about the algorithm.”