Addressing Age-Related Bias in Sentiment Analysis

By Mark Diaz, Isaac L. Johnson, Amanda Lazar, Anne Marie Piper, Darren Gergle. 2018

227 citations recorded by the source corpus.

Computational approaches to text analysis are useful in understanding aspects of online interaction, such as opinions and subjectivity in text. Yet, recent studies have identified various forms of bias in language-based models, raising concerns about the risk of propagating social biases against certain groups based on sociodemographic factors (e.g., gender, race, geography). In this study, we contribute a systematic examination of the application of language models to study discourse on aging. We analyze the treatment of age-related terms across 15 sentiment analysis models and 10 widely-used GloVe word embeddings and attempt to alleviate bias through a method of processing model training data. Our results demonstrate that significant age bias is encoded in the outputs of many sentiment analysis algorithms and word embeddings. We discuss the models' characteristics in relation to output bias and how these models might be best incorporated into research.

Subfields: Artificial Intelligence, Social Psychology, Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

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

In Artificial Intelligence

In Social Psychology

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

Language (Technology) is Power: A Critical Survey of “Bias” in NLP (2020, background): “…al. (2016b); Hovy and Spruit (2016); Blodgett and O’Connor (2017); Larson (2017); Schnoebelen (2017); Blodgett et al. (2018); Curry and Rieser (2018); Díaz et al. (2018); Dixon et al. (2018); Kiritchenko and Mohammad (2018); Park et al. (2018); Shen et al. (2018); Thelwall (2018); Zhao et al.…”
Re-contextualizing Fairness in NLP: The Case of India (2022, background): “, 2020), age (Diaz et al., 2018), and sexual orientation (Garg et al.”
Debiasing Community Detection: The Importance of Lowly Connected Nodes (2019, background): “There is a body of work surrounding different types of biases in social networks, such as the retweet network [7, 13].”
Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need? (2018, background): “Other work has focused on auditing widely-used ML products from the outside [8, 19, 26, 58], often coupled with high-level calls to action aimed at those responsible for developing and maintaining these systems or for regulating their use [33, 86, 88].”
On Releasing Annotator-Level Labels and Information in Datasets (2021, methodology): “, 2020), sentiment (Díaz et al., 2018), and emotion (Demszky et al.”
Fairness in Machine Learning: A Survey (2020, unknown): “[82]), dictionary modification (e.”
Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations (2021, background): “…moral values, and lived experiences often inﬂuence their interpretations, especially in subjective tasks such as identifying political stances (Luo et al., 2020), sentiment (Díaz et al., 2018), and online abuse and hate speech (Cowan and Khatchadourian, 2003; Waseem, 2016; Patton et al., 2019).”
Debiasing Word Embeddings from Sentiment Associations in Names (2020, background): “This type of intervention is done either as a post-processing step [3], pre-processing step [5], or directly during training [31].”
A Study of Implicit Bias in Pretrained Language Models against People with Disabilities (2022, methodology): “We base our work on Díaz et al. (2018) who studied discourse on ageing in sentiment models.”
Mitigation of Unintended Biases against Non-Native English Texts in Sentiment Analysis (2019, methodology): “Others have also observed that gender can signiﬁcantly aﬀect results [7,8,15].”
Biases in Large Language Models: Origins, Inventory, and Discussion (2023, background): “Age bias has been found in sentiment analysis tools, which tended to consider young more positive than old, mostly due to the data they were trained on [47].”
How True is GPT-2? An Empirical Analysis of Intersectional Occupational Biases (2021, background): “There have been many attempts to identify, quantify, and de-bias context-independent word embeddings such as word2vec and GloVe (Bolukbasi et al., 2016; Zhao et al., 2019; Diaz et al., 2018).”
On Measures of Biases and Harms in NLP (2021, background): “However, this task has been demonstrated to be stereotypically influenced by demographic characteristics such as race and gender (Kiritchenko and Mohammad, 2018), age (Díaz et al., 2018) and names of individuals (Prabhakaran et al., 2019).”
“You Are Grounded!”: Latent Name Artifacts in Pre-trained Language Models (2020, background): “Bias can perpetuate to downstream tasks such as coreference resolution (Webster et al., 2018; Rudinger et al., 2018; Zhao et al., 2018), natural language inference (Rudinger et al., 2017), machine translation (Stanovsky et al., 2019), and sentiment analysis (Dı́az et al., 2018).”
Multilingual Twitter Corpus and Baselines for Evaluating Demographic Bias in Hate Speech Recognition (2020, background): “While existing work (Davidson et al., 2019; Diaz et al., 2018) infers user demographic information (white/black, young/old) from the text, such inference is still likely to cause confound- The work was partially done when the first author worked as an intern at Adobe Research.”
Equity Beyond Bias in Language Technologies for Education (2019, background): “…algorithms are known to produce discriminatory results, with disparate outcomes by gender (Volkova et al., 2013), race (Kiritchenko and Mohammad, 2018), and age (D´ıaz et al., 2018), perpetuating a quantiﬁable trend of disproportionate surveillance impact for people of color (Voigt et al., 2017).”