The DOI matches its Crossref record; 71 of the 79 references it cites were found in a registry or an index; 3 of the 14 sources its claims cite were checked for support.
The Effects of Twitter Sentiment on Stock Price Returns
Open accessAbstract
Social media are increasingly reflecting and influencing behavior of other complex systems. In this paper we investigate the relations between a well-known micro-blogging platform Twitter and financial markets. In particular, we consider, in a period of 15 months, the Twitter volume and sentiment about the 30 stock companies that form the Dow Jones Industrial Average (DJIA) index. We find a relatively low Pearson correlation and Granger causality between the corresponding time series over the entire time period. However, we find a significant dependence between the Twitter sentiment and abnormal returns during the peaks of Twitter volume. This is valid not only for the expected Twitter volume peaks (e.g., quarterly announcements), but also for peaks corresponding to less obvious events. We formalize the procedure by adapting the well-known "event study" from economics and finance to the analysis of Twitter data. The procedure allows to automatically identify events as Twitter volume peaks, to compute the prevailing sentiment (positive or negative) expressed in tweets at these peaks, and finally to apply the "event study" methodology to relate them to stock returns. We show that sentiment polarity of Twitter peaks implies the direction of cumulative abnormal returns. The amount of cumulative abnormal returns is relatively low (about 1-2%), but the dependence is statistically significant for several days after the events.
Overview
CoChat Summary
TL;DR
The paper finds that across the full 15-month period, Twitter sentiment and stock returns are only weakly related by standard correlation/Granger tests, but sentiment becomes informative around detected Twitter-volume peaks. Using an event-study design on 30 DJIA stocks and over 1.5 million tweets, the authors report statistically significant cumulative abnormal returns that follow the polarity of Twitter sentiment.
Key findings
- On the full sample, Pearson correlations between Twitter sentiment and stock-price variables are “small,” and only 3 companies pass the Granger test for sentiment polarity predicting price return.
- The volume of tweets Granger-causes absolute price return for one third of the companies, suggesting Twitter attention is informative about volatility.
- The event study detects 260 events in total; 118 of 151 EA events are detected, giving 78% recall for earnings-announcement events.
- For positive sentiment events, cumulative abnormal returns are significant at the 1% level for 10 days after the event; for negative sentiment events, the absolute CAR is twice as large. For non-EA events, CAR remains significant at the 1% level for 4 days after positive events and 8 days after negative events.
Methods
- Studied 30 DJIA stocks over 15 months (June 1, 2013 to September 18, 2014) and collected over 1.5 million tweets using Twitter Search API cash-tags.
- Built a supervised sentiment classifier from 103,262 annotated tweets labeled by 10 financial experts; 6,143 tweets were double-annotated; classification used an SVM with bag-of-words features (tokenization, lemmatization, unigrams/bigrams, TF-IDF).
- Analyzed relationships using Pearson correlation, Granger causality, and an event study with Twitter peaks detected by a sliding-window median rule; event impact was measured with cumulative abnormal returns (CAR) under a market model using the DJIA index and a 120-day estimation window.
Limitations
- The paper itself notes that when sentiment and returns are measured over the entire time period, the correlation is low and only a few companies pass Granger causality, implying limited predictive power outside event windows.
- The event-polarity thresholds are acknowledged as somewhat arbitrary; the authors explicitly say that improving the baseline event-class distribution “remains a subject of further research.”
- Evident from the text: the study focuses only on DJIA 30 stocks and one 15-month period, so broader generalization is not established in the paper.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures





Verification
Claims the paper makes
1 Valid · 1 Not valid · 8 Not checked · 46 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“The computed coefficients are small, but are in line with the result of [30].”
§ The Effects of Twitter Sentiment on Stock Price Returns > Results > Correlation and Granger causality
What the sources say
- Disputedcites [30] Mao H, Counts S, Bollen J. Predicting financial markets: Comparing survey, news, twitter and search engine data. arXiv preprint. 2011; <comment>Available from: http://arxiv.org/abs/1112.1051</comment> sourceExperimental — accuracy not yet measured
The cited paper describes its weekly search-volume Pearson correlations as “relatively strong correlations in most cases” (Search Volume and Financial Indicator Correlations, p. 4). Table III on the same page reports coefficients as high as 0.88 in magnitude, contrary to the claim that its computed correlations are small. (LLM, conf=0.87)
Checked by a reasoning model — may be wrong.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Disputed
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The classifier has reached the annotator agreement in all three measures.”
§ The Effects of Twitter Sentiment on Stock Price Returns > Results > Twitter sentiment classification
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The amount of cumulative abnormal returns is relatively low (about 1–2%), but the dependence is statistically significant for several days after the events.”
§ The Effects of Twitter Sentiment on Stock Price Returns > Abstract
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, there are also other peaks in the Twitter activity, which do not correspond to EA, abbreviated as non-EA events.”
§ The Effects of Twitter Sentiment on Stock Price Returns > Methods > Event study
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“From our data we detected 260 events.”
- Also stated in: § Event study · plain
“The total number of peaks that our procedure detects in the period of the study is 260.”
- Also stated in: § Event study · plain
“In total, 260 events are detected, 118 out of these are the EA events.”
§ The Effects of Twitter Sentiment on Stock Price Returns > Methods > Event study
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § Event study · plain
26 more claims
Create a free account to see all 56 claims and every source verdict.
- It may state two separate claims as one. 16
- Probably not a substantial claim. 8
- The quote only partly states this claim. 2
Citations
Cites
The Nature of Statistical Learning Theory
Vapnik, V. N. (1995) · doi:10.1007/978-1-4757-2440-0 · Cited by 39,138
Data Mining: Practical Machine Learning Tools and Techniques
(2011) · Elsevier eBooks · doi:10.1016/c2009-0-19715-5 · Cited by 25,437
Investigating causal relations by econometric models and cross-spectral methods
Granger, C. W. J. (1969) · Econometrica · doi:10.2307/1912791 · Cited by 22,989
EFFICIENT CAPITAL MARKETS: A REVIEW OF THEORY AND EMPIRICAL WORK*
Malkiel, B. G., & Fama, E. F. (1970) · The Journal of Finance · doi:10.1111/j.1540-6261.1970.tb00518.x · Cited by 15,699 · cited 2×
“tions must be aggregated in order to draw overall conclusions for the events of interest. The aggregation is along two dimensions: through time and across stocks. By aggregating across all the stocks [52], we get: ARτ = (1/N) XN i=1 ARi,τ . (10) The cumulative abnormal return (CAR) from time τ1 to τ2 is the sum of the…”
Opinion Mining and Sentiment Analysis
Pang, B., & Lee, L. (2008) · now publishers, Inc. eBooks · doi:10.1561/9781601981516 · Cited by 6,720
Opinion mining and sentiment analysis
Pang, B., & Lee, L. (2008) · Foundations and Trends® in Information Retrieval · doi:10.1561/1500000011 · Cited by 6,372 · cited 1×
“crossvalidation and compared to the inter-annotatoragreement, and (iv) the classifier is applied to the whole set of collected tweets. In this paper, as is common in the sentiment analysis literature [44], we have approximated the sentiment of tweets with an ordinal scale of three values: negative (−), neutral (0), an…”
MethodThe Econometrics of Financial Markets
Campbell, J. Y., Lo, A. W., & MacKinlay, A. C. (2012) · Princeton University Press eBooks · doi:10.2307/j.ctt7skm5 · Cited by 5,875
Twitter mood predicts the stock market
Bollen, J., Mao, H., & Zeng, X. (2011) · Journal of Computational Science · doi:10.1016/j.jocs.2010.12.007 · Cited by 5,206
The Econometrics of Financial Markets
Campbell, J. Y., Lo, A. W., & MacKinlay, A. C. (1997) · doi:10.1515/9781400830213 · Cited by 5,008
Giving content to investor sentiment: The role of media in the stock market
TETLOCK, P. C. (2007) · The Journal of Finance · doi:10.1111/j.1540-6261.2007.01232.x · Cited by 4,827 · cited 1×
“They study: (i) the connection of exogenous news with price movements [7], (ii) the stock price reaction to news [8, 9]; (iii) the relations between mentions of a company in financial news [10], or the pessimism of the media [11], and trading volume; (iv) the relation between the sentiment of news, earnings and return…”
Research works printed but not linked: 2
- Saif H, Fernandez M, He Y, Alani H. Evaluation datasets for twitter sentiment analysis. In: Proc. 1st Workshop on Emotion and Sentiment in Social and Expressive Media. Turin, Italy; 2013.
- Smailović J. Sentiment analysis in streams of microblogging posts. PhD Thesis, Jozef Stefan International Postgraduate School. Ljubljana, Slovenia; 2015.
Cited by
Kralj Novak, P., Smailović, J., Sluban, B., & Mozetič, I. (2015) · PLOS ONE · doi:10.1371/journal.pone.0144296 · Cited by 805
“An object of Twitter posts to which we attribute sentiment (an emoji in our case, but it can also be a stock [37], a political party [35], a discussion topic [26,36], etc.) occurs in several tweets.”
BackgroundMethodSelf-citationIntraday online investor sentiment and return patterns in the U.S. stock market
Renault, T. (2017) · Journal of Banking & Finance · doi:10.1016/j.jbankfin.2017.07.002 · Cited by 439
Sentiment analysis of Twitter data during critical events through Bayesian networks classifiers
Ruz, G. A., Henríquez, P. A., & Mascareño, A. (2020) · Future Generation Computer Systems · doi:10.1016/j.future.2020.01.005 · Cited by 332
Twitter and Research: A Systematic Literature Review Through Text Mining
Karami, A., Lundy, M., Webb, F., & Dwivedi, Y. K. (2020) · IEEE Access · doi:10.1109/access.2020.2983656 · Cited by 278
“…as a useful information resource for financial market activity [176], exploring the relationship between the Twitter daily happiness trend and the stock market trends [177], and examining the impact of positive, negative, and neutral tweets on price returns [178] and renewable energy stocks [179].”
BackgroundRouf, N., Malik, M. B., Arif, T., Sharma, S., Singh, S., Aich, S., & Kim, H.-C. (2021) · Electronics · doi:10.3390/electronics10212717 · Cited by 273
Metadata
- Authors
- Ranco, G., Aleksovski, D., Caldarelli, G., Grčar, M., & Mozetič, I.
- Affiliations
- IMT School for Advanced Studies Lucca (IT)
- Jožef Stefan Institute (SI)
- Corresponding author
- Aleksovski, Darko
- Venue
- PLoS ONE
- Publisher
- Public Library of Science (PLoS)
- Published
- 2015-09-21
- Volume
- 10
- Issue
- 9
- Pages
- e0138441
- ISSN
- 1932-6203 (ISSN-L)
- PMC
- PMC4577113
- OpenAlex
- W1947985544
- Cited by
- 395
- References
- 79
- Citations by year
- 2026: 26 · 2025: 36 · 2024: 42 · 2023: 33 · 2022: 43 · 2021: 41 · 2020: 54 · 2019: 45 · 2018: 29 · 2017: 24 · 2016: 19 · 2015: 3
- FWCI
- 47.78
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLoS ONE (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DiRROS repository (University of Maribor) (publishedVersion, other-oa)
- IMT Institutional Repository (IMT School for Advanced Studies Lucca) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- ARCA (Università Ca' Foscari Venezia) (submittedVersion, cc-by-nc-nd)
- arXiv (Cornell University)
- Funders
- Javna Agencija za Raziskovalno Dejavnost RS (P2-103)
- Topics
- Social Sciences › Economics, Econometrics and Finance › Finance › Financial Markets and Investment Strategies
- Social Sciences › Decision Sciences › Management Science and Operations Research › Stock Market Forecasting Methods
- Physical Sciences › Computer Science › Artificial Intelligence › Sentiment Analysis and Opinion Mining
- Keywords
- Twitter sentiment, stock returns, social media and financial markets, Dow Jones Industrial Average, abnormal returns, event study, sentiment polarity, Granger causality, microblogging
- MeSH terms
- Commerce, Financial Management, Humans, Social Support, Internet, Blogging, Social Media
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: The Effects of Twitter Sentiment on Stock Price Returns. https://app.cochat.ai/papers/10.1371/journal.pone.0138441
- The paper
- Ranco, G., Aleksovski, D., Caldarelli, G., Grčar, M., & Mozetič, I. (2015). The Effects of Twitter Sentiment on Stock Price Returns. PLoS ONE. https://doi.org/10.1371/journal.pone.0138441
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.