You Are What You Read: Inferring Personality From Consumed Textual Content

Adam  Sutton; Almog Simchon; Matthew  Edwards; Stephan Lewandowsky

You Are What You Read: Inferring Personality From Consumed Textual Content

Adam Sutton^*
, Almog Simchon
, Matthew Edwards
, Stephan Lewandowsky

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

Abstract

In this work we use consumed text to infer Big-5 personality inventories using data we have collected from the social media platform Reddit. We test our model on two datasets, sampled from participants who consumed either fiction content (N = 913) or news content (N = 213). We show that state-of-the-art models from a similar task using authored text do not translate well to this task, with average correlations of r=.06 between the model's predictions and ground-truth personality inventory dimensions. We propose an alternate method of generating average personality labels for each piece of text consumed, under which our model achieves correlations as high as r=.25 when predicting personality from the text being read.

Original language	English
Title of host publication	Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis
Publisher	Association for Computational Linguistics
Pages	28-38
Number of pages	11
ISBN (Electronic)	978-1-959429-87-6
Publication status	Published - 14 Jul 2023
Event	The 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis - WASSA, Toronto, Canada Duration: 14 Jul 2023 → … https://wassa-workshop.github.io/

Workshop

Workshop	The 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis
Abbreviated title	WASSA 2023
Country/Territory	Canada
City	Toronto
Period	14/07/23 → …
Internet address	https://wassa-workshop.github.io/

Research Groups and Themes

Cyber Security
TeDCog

Access to Document

https://aclanthology.org/2023.wassa-1.4Licence: CC BY

Handle.net

Persistent link

Protecting the Democratic Information Space in Europe
Lewandowsky, S. (Principal Investigator), Westaway, R. M. (Administrator) & Carrella, F. (Researcher)
1/10/21 → 30/09/26
Project: Research, Parent

Cite this

@inproceedings{fa8947e8c41140eda38cb066c1674837,

title = "You Are What You Read: Inferring Personality From Consumed Textual Content",

abstract = "In this work we use consumed text to infer Big-5 personality inventories using data we have collected from the social media platform Reddit. We test our model on two datasets, sampled from participants who consumed either fiction content (N = 913) or news content (N = 213). We show that state-of-the-art models from a similar task using authored text do not translate well to this task, with average correlations of r=.06 between the model's predictions and ground-truth personality inventory dimensions. We propose an alternate method of generating average personality labels for each piece of text consumed, under which our model achieves correlations as high as r=.25 when predicting personality from the text being read.",

author = "Adam Sutton and Almog Simchon and Matthew Edwards and Stephan Lewandowsky",

year = "2023",

month = jul,

day = "14",

language = "English",

pages = "28--38",

booktitle = "Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, \& Social Media Analysis",

publisher = "Association for Computational Linguistics",

address = "United States",

note = "The 13th Workshop on Computational Approaches to<br/>Subjectivity, Sentiment, \& Social Media Analysis<br/>, WASSA 2023 ; Conference date: 14-07-2023",

url = "https://wassa-workshop.github.io/",

}

Sutton, A, Simchon, A, Edwards, M & Lewandowsky, S 2023, You Are What You Read: Inferring Personality From Consumed Textual Content. in Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis. Association for Computational Linguistics, pp. 28-38, The 13th Workshop on Computational Approaches to
Subjectivity, Sentiment, & Social Media Analysis
, Toronto, Canada, 14/07/23. <https://aclanthology.org/2023.wassa-1.4>

You Are What You Read: Inferring Personality From Consumed Textual Content. / Sutton, Adam ; Simchon, Almog; Edwards, Matthew et al.
Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis. Association for Computational Linguistics, 2023. p. 28-38.

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - You Are What You Read

T2 - The 13th Workshop on Computational Approaches to<br/>Subjectivity, Sentiment, & Social Media Analysis<br/>

AU - Sutton, Adam

AU - Simchon, Almog

AU - Edwards, Matthew

AU - Lewandowsky, Stephan

PY - 2023/7/14

Y1 - 2023/7/14

N2 - In this work we use consumed text to infer Big-5 personality inventories using data we have collected from the social media platform Reddit. We test our model on two datasets, sampled from participants who consumed either fiction content (N = 913) or news content (N = 213). We show that state-of-the-art models from a similar task using authored text do not translate well to this task, with average correlations of r=.06 between the model's predictions and ground-truth personality inventory dimensions. We propose an alternate method of generating average personality labels for each piece of text consumed, under which our model achieves correlations as high as r=.25 when predicting personality from the text being read.

AB - In this work we use consumed text to infer Big-5 personality inventories using data we have collected from the social media platform Reddit. We test our model on two datasets, sampled from participants who consumed either fiction content (N = 913) or news content (N = 213). We show that state-of-the-art models from a similar task using authored text do not translate well to this task, with average correlations of r=.06 between the model's predictions and ground-truth personality inventory dimensions. We propose an alternate method of generating average personality labels for each piece of text consumed, under which our model achieves correlations as high as r=.25 when predicting personality from the text being read.

M3 - Conference Contribution (Conference Proceeding)

SP - 28

EP - 38

BT - Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis

PB - Association for Computational Linguistics

Y2 - 14 July 2023

ER -

You Are What You Read: Inferring Personality From Consumed Textual Content

Abstract

Workshop

Research Groups and Themes

Access to Document

Handle.net

Fingerprint

Projects

Protecting the Democratic Information Space in Europe

Cite this