Development and Validation of the Personal Values Dictionary: A Theory‐Driven Tool for Investigating References to Basic Human Values in Text. Issue 5 (3rd August 2020)

• Record Type: Journal Article
• Authors: Ponizovskiy, Vladimir; Ardag, Murat; Grigoryan, Lusine; Boyd, Ryan; Dobewall, Henrik; Holtz, Peter
• Editors: Rauthmann, John
• Other Names: Rauthmann John F. guestEditor.
• Is Part Of: European journal of personality. Volume 34:Issue 5(2020)
• Page Start: 885
• Page End: 902
• Publication Date: 2020-08-03
• Subjects: basic human values -- measurement -- language -- dictionary -- computerized text analysis; Personality -- Periodicals; 155.2
• DOI: 10.1002/per.2294
• Languages: English
• ISSNs: 0890-2070

Abstract

Abstract: Estimating psychological constructs from natural language has the potential to expand the reach and applicability of personality science. Research on the Big Five has produced methods to reliably assess personality traits from text, but the development of comparable tools for personal values is still in the early stages. Based on the Schwartz theory of basic human values, we developed a dictionary for the automatic assessment of references to personal values in text. To refine and validate the dictionary, we used Facebook updates, blog posts, essays, and book chapters authored by over 180 000 individuals. The results show high reliability for the dictionary and a pattern of correlations between the value types in line with the circumplex structure. We found small to moderate ( r s = .1–.4) but consistent correlations between dictionary scores and self‐reported scores for 7 out of 10 values. Correlations between the dictionary scores and age, gender, and political orientation of the author and scores for other established dictionaries mostly followed theoretical predictions. The Personal Values Dictionary can be used to assess references to value orientations in textual data, such as tweets, blog posts, or status updates, and will stimulate further research in methods to assess human basic values from text.