Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Psychometrics›Wordscores
Latent structureText Scaling

Wordscores

Wordscores is a text-based scaling method developed by Laver, Benoit, and Garry (2003) that estimates the policy positions of political actors based on word frequencies in their texts. By comparing word usage in reference texts of known positions with test texts, the method infers the latent political dimension of any document without requiring manual coding or training data.

ScholarGate
  1. Latent structure
  2. v1
  3. 3 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Wordscores
Exploratory Structural E…Latent Transition Analys…Multiple Factor AnalysisPartial Least Squares St…WordfishIdeal Point EstimationManifesto CodingRedundancy AnalysisWordfish Scaling

When to use it

Apply Wordscores when you have reference texts with known positions and want to scale new documents on the same latent dimension. Ideal for analyzing party manifestos, legislative speeches, or news archives where reference materials exist. The method works best when reference texts are truly representative of their positions and when vocabulary is stable across documents.

Strengths & limitations

Strengths
  • Unsupervised scaling: no manual coding required after reference texts are selected
  • Fast and scalable: easily processes thousands of documents
  • Domain-general: works across different text types (manifestos, speeches, newspaper articles)
  • Theory-driven: leverages known reference positions to interpret the latent dimension
  • Straightforward interpretation: positions are directly comparable to reference scales
Limitations
  • Depends on reference texts: quality and representativeness of reference texts directly affect accuracy
  • Assumes single dimension: typically estimates one latent dimension (e.g., left-right) rather than multiple dimensions
  • Vocabulary sensitivity: rare words and new vocabulary in test texts are not scored, potentially introducing bias
  • Static references: assumes reference positions are constant and comparable to test documents across time

Frequently asked

How do I choose appropriate reference texts?

Choose reference texts that are clearly and unambiguously representative of the extreme positions on your latent dimension. For political analysis, use manifestos of clearly left and right parties. Ensure they are substantial in length (hundreds or thousands of words) and typical of the document genre.

What if a word appears in my test text but not in the reference texts?

Words absent from reference texts receive no score (treated as missing). Many implementations handle this by excluding rare words entirely or using smoothing techniques. Some methods interpolate or use related words, but this adds assumptions.

Can Wordscores estimate multiple dimensions at once?

Standard Wordscores is designed for single-dimension scaling. To estimate multiple dimensions (e.g., left-right and social-liberal), you need multiple pairs of reference texts or move to methods like Wordfish or topic modeling.

How stable are Wordscores across time?

Wordscores assumes the meaning of words is stable, which breaks down if vocabulary shifts dramatically. Over long time periods or with evolving policy domains, reference-based methods may need updating or be supplemented with relative scaling methods like Wordfish.

What metrics should I use to validate Wordscores results?

Compare estimated positions of reference texts themselves (should recover known positions), correlate with external measures (expert surveys, vote records), or perform holdout validation (remove some reference texts and see if Wordscores still predicts them accurately).

Sources

  1. Laver, M., Benoit, K., & Garry, J. (2003). Extracting policy positions from political texts using words as data. American Political Science Review, 97(2), 311-331. DOI: 10.1017/s0003055403000698 ↗
  2. Benoit, K., & Laver, M. (2012). The basic arithmetic of legislative decisions. Journal of Political Institutions and Political Economy, 1(1), 1-29. link ↗
  3. Klemmensen, R., Hobolt, S. B., & Hansen, M. E. (2007). Estimating policy positions using political texts: A scaling approach. Electoral Studies, 26(4), 746-755. link ↗

How to cite this page

ScholarGate. (2026, June 3). Wordscores. ScholarGate. https://scholargate.app/en/psychometrics/wordscores

Related methods

Exploratory Structural Equation ModelingLatent Transition AnalysisMultiple Factor AnalysisPartial Least Squares Structural Equation ModelingWordfish

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • Exploratory Structural Equation ModelingPsychometrics↔ compare
  • Latent Transition AnalysisPsychometrics↔ compare
  • Multiple Factor AnalysisPsychometrics↔ compare
  • Partial Least Squares Structural Equation ModelingPsychometrics↔ compare
  • WordfishPsychometrics↔ compare
Compare side by side →

Referenced by

Exploratory Structural Equation ModelingIdeal Point EstimationManifesto CodingPartial Least Squares Structural Equation ModelingRedundancy AnalysisWordfishWordfish Scaling

Similar methods

WordfishWordfish ScalingDictionary-Based Text Analysis in PoliticsPolitical Ideology ScalingManifesto CodingDictionary-Based Text AnalysisSupervised Text ClassificationContent Analysis of Political Speeches

Related reference concepts

Topic Modeling and Text MiningText ClusteringText Classification and Sentiment AnalysisEvaluation and AnnotationText ClassificationLatent Semantic and Topic Models

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Wordscores (Wordscores). Retrieved 2026-07-21 from https://scholargate.app/en/psychometrics/wordscores · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Michael Laver, Kenneth Benoit, John Garry
Subfamily
Text Scaling
Year
2003
Type
Text analysis and dimension reduction
Related methods
Exploratory Structural Equation ModelingLatent Transition AnalysisMultiple Factor AnalysisPartial Least Squares Structural Equation ModelingWordfish
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account