---
license: mit
language:
  - en
tags:
  - text-classification
  - social-science
  - politics
  - sentiment-analysis
  - benchmark
task_categories:
  - text-classification
task_ids:
  - multi-class-classification
  - sentiment-classification
annotations_creators:
  - human-annotated
multilinguality: monolingual
---

[![GitHub Repo](https://img.shields.io/badge/GitHub-View%20Repo-black?style=flat-square&logo=github&logoColor=white)](https://github.com/iqss-research/readme-software)

# Automated Nonparametric Content Analysis Datasets

This repository provides the four benchmark datasets used in:

> Connor T. Jerzak, Gary King, and Anton Strezhnev. **An Improved Method of Automated Nonparametric Content Analysis for Social Science.** *Political Analysis*, 31(1): 42–58, 2023.

Each dataset is formatted for easy loading in Python and R (CSV). Labels are integer-coded from `1,...,K`; text is provided as raw strings.

## Datasets

| Name            | Documents | Categories | Source & Description                                                                                                                                             |
| --------------- | --------: | ---------: | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **enron.csv**       |     1,426 |          5 | Corporate emails from the Enron corpus, hand-coded into five thematic categories (e.g., business, personal, legal)                                               |
| **immigration.csv** |       462 |          5 | Newspaper editorials on immigration policy, hand-coded into five sentiment/policy categories; originally used in Hopkins & King (2010) and Jerzak et al. (2023)  |
| **clinton.csv**     |     1,938 |          7 | Blog posts about Hillary Clinton from 2008, hand-coded into seven topical categories; feature space of \~3,623 word stems                                        |
| **stanford.csv**    |    11,855 |          5 | Sentences from the Stanford Sentiment Treebank, labeled on a five-point sentiment scale; commonly used in text quantification research                           |


---

### Citation

Connor T. Jerzak, Gary King, Anton Strezhnev. *An Improved Method of Automated Nonparametric Content Analysis for Social Science*. Political Analysis, 31(1): 42–58, 2023. [\[PDF\]](https://gking.harvard.edu/sites/scholar.harvard.edu/files/gking/files/div-class-title-an-improved-method-of-automated-nonparametric-content-analysis-for-social-science-div.pdf)

```
@article{JSK-readme2,
  title={An Improved Method of Automated Nonparametric Content Analysis for Social Science},
  author={Jerzak, Connor T. and Gary King and Anton Strezhnev},
  journal={Political Analysis},
  year={2023},
  volume={31},
  number={1},
  pages={42-58}
}
```