Integrating Unstructured Text into Causal Inference: Empirical Evidence from Real Data

TL;DR

This paper introduces a framework that uses transformer-based language models to perform causal inference with unstructured text, demonstrating comparable results to traditional structured data methods and expanding causal analysis to text-only scenarios.

Contribution

The paper presents a novel approach that leverages unstructured text for causal inference, validated by empirical comparisons with structured data across multiple levels.

Findings

Unstructured text can produce causal estimates similar to structured data.

The framework extends causal inference applicability to text-only data scenarios.

Results are consistent across population, group, and individual levels.

Abstract

Causal inference, a critical tool for informing business decisions, traditionally relies heavily on structured data. However, in many real-world scenarios, such data can be incomplete or unavailable. This paper presents a framework that leverages transformer-based language models to perform causal inference using unstructured text. We demonstrate the effectiveness of our framework by comparing causal estimates derived from unstructured text against those obtained from structured data across population, group, and individual levels. Our findings show consistent results between the two approaches, validating the potential of unstructured text in causal inference tasks. Our approach extends the applicability of causal inference methods to scenarios where only textual data is available, enabling data-driven business decision-making when structured tabular data is scarce.