[DSRP Evidence](https://dsrpevidence.org/)

# Evaluating Context in LLM Prompts for Causal Inference in Empirical Sustainability Studies

## Details

**Authors** Zhang and Fan

**Year** 2026

**Publisher** Environmental Research Communications

**Discipline** Engineering

**Secondary disciplines** Computer Science & AI, Environment & Sustainability

[Read it at the publisher](https://doi.org/10.1088/2515-7620/ae9d05) 
10.1088/2515-7620/ae9d05

## In authors' words

### What they found (results)

Benchmarking GPT-5 against five peer-reviewed sustainability studies, the authors found the model recovers core causal edges with moderate accuracy (0.33-0.85) and assigns direction reliably, but shows a strong tendency toward over-connecting variables with spurious causal links.

## Commentary

### In short

An AI system can correctly identify causal relationships and their direction yet fail at the accompanying distinction task of bounding which connections genuinely belong, revealing that inferring a relationship and correctly delimiting it are separable competencies.

**Patterns it shows** D, R

**Added** 2026-08-23

**How to cite this** Zhang and Fan (2026). Evaluating Context in LLM Prompts for Causal Inference in Empirical Sustainability Studies. Environmental Research Communications.