STAQ: Synthetic Technology Assistance Queries

STAQ: Synthetic Technology Assistance Queries STAQ is a synthetic dataset of technology-support queries designed to reflect communication patterns observed among older adults. Each synthetic query is paired with a clearer rephrased version that preserves the intended meaning. The dataset contains 514 query pairs across four communication categories: Incomplete: missing essential contextual information Underspecified: lacking sufficient detail to identify or resolve the problem Overspecified: containing unnecessary situational or contextual details Verbose: containing excessive narrative, repetition, or redundant information Each record includes the following fields: id: unique identifier query: synthetic technology-support query rephrased_query: clearer reformulation of the query category: communication category expert_rating: expert assessment of whether the query plausibly resembles a technology-support query communicated by an older adult A random subset of 50 synthetic queries was evaluated by domain experts for face validity. Expert ratings are encoded as: 0: Unlikely 1: Likely 2: Maybe N/A: Not rated This dataset accompanies the paper: Helping the Helper: LLM-Assisted Problem Articulation for Older Adults Seeking Technology Support Accepted at the 28th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2026).