"It makes sense to me": Examining data file column names in the visual cognitive literature

The open science movement has expanded expectations for transparency and reproducibility, yet the usability of shared datasets remains an underexplored barrier to cumulative science. In this study, we systematically examined 115 datasets from recent visual cognition publications to evaluate how core variables are labeled and documented. Across these datasets, we identified more than 3,000 unique column names, with most appearing only once, reflecting a lack of shared conventions. Even for foundational measures common to most experiments in the field (which we refer to as the "Big Four," i.e., participant identifiers, trial identifiers, response accuracy, and response times) we observed striking variability. Many datasets appeared to be un-curated exports from data collection software, often containing redundant or irrelevant variables, inconsistent coding schemes, or ambiguous column headings. Accessibility was also a recurring issue, with 28 datasets excluded from analysis due to broken links, restricted access, and interoperability issues arising from the use of closed file formats. To address these challenges, we propose concrete recommendations for standardizing column names, with specific guidelines for the Big Four variables, alongside broader suggestions for dataset curation, accessible file formats, and minimal documentation. We also introduce Output It Forward, a Chrome extension developed to streamline the identification of data availability statements and repository links. By highlighting inconsistencies in current practices and offering practical recommendations, our findings underscore that data sharing must go beyond availability to ensure usability. Clearer conventions and community standards will enhance the transparency, interpretability, and long-term value of shared datasets in visual cognition and beyond.