Yearly salary filters can quietly reshape who appears in pay benchmarks

A developer explains that adding a yearly minimum salary filter to a job-listings scraper can change which postings enter a salary benchmark, even when the data pipeline succeeds. The filter requires a disclosed yearly minimum and compares amounts in each job's own currency, so it alters group membership without resolving currency comparability. The article outlines how to tell apart real pay changes from artifacts like missing currency fields, mixed pay periods, or hitting the collection limit. It recommends collecting a disclosure cohort first and applying analytical eligibility rules afterward. It also warns against filling missing values with zeros or inferred currencies, since that would manufacture misleading salary data.