Can ‘pure‐path’ models explain free oscillation data?