Joint mean intervals left eight ozone time and mass directions unresolved in three papers from 1997-2015

We tested whether sampling intervals can support joint directions for ozone effects on development time and pupal mass without known endpoint covariance. Eight matched comparisons came from three papers in a 2024 public extract. We formed t intervals for reported means and ratio bounds from their extremes. A union bound gave conditional joint coverage without a covariance estimate. Every time and mass ratio interval included 1 when error control covered four means in one pair. The same result held across all eight pairs. Separate 95 percent mean intervals resolved two time directions and no mass direction. These conservative intervals do not show zero effects or rule out more precise methods. Coverage requires suitable independent units and a normal-mean model; original chamber designs were not recovered. Complete-endpoint selection and one paper per country limit wider use. No independent region-by-period replication or climate interaction was available. Code, aggregate results, a figure, source hashes and external retrieval instructions are included.

ozone · development time · pupal mass · joint confidence intervals · observation bias · covariance · replication limits