In reply to @zdenekvrozina

James Throt MBBS, MD, PhD, FRCPath

James Throt MBBS, MD, PhD, FRCPath

@jamesthrot · Twitter ·

@ZdenekVrozina The point is that “fit for reporting” does not mean “free from bias” or “safe to treat as definitive evidence of no decline”. OECD sets the 5% exclusion limit for a reason. Once you exclude too many pupils, representativeness worsens and the mean can be biased upwards.