O-92 - OPEN protein under-reporting was ~27-34% on FFQ and ~11-15% on 24HR, but protein density was barely biased
Protein absolute-intake under-reporting parallels energy (FFQ worse than 24HR). The near-absence of bias in protein DENSITY is important: much of the absolute under-reporting is a scaling/energy effect that partly cancels in ratios, which is why energy-adjusted or density exposure measures are less biased than absolute intakes.
Link to original
Why this is evidence
Absolute protein is under-reported ~30% yet protein density (a ratio) barely at all. H-41 reads this as the empirical basis for energy-adjustment/ratio measures removing most measurement bias (the error is manageable); H-31 emphasizes the large absolute-intake error. Discriminates toward the defence pole.