Rice finding 1170

Earlier matched collections uses the latest selected year and the earliest earlier year with common region × season groups. We calculate a median within each group, then the median across the same groups in each year, with equal group weight. Unmatched composites remain in the source tables. Other collection regions matches fiscal year and season and balances the peer regions. These are descriptive sample comparisons.

A reference middle 50% is the earlier groups' Q1–Q3, with linear interpolation. It describes that reference set; it is not a normal range, a safety threshold or an outlier test. A single earlier measurement has no meaningful reference distribution.

Regional results are prepared-food composites from collection markets, not growing origins or individual brands. Fiscal-year coverage is uneven. Matched comparisons retain common region/season strata, with equal stratum weight. Different sampled composites are not a longitudinal panel or a demonstrated population trend.

FDA regional foods combine purchases from three cities into one prepared-food composite. Row-specific administrative reporting limits are preserved. A non-detect is bounded by its reporting limit; no concentration is imputed. The multi-analyte rice records contain winter-only FY2020 collections; this does not mean every analyte was tested in that year. Matching also addresses unequal regional coverage, including the absence of West in the FY2018 winter subset.

FDA collection and preparation design · Original data key

Unmatched source-year summaries describe every reported sample in that year. The selected comparison uses only its declared matching groups. Different fiscal years can include collections from the same calendar year.

Collection markets do not identify growing or ingredient origin. Other collection regions compares samples matched by year and season.

Source-value tables retain public category-level sample identifiers, fiscal year, reported result and original locator. A zero lower interval endpoint is never displayed as an observed concentration.

Comparisons concern the selected historical dataset and are not current supplier rankings, compliance findings or independent measurements. They are descriptive sample comparisons and do not establish a population trend.