Other pages in this series (5)
A2
Revisions outside the previous range
Share of revisions in which the new value left the previous recommendation's uncertainty range
Takeaway
- CIAAW uncertainties are not standard uncertainties (1σ) but expanded ranges (wide ranges that normal materials fall in with great certitude); so only 12 of 204 revisions (5.9%) put the new value outside the previous range.
- In 2 revisions the two ranges did not overlap at all (Br 1965, Zn 2007). CIAAW's FAQ describes a new value outside the previous range as a single case, zinc in 2007, but recounting from the historical table does not reproduce that single case under either criterion.
- Definition
- Only revisions whose previous recommendation carried an uncertainty. Two criteria: (1) the new value lies outside the previous value ± previous uncertainty; (2) the new range and the previous range do not overlap at all. CIAAW uncertainties are not standard uncertainties (1σ) but expanded ranges, within which the atomic weight of normal materials is expected to lie 'with great certitude'. They are not expanded uncertainties with a stated coverage factor or confidence level.
- Formula
- (1) |x_new − x_old| > U_old (2) x_old + U_old < x_new − U_new or x_new + U_new < x_old − U_old. Values before 1961 are multiplied by f = 15.9994/16 to put them on the ¹²C scale (15.9994 is oxygen in the 1961 table).
- Denominator
- 204 revisions whose previous recommendation (from 1951 on) carried an uncertainty and whose new recommendation is a single value (with or without an uncertainty). Switches to an interval (A5) and withdrawals are excluded. A new value without an uncertainty is treated as a point of zero width under criterion (2) (one case, Cr 1967).
- How to read it
- Criterion (1): 12 revisions (5.9%); criterion (2): 2 (Br 1965, Zn 2007). Of the 12 under (1), 6 revise recommendations made in 1961–1969. CIAAW's FAQ describes a new value outside the previous recommendation's uncertainty range since 1951 as a single case, zinc in 2007. That is criterion (1); recounting from the same body's historical table (CSV) gives 12 under (1) and 2 under (2), so the FAQ's single case is not reproduced under either criterion. Why the FAQ's count differs from the historical table (what it counted, and as of when) cannot be determined from the public material.
- Caveats
- CIAAW ranges are deliberately wide expanded ranges, so no normal-distribution expectation (31.7% outside 1σ) applies, and the figures are not compared directly with the miss rates of the physical-constants chapter (standard uncertainties). The CSV gives 28.9815 for aluminium in 1969, inconsistent with the 1961 and 1971 rows of the same table (26.9815, 26.981 54); it is treated as a slip for 26.9815 (not yet checked against the original report; recorded in the corrections file).
| Symbol | Previous | New | z_old | Ranges disjoint |
|---|---|---|---|---|
| Br | 1961: 79.909(2) | 1965: 79.904(1) | −2.500 | yes |
| C | 1961: 12.011 15(5) | 1969: 12.011(1) | −3.000 | |
| H | 1961: 1.007 97(1) | 1969: 1.0080(3) | 3.000 | |
| S | 1961: 32.064(3) | 1969: 32.06(1) | −1.333 | |
| Bi | 1969: 208.9806(1) | 1971: 208.9804(1) | −2.000 | |
| K | 1969: 39.102(3) | 1971: 39.098(3) | −1.333 | |
| Ge | 1985: 72.61(2) | 1999: 72.64(1) | 1.500 | |
| Pt | 1995: 195.078(2) | 2005: 195.084(9) | 3.000 | |
| Zn | 2001: 65.409(4) | 2007: 65.38(2) | −7.250 | yes |
| Th | 2005: 232.038 06(2) | 2013: 232.0377(4) | −18.000 | |
| Yb | 2007: 173.054(5) | 2015: 173.045(10) | −1.800 | |
| Lu | 2007: 174.9668(1) | 2024: 174.966 69(5) | −1.100 |
Source: Standard atomic weights © IUPAC/CIAAW 2016–2024. Commission on Isotopic Abundances and Atomic Weights, Historical changes of standard atomic weights (compiled by J. Meija), https://www.ciaaw.org/historical-atomic-weights.htm