Analysis of Elections Conducted in Representation Cases Closed, Fiscal Year 1990

p. 175. Grid is 17 columns: stub plus four panels (All R, RC, RM, RD) of four columns each. "Elections conducted" is centered over all four columns of each panel, so it is repeated 16 times in header level 2. The footnote marker 1 is printed only on the All R panel's "Resulting in certification" heading. The "Rerun required" / "Runoff required" sub-rows carry a value only in each panel's "Resulting in a rerun or runoff" column; every other cell in those rows is a printed dash rule (------) and is transcribed as "". Transcribed as printed, without correction; several arithmetic inconsistencies are present: - Stipulated elections, RC panel, "Total elections": printed 2,881. This does not reconcile either across its own row (1 + 50 + 2,930 = 2,981) or down the RC total column (44 + 2,881 + 579 + 2 + 1 = 3,507 vs. the printed All types figure of 3,607). Both checks point to a misprint for 2,981. The scan is clear enough that I read the digit as an 8, but this is the single cell in the table I would flag as most likely a printing error rather than a reading error. - Last row ("Runoff required" under Expedited—Sec. 8(b)(7)(C)), RM panel: printed 1, although the Expedited row shows 0 RM elections in all four RM columns. This stray 1 is what makes the RM "Resulting in a rerun or runoff" column sum to 17 against the printed All types figure of 18, and the All R rerun-or-runoff column sum to 97 against the printed 98. I re-checked the cell position and the 1 does sit under the RM "Resulting in a rerun or runoff" column. - Board-directed row: the All R figures (4 total, 3 resulting in certification) exceed the sum of the three sub-panels (RC 2 + RM 0 + RD 1 = 3 total; RC 2 + RM 0 + RD 0 = 2 certified). Correspondingly the RM "Total elections" column sums to 104 against the printed All types figure of 105. All other row and column cross-checks in the table balance. The rightmost cell of the "Expedited—Sec. 8(b)(7)(C)" row (RD panel, "Resulting in certification") sits right at the edge of the crop; I read it as 1, and it is the value needed to balance both that row (0 + 0 + 1 = 1) and the RD certification column (11 + 504 + 71 + 0 + 1 = 587). The scan uses a dotless i throughout (certification, Director, in); normalized to ordinary spelling in headings, labels and the footnote.