Size of Units in Collective-Bargaining and Decertification Elections, Fiscal Year 1954

p. 182. Structure: the page carries one table with two stacked sections, "A. COLLECTIVE-BARGAINING ELECTIONS" (printed above the column headings) and "B. DECERTIFICATION ELECTIONS" (printed between the two bodies). Both are transcribed as label-only section rows. Section B repeats section A's eleven columns but prints no headings of its own, so the headings are given once. Decimal points: this is a 1-bit scan in which the decimal point of the percent columns very often drops out, so values print as "100 0", "23 1", "11 5". I have restored the point throughout ("100.0", "23.1", "11.5"), as with the comma-space rule. By the same logic a bare right-aligned digit in a percent column has been read as a leading-point fraction: panel A 400-499 and 500-599 "Percent of total" = ".9"; 2,000-2,999 "Percent of total" = ".3" and A.F.L. Percent = ".2"; 3,000-3,999 and 5,000-9,999 "Percent of total" = ".1"; C.I.O. Percent ".4" (2,000-2,999) and ".1" (3,000-3,999); Unaffiliated Percent ".3" (4,000-4,999). Each of these is consistent with the printed counts, so I am confident in the reading, but the points themselves are not legible on this scan. Zero cells: in the percent columns a zero is printed sometimes as "0" and sometimes as ".0". At this resolution the presence or absence of the leading point on a zero is the least certain detail in the grid; I transcribed each as it appeared, but individual zero cells (notably panel B, C.I.O. Percent and Unaffiliated Percent) could go either way. "(^1)" cells: at panel A row 3,000-3,999 (A.F.L. Percent) and row 10,000-15,000 (Percent of total) the value position holds a parenthesised raised 1, i.e. the reference to footnote 1 ("Less than one-tenth of 1 percent") standing in place of a number, not a printed value. Headings: the periods in "A. F. L. affiliates" and "C. I. O. affiliates" are faint or broken on the scan (the C.I.O. line reads roughly "C I O. affili-ates"); I have set them in the conventional form. The em dash closing "won by—" is printed. Doubtful digits: panel B, "400 and over", "Percent of total" reads as 3.4 but 5 of 150 is 3.3, and the neighbouring 200-299 row with the same count of 5 prints 3.3 — the digit is legible as 4 but may be a misprint or a scan artefact. Printed values that disagree with the counts, transcribed as printed and not corrected: panel A 10-19, "Elections in which no representative was chosen" Percent prints 18.0 where 287 of 1,603 is 17.9; panel A 90-99 and 150-199, Unaffiliated Percent prints 4.0 where 14 of 355 is 3.9; panel B 70-79 and 100-149, C.I.O. Percent prints 7.2 where 1 of 14 is 7.1, and A.F.L. Percent prints 3.5 in four rows but 3.4 in two others for the same 1 of 29. Crop: the footnote line at the foot of the page is clipped by the bottom edge of the image — the marker and the words are readable but the descenders are cut, and the terminal period is inferred. Verification: every count column was summed against its printed total. Panel A: 4,663; 1,924; 781; 355; 1,603 — all five agree. Panel B: 150; 29; 14; 5; 102 — all five agree.