Industrial Distribution of Representation Elections Held in Cases Closed, Fiscal Year 2004—Page 1 of 6

p. 248. Fifteen columns: stub, "Total elections", four columns under "Number of elections in which representation rights were won by unions", "Number of elections in which no representative was chosen", "Number of employees eligible to vote", "Total valid votes cast", four columns under "Valid votes cast for unions", "Total votes for no union", and the final "Eligible employees in units choosing representation". The stub heading carries a raised 1 ("Industrial Group^1"), but no footnote text appears on this crop; the page is cut off immediately below the "Wood Product Manufacturing" row, so the footnote (and the remainder of the table, which the caption marks as "Page 1 of 6") falls outside the image. The last column heading is typeset in a narrow column and broken as "Eligible employ- ees in units choos- ing rep- resentati on"; I have joined it and repaired the hyphens to "Eligible employees in units choosing representation". Comma usage is inconsistent in the original and is transcribed as printed: the "Number of employees eligible to vote" column and the last column are set without digit-grouping commas (e.g. 3755, 4010, 7703, 8970, 10114, 12735, 2113, 4318, 4852, 2987, 3360), while the vote columns use them. Suspected misprint, transcribed as printed: in "Leather and Allied Product Manufacturing" the total valid votes cast (501) exceeds the number of employees eligible to vote (498); its "Eligible employees in units choosing representation" is 0, consistent with 0 elections won. The subtotal row is printed as "31-Manufacturing" (with the leading sector-code digits), unlike the other subtotal rows, which are unnumbered; it is transcribed exactly as printed. Section subtotal rows (Agriculture..., Mining, Construction, 31-Manufacturing) are set off by heavier rules and are indented, but are ordinary data rows with values. All internal arithmetic checks I could perform (union + no-union votes = total valid votes; AFL-CIO + other national + other local = union total; component rows summing to each subtotal) are consistent with the figures as transcribed, except the Leather row noted above.

p. 249. Crop is page 2 of 6 of Table 16; the table continues before and after this image, so the page is a continuation even though the column headings are reprinted here. 15 columns: stub, then Total elections; the four-column block "Number of elections in which representation rights were won by unions" (Total / AFL-CIO unions / Other national unions / Other local unions); Number of elections in which no representative was chosen; Number of employees eligible to vote; Total valid votes cast; the four-column block "Valid votes cast for unions" (Total / AFL-CIO unions / Other national unions / Other local unions); Total votes for no union; Eligible employees in units choosing representation. The last heading is broken across lines as "representati / on" and has been joined to "representation". "AFL-CIO" is printed with an en dash (AFL–CIO) in both heading blocks; transcribed with a plain hyphen. The subtotal row "32-Manufacturing" does not equal the sum of the six detail rows shown above it on this page (e.g. elections 139 vs 158 printed; employees eligible 10,254 vs 11,412 printed). This is not a misprint: NAICS 32 also includes a Wood Product Manufacturing line that falls on the preceding page of the table, outside this crop. The residual is internally consistent (one extra line of 19 elections, 8 won, 11 with no representative). The "33-Manufacturing" and "Wholesale Trade" subtotals do reconcile exactly with the detail rows printed on this page, in every column. "Fabricated Metal Product Manufacturing" is followed by a single period that is the start of the leader-dot run; it has been dropped, consistent with the other labels. Rule weights in the image mark 32-Manufacturing, 33-Manufacturing and Wholesale Trade as subtotal rows (double rules below each); they are transcribed as ordinary body rows with their values. All cells were legible; no blanks, dashes or footnote markers appear in the body.

p. 250. 15 columns. The stub heading carries a raised "1" (Industrial Group^1), but no footnote text is printed on this crop — the footnote appears elsewhere in the six-page table, so `footnotes` is empty. This is "Page 3 of 6" of a continuing table. The 44-Retail Trade subtotal row covers industries that begin on the previous page: the seven 44- rows printed here sum to 88 elections, not the printed 140 (likewise 1006 vs. the printed 1403 in the last column). The printed subtotal is internally consistent (73 + 67 = 140; 1,348 + 1,788 = 3,136; 1,313 + 26 + 9 = 1,348), so this is a crop artifact, not a misprint. The 45-Retail Trade and 48-Transportation subtotals do reconcile with the rows shown above them. Both subtotal rows for NAICS 44 and 45 are labeled "Retail Trade" as printed. Digit-grouping commas are applied inconsistently in the source: the "Number of employees eligible to vote" and "Eligible employees in units choosing representation" columns print four-digit values without commas (1699, 2305, 2531, 2605, 7422, 1373, 11844, 3640, 1403, 1354, 1368, 3457, 5983), while the other columns use commas. Transcribed as printed. "AFL–CIO" is set with an en dash in the headings; reproduced as printed. All rows satisfy the internal checks (union votes + no-union votes = total valid votes; Total = AFL–CIO + other national + other local; elections won + elections with no representative = total elections), which supports the column assignments.

p. 251. 15 columns: stub plus 14 numeric columns. The crop is page 4 of 6 of a continuing table; no footnote text is printed on this page, although the stub heading carries marker 1 (footnote text appears elsewhere in the report). Indented subtotal lines (49-Transportation; Information; Finance and Insurance; Real Estate and Rental and Leasing) are transcribed with their printed wording; "49-Transportation" begins with the printed sector code "49-". All four subtotals add up exactly from their component rows in every one of the 14 columns, which supports the digit readings. Thousands separators are inconsistent in the source: the "Number of employees eligible to vote" column and the "Eligible employees in units choosing representation" column print four-digit values without commas (e.g. 4937, 5721, 5421, 4944, 5336, 1369; 1345, 1879, 1401, 1140), while neighbouring columns use commas. Transcribed as printed. Possible source anomaly, not a reading error: the "Insurance Carriers and Related Activities" row shows 0 elections won by unions and 3 elections with no representative chosen, yet reports 964 in "Eligible employees in units choosing representation" and 2,113 valid votes for unions against 1,773 for no union. These figures are printed as shown and are what the Finance and Insurance subtotal is built from. Dashes: "AFL–CIO" in the headings and "Monetary Authorities–Central Bank" in the stub are printed with en-dashes; "Non-Financial" is a hyphenated compound broken across printed lines and rejoined. The table body continues past the bottom of the crop below the "Professional, Scientific and Technical Services" row.

p. 252. Fifteen columns, two heading levels; the spanner "Number of elections in which representation rights were won by unions" covers columns 2-5 and "Valid votes cast for unions" covers columns 9-12, so those headings are repeated across their spans. The stub heading and the six tall single headings are repeated in both levels. The last column's heading is printed hyphenated and broken across lines ("Eligible employ- ees in units choos- ing rep- resentati on") and is joined here as "Eligible employees in units choosing representation". This crop is page 5 of 6 of Table 16, so the table begins before and continues after this image; there is no grand total row here. The caption, which carries the "Page 5 of 6" designation, is omitted per instructions. Typesetting quirk, transcribed as printed: the two "eligible employees" columns (col 7, "Number of employees eligible to vote", and col 14, "Eligible employees in units choosing representation") are printed WITHOUT digit-grouping commas (e.g. 10276, 27767, 52179, 18072, 35137), while the vote columns use commas. No commas were added. Row labels that wrap over two or three printed lines were joined: "Management of Companies and Enterprises"; "Waste Management and Remediation Services"; "Administrative and Support, Waste Management and Remediation Services"; "Performing Arts, Spectator Sports and Related Industries"; "Museums, Historical Sites and Similar Institutions"; "Amusement, Gambling and Recreation Industries". Several labels are indented as subtotal lines (Management of Companies and Enterprises; the Administrative/Waste combined line; Health Care and Social Assistance; Arts, Entertainment and Recreation; Accommodation and Foodservices); indentation is not reproduced in the cell text. No illegible cells. Arithmetic checks (component union columns summing to their Total, union total plus "no union" equalling total valid votes cast, election components summing to elections won, elections won plus no-representative equalling total elections, and the Health Care and Social Assistance subtotal against its four component rows) are internally consistent throughout; nothing appears misprinted.

p. 253. This is the last page (page 6 of 6) of the table; it continues the industry list from earlier pages, so only the tail of the industrial groups plus the grand total appear here. Column headings are reprinted in full on this page, so header_rows is populated. 15 printed columns. The stub heading carries a superscript 1 ("Industrial Group^1"), but no footnote text is printed under the table on this crop (it is presumably on an earlier page of the table), so footnotes is empty. Two heading cells wrap heavily in the original and have been joined/de-hyphenated: "Number of elections in which no representative was chosen" and the last column, printed as "Eligible employ-/ees in/units/choos-/ing rep-/resentati/on" → "Eligible employees in units choosing representation". Formatting is inconsistent in the source and has been kept as printed: the "Number of employees eligible to vote" column and the last column print four- to six-digit figures without digit-grouping commas (1467, 1390, 4065, 1043, 1664, 1672, 192537; 1891, 1192, 95778), while the neighboring columns use commas (1,265; 3,677; 161,971). Nothing was normalized. The three rows "Other Services (except Public Administration)", "Public Administration" and "Unclassified Establishments" are indented subtotal/summary lines set off by horizontal rules; their labels are transcribed as printed. All digits were legible; no doubtful cells.