New Detailed Birthplace Codes for Germany for 1860, 1870 & 1900

The final versions of the 1860, 1870, and 1900 IPUMS samples include new birthplace codes for Germany. These new codes are based on Germany's administrative areas and their broader regional categories circa 1900, a convenient midpoint. The new classification scheme reduces the number of codes associated with sub-regions of Germany, especially for the various sub-areas of Saxony, Hesse, Hesse-Nassau, and Rhineland. The division of West and East Germany is maintained for consistency with earlier versions of IPUMS, and the codes for Alsace-Lorraine in France and German areas in Poland do not change.

The new German detailed birthplace codes are as follows:

Detailed birthplace
453 00 Germany, nec (not elsewhere classified)
453 01 Berlin
453 02 West Berlin
453 03 East Berlin
453 10 West Germany, nec
453 11 Baden (Mannheim, Karlsruhe, Freiburg, Konstanz)
453 12 Bavaria (Mittelfranken, Niederbayern, Oberbayern, Oberfranken, Oberpfalz, Pfalz, Schwaben, Unterfranken)
453 13 Braunschweig (three parts--between Hannover, Hesse-Nassau, and Brandenburg)
453 14 Bremen
453 15 Hamburg
453 16 Hannover (Aurich, Hannover, Hildesheim, Luneburg, Osnabrück, Stade)
453 17 Hesse (Oberhessen, Starkenburg)
453 18 Hesse-Nassau (Kassel, Wiesbaden)
453 19 Lippe
453 20 Lübeck
453 21 Oldenburg (three parts--one in Rhineland {Birkenfeld}, one between Schleswig and Lübeck {F. Lübeck})
453 22 Rhineland (Aachen, Düsseldorf, Koblenz, Köln, Saarland, Trier)
453 23 Schaumburg-Lippe
453 24 Schleswig
453 25 Sigmaringen (Hohenzollern)
453 27 Westphalia (Arnsberg, Minden, Münster)
453 28 Württemberg (Donaukreis, Jagstkreis, Neckarkreis, Schwarzwaldkreis)
453 40 East Germany, nec
453 41 Anhalt
453 42 Brandenburg (Magdeburg, Potsdam)
453 44 Kingdom of Saxony (Dresden, Leipzig, Zwickau)
453 45 Mecklenburg
453 46 Saxony (Erfurt, Merseburg)
453 47 Thüringen
453 60 Prussia, nec

Modified Urban/Rural Definition

The final versions of the 1860 and 1870 samples rely on a slightly modified urban/rural definition compatible with the Census Bureau's historical series. In the preliminary samples, all New England towns containing 2,500 or more inhabitants were classified as urban. The final samples use the original, unpublished worksheets prepared at the U.S. Census Bureau in the 1930s to define urban places. In addition to the regularly incorporated cities and municipalities containing 2,500 or more individuals, the Bureau considered as urban only those New England towns in which there was a village or thickly settled area having more than 2,500 inhabitants and comprising--either by itself or when combined with other villages within the same town--more than 50 percent of the total population of the town. As a result, the percentage of sample individuals classified as urban in the final samples is slightly lower than that reported in the preliminary samples. Future versions of the 1850 and 1880-1920 IPUMS datasets will use this modified urban definition.

Changes To OCC1950 Imputation Procedures

A few changes were made to the occupation imputation procedures in the final versions of the 1860 and 1870 IPUMS samples. In the preliminary versions of the datasets and in the existing 1850 and 1880 IPUMS samples, all laborers in households containing a farmer were reclassified as farm laborers (OCC1950 code '820'). In the final samples and in future versions of the other historical samples, we will rely on a modified procedure. Laborers with a specific non-agriculture industry noted (e.g. "Laborer Brickyard") will no longer be reclassified as farm laborers if a farmer was in the household. Although this change effects only a handful of cases, it increases the accuracy of the imputation.

We have also revised our imputation procedures for housekeepers. As in the preliminary versions of the samples, the OCC1950 code of '700'--"Housekeeper" is reserved for individuals in the paid labor force, and the OCC1950 code of '980'--"Keeps house" is reserved for individuals who keep their own home and are therefore not in the paid labor force. Unfortunately, the sheer number of "housekeepers" enumerated in the nineteenth century--many rural areas have a "housekeeper" in every household--suggests widespread enumeration error. In the past, we reclassified women enumerated as "housekeeper" as "keeping house" (i.e., we removed them from the paid labor force) if they were related to the head of the household in which they resided. These women were given an OCC1950 code of '981'--"Imputed keeping house" to distinguish them from women whose occupational response was "keeping house." In the final version of the 1860 and 1870 IPUMS samples we reversed the procedure. All women with the occupational response of "housekeeper" were initially given an OCC1950 code of '981' and later reclassified as '700' if they were unrelated to the household head and had no spouse present. An exception is individuals with an occupational response that clearly indicated they were in the paid labor force (e.g., "Housekeeper for the Buckly family"). These individuals were initially given the OCC1950 code of '700'. Unlike the preliminary samples, these individuals will not be reclassified as '981,' even if they are related to the household head and have a spouse present. In the end, the reversed imputation procedures changed very few cases. As before, however, users interested in female labor force participation should be aware of the severe limitations inherent in the data.

These changes affect the value labels as follows:

New and Deleted Cases

Finally, a review of sample densities by county in the preliminary releases of the 1860 and 1870 samples revealed a few missed counties and a few over-sampled counties, which have been corrected in the final versions. The details are as follows:

Data additions/subtractions for 1860

  1. Our original copy of the microfilm reel containing the manuscript pages for Washington County, Rhode Island was missing a number of pages. We obtained a new copy of the reel and sampled the missing pages. Twenty households containing 142 individuals--residing in Kingston, Richmond, and Westerly, RI--were added to the dataset.
  2. We discovered that the manuscript pages for Marshall County, IL were duplicated on microfilm reels 209 and 210. Since the pages should have only been on reel 210, we deleted the 29 households and 135 person records we had sampled from reel 209.
  3. Likewise, the manuscript pages for Mason County, IL were present on both reels 209 and 210. Since the pages should have only been on reel 210, we deleted the 21 households and 123 person records initially obtained from reel 209.
  4. Twelve households containing 75 individuals from Talladega County, AL were excluded from the preliminary dataset by mistake. The data was added to the final dataset.

Data additions/subtractions for 1870

  1. Microfilm reel 1165 (Wayne and Wilkes Counties, NC) was incorrectly duplicated and numbered reel 1164 (Warren, Washington, and Watauga Counties), resulting in the former being entered twice and the latter not being included in the preliminary sample. For the final sample we deleted 62 households and 272 person records from Wayne County, 41 households and 226 person records from Wilkes County, and added new cases sampled from Warren, Washington, and Watauga counties.
  2. Two different enumerations of Liberty County, GA are included on microfilm reel 162. Since the U.S. Census Office considered only the second enumeration (pages 148 on) as the official count, we deleted the 34 households and 143 individuals initially obtained from the first enumeration.
  3. Our original microfilm reel 923 for Delaware County, NY was incomplete. We obtained a new reel and sampled the missing pages.

Back to Top