Quality Control Documentation
Download the 254 page Quality Control Documentation (PDF) that covers the assessment of data, issues of quality control, and detailed analysis of matches between database records.
The map is a composite of over 100 Sanborn maps of the Bethlehem area. The Sanborn maps were provided by the National Canal Museum, and were taken from an atlas owned by the Bethlehem Steel Corporation, which is now in the museum's collection. Apparently, Bethlehem Steel made the atlas by binding together maps which were not all from the same year. According to the date of the individual maps, the majority of them were created in the early 1920s. These map images were placed upon a geodatabase of the streets of the Bethlehem area, provided by the Historic Bethlehem Partnership.
The georeferenced points selectable on the map received address data from tables of the associated database.
The Bethlehem area consisted of five separate municipalities, which merged into the City of Bethlehem in 1917. During this time, many street names were changed throughout the area, especially in South Bethlehem. The GIS system will display an address at its location in 1900, even though the Sanborn map may give a different street name.
The Bethlehem Area Public Library supplied microfilm of the Sholes' Directory of the Bethlehems, 1900-1901 from their holdings. The microfilm was digitized, and transcribed into a Microsoft EXCEL spreadsheet. This information was added into a table on a MySQL database. While in the database, the records were changed only to remove abbreviations or correct known spelling mistakes. This was done to facilitate the use of database searches. The digital versions of the pages have been linked to the database records to allow the user to view an image of the primary source of each database record.
The four documents regarding the employees of Bethlehem Steel were provided by the Hagley Museum and Archives, and similarly digitized into a text format and transcribed into corresponding tables on the database. The data in these records have not been altered, although a separate guide to known Bethlehem Steel department codes is being made for the Manufacturing Department Employees (1902). Like the Sanborn directories above, links have been made to images of the primary sources. Please note that these records seem to be solely regarding workers in Bethlehem Steel; upper management is excluded.
The data from the United States Census was generated by finding everyone living in a household of someone mentioned in the above cited Bethlehem Steel employee records. Usually a reference in the census, such as occupation, was sought to confirm that the individual in the census was the same person listed in the company records.
The records in the Sholes' and census tables have the added ability of allowing the user to click on the residential address of a returned record and get a list of links to everyone known living at that address. This feature is very helpful in identifying family members and borders.
Comparing records from different sources
Although the records in the database are within a span of two years and located in the same geographic area, it can be difficult to match an individual in one record with the same person in another. For example, the Sholes' Directory lists nine Charles Smiths (some using a middle initial). Three of these Mr. Smiths list their employer as Bethlehem Steel. The records from the 1901 employees of Machine Shop #2 list one Charles Smith, but as his listed occupation does not match with the three Smiths from the Sholes' Directory who reportedly worked for the Steel it is almost impossible to determine which, if any, of the Charles Smiths in the Sholes' Directory was the one listed in the Bethlehem Steel record.
To help researchers track individuals among the various records, Lehigh University's staff undertook the imposing task of comparing fields in each record to potential matches in corresponding records. The results of this endeavor are currently in the project's Quality Control Documentation (PDF), but are also being placed within fields of their corresponding database tables.