Wikipedia:Reliability of GNIS data/Ramsay Place-Name Card Collection

One source of GNIS data is labelled in GNIS records as the Ramsay Place-Name Card Collection, with the following annotation in the citation field of the record: "A collection of approximately 32,000 place names on filing cards. Three sets exits: one at USGS-GNIS in Reston, Virginia; one in the Western Historical Manuscripts Collection at the University of Missouri-Columbia; one in the State Historical Society of Missouri at the University of Missouri-Columbia. The last set was used for Phase II compilation because it was organized by counties. Names were collected by students writing theses during the period 1929 to 1945. The date on the Phase II entry refers to the year of completion of the thesis that covered a given county. SHS-MO/194"

Source
Robert Lee Ramsay (1880–1953) was a professor at the University of Missouri who organized a bunch of M.A. theses in the 1930s and 1940s to document place names in Missouri. He intended to publish a full dictionary of Missouri place names, but that never got done. Instead what we have are set of 4 x place-name cards in the aforementioned institutions indexing these theses.

The State Historical Society of Missouri, for one, has its place-name card collection available on-line at https://collections.shsmo.org/manuscripts/columbia/c2366. At the bottom of the page are links to the collections for individual counties. Note that the layout of the SHS-MO WWW site has changed since many mass-GNIS-import articles were written, and many external hyperlinks in Wikipedia have been archived through the Wayback Machine rather than corrected to follow the reorganization.

The actual original theses are available on-line directly from the University of Missouri, as single or split into multiple PDF files. Here is an incomplete list:

Example entry
An example of such an entry is which refers to the Reynolds County, Missouri card set, a digitized version of which is at https://collections.shsmo.org/manuscripts/columbia/C2366/reynolds-county. The card itself reads:

This is citing and the "(Parks, CENTERVILLE REFORMER (1904-07), New International Atlas 1930)" part is actually Hamlett's own citation cross-reference list in the thesis itself. In fact, this is one of Ramsay's systematic errors. It is in reality from the later work by the same author,, which reads:

This is Logan Township, Reynolds County, Missouri, and the sources provided later in the dissertation are categorized under headings: Hamlett does not provide a citation for the 1930 New International Atlas but presumably it would have been cited alongside the others in '''Geographical references. Maps and Atlases''' on page 397.
 * Personal interviews:"Parks R.L., Reynolds co. County Collector."
 * Secondary sources. Histories. Newspaper articles:" The Centerville Reformer, Centerville, Missouri, 1904–1907."

Problems
The GNIS records sourced to Ramsay were imported into the GNIS over the years 1988 to 1991. This "more than doubled" the GNIS data-set for Missouri. The problems with them are:
 * The GNIS records have been bulk transcribed and information about the place has been lost. For example: Chitwood, a railroad stop in the original source, and on the card, has become "Populated Place" in the GNIS record (even though the GNIS classificiation "Locale" exists for railroad stations).  The GNIS mass-importers to Wikipedia gave us an "unincorporated community" article as a consequence.
 * The cards mis-cite the theses sometimes.
 * The theses are of variable quality, according to their individual authors.
 * Not every place in a thesis is as well supported as the others.
 * Some places are cited to only a single source; whereas others (such as Chitwood) are cross-referenced to multiple sources.
 * The personal interviews have variable quality sources; some people are postmasters, some are otherwise ordinary citizens, and some simply do not have their credentials given at all. Source "Bert Williams" for "Butler Hill" in  is just "Bert Williams, Knob Lick" for example.
 * Some personal interview citations do not even identify the person, just giving initials and surname for example.
 * None of this is in-depth history/geography/demographics/politics/&c. of a place, just a source for its name. The whole idea was to make a dictionary of names, Ramsay being a founder member of the American Name Society.