Next / Previous / Shipman's Home Sweet Homepage / Site map

How to read the Christmas Bird Count correction reports

These pages describe corrections to the Christmas Bird Count Database. These errors were committed during Shipman's original data entry, so they still need to be fixed in the Audubon database.

For each discrepancy between the published accounts and the database as entered, there is a small table with four elements:

Here's a small example:

#079-0587 AL: Birmingham

Wrong Context Right
Hairy Woodpecker 4
Downy Woodpecker 7
Downy Woodpecker 68
Red-cockaded Woodpecker 7
Eastern Phoebe 3

This example shows count #587 from the 79th CBC. Just after the record "Hairy Woodpecker 4", the data entry operator's eye mistracked and entered "Downy Woodpecker 7", then got back on track with "Eastern Phoebe 3". But in place of that one Downy Woodpecker record there should be two records, "Downy Woodpecker 68; Red-cockaded Woodpecker 7".

The context records are given so you can be sure you're working on the correct circle. If you don't find the "Hairy Woodpecker 4" record, this probably isn't the right circle.


Audubon's database and duplicate records

When Audubon's original database was loaded from Shipman's version in the late 1990s, any duplicate codes within a circle were kicked out. So if Shipman's database had two records for the same species in a circle, neither will appear, whether their numbers agreed or not.

This means that when the data entry operator's eye looped back and re-entered one or more records in a circle, neither of the records for a given form will appear in the Audubon database.

The way these particular errors are notated in the reports is slightly different. Here's an example:

#080-0460 SC: Columbia

Wrong Context Right
Merlin 5
American Kestrel 32
Northern Bobwhite 21
American Coot 5
American Kestrel 32
Northern Bobwhite 21
American Coot 20
Killdeer 42

Record "merlin5" is correct, but the next three records are duplicates. The third one, "amecoo5", has the wrong count (the 5 from Merlin) because the operator's eye mistracked. The next three records are correct, and then the entry gets back on track with the "killde42" record.

If you go to the Audubon database and look this circle up (it is count code "SC0C", not "SCCO"), you will find the "Merlin 5" and the "Killdeer 42", but no records at all for American Kestrel, Northern Bobwhite, or American Coot, because those were duplicated in Shipman's database.

So, whenever there is more than one record in the bottom of the Context column, all but the last record must be added to the Audubon database.


Audubon's database and "minus one" records

In Audubon's database, records from Shipman's database with a count of -1 (minus one, meaning count unknown) will show up with a count of zero. This generally occurs with count-week records, but there are some records like this from count day. There aren't many of these, maybe a few hundred in the whole database.


Status of Audubon's corrections

At this writing, here is the status of Audubon's database with regard to Shipman's proofreading project.


Next: Errata to 1st Christmas Bird Count database
See also: Christmas Bird Count Database corrections
Site map
John W. Shipman, john@nmt.edu
Last updated: 2011/10/16 02:26:03
URL: http://www.nmt.edu/~shipman/z/cbc/howproof.html