Table 2 This table describes the type of attributes we constructed for each potential location in step 2.

From: Unveiling the geography of historical patents in the United States from 1836 to 1975

Variable Type Comment
Min. Location Integer Candidate Location (CL) first appearance in the document.
Street Dummy 1 if CL is located to the right of the words ‘ST’, ‘AVENUE’, ‘ROAD’, ‘RD’, ‘BLVD’, or ‘AVE’. 0 otherwise.
City Dummy 1 if ‘CITY’ is part of the CL name.
Frequency Integer Number of times the CL was found.
State Dummy 1 if the name of the corresponding state was found. 0 otherwise
Min. Location State Integer State first appearance in the document.
State Distance Set of dummy variables Dummies corresponding to intervals of character distances between CL and the state name. The base category 0 includes also cases where the state name hasn’t been found.
Countries Dummy 1 if the following words appear in the document: ‘GERMANY’, ‘ENGLAND’, ‘FRANCE’, ‘GREAT BRITAIN’, ‘SCOTLAND’, ‘IRELAND’, or ‘CANADA’. O otherwise.
Country Distance Dummy 1 if a country name as specified above is found close to the CL. 0 otherwise.
Cutoff Dummy 1 if the CL or the state has been found after the 50% of the document length. 0 otherwise.
Substring Dummy 1 if the CL is a substring of another CL within the same patent. 0 otherwise. (i.e., York for New York).
Nchar Integer Number of characters of the CL.
Detected Name Dummy 1 if the CL matches any part of the inventor or assignee name. 0 otherwise.
W State Dummy 1 if at least 1 state name has been found for other CL within the same patent document. 0 otherwise.
Rel. Min. Location Countinuous Min Location over the length of the document. Varies between 0 and 1.
City Dummy 1 if the CL was found next to the word ‘CITY’. 0 otherwise.
County Dummy 1 if the CL was found next to the word ‘COUNTY’. 0 otherwise.
COC Dummy 1 if more than one CL of the same county ID co-occur within the same patent document. 0 otherwise.
WX Continuous Index constructed with all the aforementioned variables for ‘competing’ CL within the same patent, as it is usual in spatial settings.