Journal on Policy & Complex Systems Volume 3, Issue 1, Spring 2017 | Page 24

Growing Collaborations
2002 ) within an ABM , produces plausible forecasts of network growth . However , our algorithm does still struggle to forecast the correct amount of link accretion and link decay . Using two periods of data allows for measuring forecast error . The measured forecast error , the amount of over- or underforecasting of link accretion and link decay , could then be used to calibrate our algorithm to a specific network . Measured error from initial model runs could be repurposed as a new parameter that either increases or decreases the amounts of accretion and decay . This more refined version of our algorithm could then be expected to produce more accurate forecasts of network growth and development .
Appendix A : Rules
Attribute Coding
List numbers indicate the assigned ordinal value for each attribute . Homophily in one attribute occurs when two organizations who share a link also share the same value on that attribute .
Capacity Attribute ( Scheinert , 2012 )
• 1 : very small organizations ( all
jurisdictions )
o
Those with expected staffs of around
10 members or less
o
Groups
of
insufficient
size
to
maintain a website or web presence
• 2 :
small
organizations
( all
jurisdictions )
o
Organizations that would staff up to
about 50 people or budgets in thousands to
10s of thousands of dollars , annually
o
Includes
local-level
government
offices
o
Small staff state-level organizations
o
INGO ’ s that only function or can
only function in one location at a time
• 3 : medium sized organizations ( all
jurisdictions ) o Staffs between 50 and 100 or budgets at or exceeding $ 500,000 / year o Includes state-level political parties o State-level agencies o INGO ’ s working in a limited number of countries
• 4 : large domestic organizations and large international organizations ( not government backed ) o Large staffs or budgets exceeding $ 5,000,000 / year o Organizations backed by a single foreign government o INGO ’ s working in many countries
• 5 : large international organizations ( public only ) that draw on resources of multiple governments , allowing for budgets of many millions of dollars per year
Sector Attribute ( Koliba et al ., 2015 )
1 .
Federal Governing Agents
2 .
State / Provincial Governing Agents
3 .
Regional / Geo-governing Agents
4 .
Local / Municipal Governing Agents
5 .
Private ( For-Profit ) Enterprise
6 .
Non-Governmental and Non-Profit
Organizations
7 .
Citizen Agents
8 .
Research Agents
9 .
International Governing Agents
Jurisdiction Attribute ( Koliba et al ., 2015 )
1 .
Vermont
2 .
New York
3 .
Quebec
4 .
US
5 .
Canada
6 .
International
Last resort codes :
7 .
( Aggregate Sector-Identified Agents
( Identifiable Sector but no jurisdiction ))
8 .
( Unclear or otherwise unusable
jurisdiction information )
21