mpn_cow combines interstate trade and membership in intergovernmental
organisations (IGOs) into a single multilevel network of 116 states and 40 IGOs.
It holds two tie layers:
directed, weighted ties among the states, in 2009
undirected ties from states to the 40 IGOs
The mode node attribute separates the "states" from the "IGOs".
The polity2 node attribute is carried over for the states, and the scope
node attribute records whether an IGO is "global" or "regional".
Seven are global: IOMig, OECD, OIC, OPEC, OSCE, UN, and WTO.
Four states appear in the trade data but not in the IGO data:
Bosnia and Herzegovina, Iceland, Lebanon, and Brunei.
Their polity2 value is NA.
data(mpn_cow)#> -- # State trade and IGO membership network ------------------------------------
#> # A labelled, weighted, multilevel network of 116 states and 40 IGOs and 10730
#> trade arcs and 839 membership ties (9958 parallel)
#>
#> -- Nodes
#> # A tibble: 156 x 4
#> label mode polity2 scope
#> <chr> <chr> <dbl> <chr>
#> 1 United States of America states 10 <NA>
#> 2 Canada states 10 <NA>
#> 3 Cuba states -7 <NA>
#> 4 Dominican Republic states 8 <NA>
#> 5 Jamaica states 9 <NA>
#> 6 Trinidad and Tobago states 9 <NA>
#> # i 150 more rows
#>
#> -- Missings
#> # A tibble: 759 x 4
#> from to weight layer
#> <int> <int> <dbl> <chr>
#> 1 31 1 NA trade
#> 2 31 2 NA trade
#> 3 31 3 NA trade
#> 4 41 3 NA trade
#> 5 59 3 NA trade
#> 6 64 3 NA trade
#> # i 753 more rows
#>
#> -- Ties
#> # A tibble: 11,569 x 4
#> from to weight layer
#> <int> <int> <dbl> <chr>
#> 1 2 1 228376 trade
#> 2 4 1 3421. trade
#> 3 5 1 501 trade
#> 4 6 1 5624. trade
#> 5 7 1 178335 trade
#> 6 8 1 3379. trade
#> # i 11,563 more rows
#>The scope attribute recovers a distinction that earlier releases stored,
misleadingly, as a tie weight on the membership relation.
No organisation held both values, so the attribute belongs to the IGOs and not
to their ties. The zeros were never absent ties: the only state with no UN tie
is Taiwan, so reading a zero as an absence would invert the membership of
every global body.
The Correlates of War Trade dataset codes missing trade values as -9.
These 759 cells are recorded as missing rather than as negative trade values,
so they appear in the object's Missings table and not among its ties.
Earlier releases stored them as -9, which made the network read as signed.
Note two current limitations of {manynet}.
A network with exactly two named modes is treated as two-mode, and therefore
as undirected, so as_matrix(mpn_cow) returns a symmetrised trade layer.
to_layer(mpn_cow, "trade") does not recover it either, because it maps
the two ends of each arc to the wrong nodes.
The ties stored in the object are complete and correctly directed.
Until both are resolved, use mpn_cow_trade for any matrix-level analysis of trade.
See stocnet/manynet#170 and stocnet/manynet#171.
Barbieri, Katherine, Omar M. G. Keshk, and Brian Pollins. 2009. “TRADING DATA: Evaluating our Assumptions and Coding Rules.” Conflict Management and Peace Science 26(5): 471-491. doi:10.1177/0738894209343887 .
Pevehouse, Jon C.W., Timothy Nordstrom, Roseanne W McManus, Anne Spencer Jamison. 2020. “Tracking Organizations in the World: The Correlates of War IGO Version 3.0 datasets”. Journal of Peace Research 57(3): 492-503. doi:10.1177/0022343319881175 .
Knoke, David, Mario Diani, James Hollway, and Dimitris C Christopoulos. 2021. Multimodal Political Networks. Cambridge University Press. Cambridge University Press.