Arctic Geopolitical Configuration
115heliport). Second, Nat_Mil_Exp (mln. 2011 USD)^58 presents national data
on military expenditures. The latter was taken from SIPRI Military
Expenditure Dataset in 2000, 2005, and 2010 for Canada, Denmark,
Finland, Iceland, Norway, Russian Federation, Sweden, and the United
States. Expenditures were recorded in millions of USD constant for 2011.
Finally, Nucl_W (binary)^59 differentiates between nuclear and non-nuclear
Arctic states. This differentiation is based on the Non-Proliferation Treaty
(in force since 1970). According to the Treaty, the United States and
Russia are nuclear states, and Canada, Denmark, Finland, Iceland, Norway
and Sweden are non-nuclear states.
The analysis is run in the Statistica 10 package using menus. Because
“it is up to the researcher to select the right method for his/her specific
application,”^60 all methods of visual representation of clusters in Statistica
10 have been tried on the 2000 data. We aimed at contrasting two
perspectives of Arctic regional development in time: geographical and
geopolitical. Hence the method of amalgamation and distance metrics, as
well as the reference line, must be the same in both cases throughout the
whole period under consideration. The reference line was set up according
to the author’s preliminary knowledge of Arctic geopolitical affairs at a
1.35 value. Such a configuration resulted in three geopolitical clusters.
Apart from several abnormalities,^61 all clusters were stable in time
(membership stability in all groups was not lower than 93 percent)
meaning that Hypothesis_1 was falsified. Since “cluster analysis is
diagnostic rather than definitive in nature,”^62 abnormalities signal that
sixteen variables are sufficient to discover patterns in data but are still not
able to produce mutually exclusive categories. We found only Ward’s
method (which uses variance and sum-of-squares techniques to calculate
distance between clusters) and percent disagreement (which allows
analysis of data from those parts which are categorical in nature) offered
apparent groups with relatively large membership within each group.^63
(^58) <milexdata.sipri.org/files/?file=SIPRI+military+expenditure+database+1988-
2012.xlsx>
(^59) NPT, “Treaty on the Non-Proliferation of Nuclear Weapons.”
(^60) StatSoft Inc., Electronic Statistics Textbook. Cluster Analysis, 1+.
(^61) Geographical clustering: Sakha/Yakutia, Yamal-Nenets; Geopolitical clustering:
Iceland, Faroe Islands.
(^62) Wolfson, Madjd-Sadjadi and James, “Identifying National Types: A Cluster
Analysis of Politics, Economics, and Conflict”, 614.
(^63) Cluster analysis loses meaningfulness if there are many small and low-
membership groups.