3.2.2. Multidimensional Proximity Data

The dependent variable measured by the multidimensional proximity model in this article is the total intensity of TWC between state actors from 1992 to 2013. To avoid statistical differences caused by territorial changes, state actors do not include historical countries and the newly independent country South Sudan, which ultimately contains 147 national actors.

For the independent variables, the study selects six specific indicators to quantify the proximities among countries, including the distance between capitals of countries, whether both countries are members of the same one international organization involved in TWC, their bilateral trade volume, whether they have the same water organization, whether they had a colonial link, whether they had a common colonizer (Table 1). These indicators are used to reflect the four proximities: geographical, economic, organizational, and colonial proximity.


**Table 1.** Definition of proximity indicators and their data sources.

Notes: Water organization is an international organization that participates in TWC and currently exists. When the actors in a certain water event involve at least one international organization and it effectively plays a cooperative role, we regard it as a "water organization". The statistics include not only its member states but also its observer states or dialogue partners.

Based on the above process, the study established a multidimensional proximity database of global TWC, including one dependent variable matrix and six independent variable matrices, with a total of 151,263 analysis units. The data covered 22 years (from 1992 to 2013) and 147 countries.
