Water-Related Intrastate Conflict and Cooperation (WARICC): A New Event Dataset Thomas Bernauer, Tobias Böhmelt, Halvard Buhaug, Nils Petter Gleditsch, Theresa Tribaldos, Eivind Berg Weibust, and Gerdis Wischnath - Replication Instructions (independent of the statistical software package used): 1. The first table gives an overview of a) the countries in the dataset, b) the downloaded media items from the BBC Monitoring pool of articles, and c) the number of events that we actually coded. While the information in the first two columns can essentially be retrieved from the variables "cname" and "ccode," the third column pertains to downloaded data files that are not part of the replication package due to their size (about 10 GB). Having said that, this third column also includes media reports that are irrelevant for our project and, hence, the forth column that exclusively contains relevant media information is of more interest. This column can be replicated by, e.g., summarizing the "case" variable by country in any stats package. Put differently then, the number of observations per country of the "case" variable as well as the total number of observations in the whole dataset directly reflect the information in the forth column. Finally, the fifth column is calculated by column 4 divided by column 3. 2. The second table gives an overview of the variables of the dataset. This information mirrors the accompanied codebook and the first row of the *.csv data file. 3. The third table gives an overview of the core variable of the dataset, i.e., the WES. In order to replicate that table, simply summarize the basic descriptive statistics of that variable (e.g., import the *.csv file into Stata and type: sum wes). 4. The forth table illustrates the WES in more detail by referring to randomly selected "real cases" of the data. Each case in Table 4 can be traced back to the original data row in the *.csv file by the "Description" column in Table 4, which mirrors the "event" variable in the *.csv file. 5. The fifth table is a combination of cross-tabs that we calculated in Stata 11. Note that two additional variables are necessary for this table: a) the polity2 variable from the 2011 Polity IV dataset (see Marshall and Jaggers 2002) and b) the variable on a country's total population from the World Bank Development Indicators (http://data.worldbank.org/indicator/SP.POP.TOTL). Against this background, one can replicate Table 5 by examining the different cross-tabs with the different combinations (e.g., 646 events in total if "conflict"=1 and an event pertains to a country that scores 7 or higher on the polity2 scale). Apart from these instructions, note that the third row is the first row divided by the second row; and the fifth row is the first row divided by the forth row. 6. The first figure is only a graphical representation of Table 3. Depending on the software used, simply plot a histogram of "wes". 7. In order to replicate the second figure, first, collapse the original data into a format that employs the country-year as unit of analysis. The WES (y-axis) and the year (x-axis) item should be collapsed with an average (unweighted mean) value. Afterwards, both R and Stata allow calculating/plotting the median band. In a final step, we identified those "top 3" countries that were the most "cooperative" and "conflictive" in each year according to the collapsed (mean) value of the WES. 8. For creating the third and forth figure, we identified those events that were either a) cooperative or b) conflictive in Jordan for 1997-2009. These events can be traced back via the following variables: "coop," "conflict," and "ccode." Then, and by using the variables "lat_coordin" and "long_coordin", we plotted the events' geographical location on a basic shape file of Jordan (e.g., from the CShapes package (http://nils.weidmann.ws/projects/cshapes)).