Generating analytical reports48 After the data has been collected, processed and entered into the database, the information can be presented in analytical reports. These reports will provide the evidence necessary to support arguments made about who was violated, responsibility for violations, how they relate to policy changes, the roles of particular security structures and officials and many other claims. These reports will be based on statistical findings. There are four main kinds of reports that can be generated from the databases in information management systems: (a) Lists – a “one-way tabulation” – i.e. a list of victims of disappearance that were killed in a particular district of a particular province during a set period of time. Various permutations of the same list can be developed. (b) Case summaries – case summaries put together all the facts for each case in a readable standard format – presenting a narrative description of the case, as well as elements of the case that were coded (i.e. violations, detail of victims, perpetrators). These cases, which draw on narrative that have more detail than others, or on multiple versions of the same incident, are often presented as ‘window’ cases. (c) Ad-hoc query results – this is where the database is able to respond to very particular questions. For example, analysts may find in a narrative a reference to an unnamed person who disappeared in a certain part of the country at a particular date. The analyst can then ask the database for the names of all victims of disappearances who fit those specifications (i.e. disappeared at that time at that place). The results may provide a starting point for subsequent inquiries. Such queries are a way for the information management system to feed questions and ideas back to human rights workers, and can help us identify incongruities between data sources, areas that need additional attention or kinds of violations that seem to be the most important. (d) Counts – statistical outputs, such as (a) the number of disappearances, by month, and location in which they occurred, (b) the number of torture cases, by month, and perpetrating organisation, which will help identify proportional levels of responsibility etc. … “there are many ways to count data. The difficult part in using statistics is not generating tables from the data: this is the mechanical application of a database after repeated data and bad data have been filtered out of the system – also known as ‘cleaning the data’. The hardest part of doing quantitative analysis is defining the proper analytical questions, and designing the appropriate tales that will best support an argument about a given question.”49 These reports can provide powerful evidentiary weight to legal and other fact-finding endeavours. 48 49 This section is based on materials drawn from Who Did What to Whom?, pp. 40–72. Who Did What to Whom?, p. 60. 38