3.2 Data Collection

In step one, standard methods for collecting and recording information are established. There are three common ways that human rights organizations receive information. They are:

3.2.1 - interviewers who elicit statements from deponents, witnesses, victims, or complainants. In-terviews of this kind are usually the most important source of information for human rights organizations.

3.2.2 - researchers who summarize their findings from documentary sources, including newspaper archives, state documents, or other human rights organizations' reports, among many other documentary sources.

3.2.3 - investigators who report the results of their fieldwork to the organization. Investigators' work encompasses a broad range of kinds of evidence, including: brief conversations with people in the field; the findings from an exhumation done by forensic anthropologists; autopsy reports of victims; and records of medical examinations of victims; among many other kinds of information.

There are two key challenges in information collection: avoiding bias and maximizing completeness. The ideas are described in the next two paragraphs.

First, the organization must ensure that all the information that is collected is unbiased. "Unbiased," in this sense, does not mean that the information is collected according to some supposedly objective rules. Rather it means, for example, that all the interviews ask the same questions and can therefore be compared to one another. If some interviewers fails to ask certain questions, those interviews will be biased against finding whatever should have been asked in the omitted questions. If the information is not collected systematically according to commonly-understood rules, then the analysis will show trends that reflect differences in collection techniques rather than real differences in the human rights reality. Avoiding bias means that any given interview may be safely compared to any other interview.

For example, consider a problem that might emerge from biased interviewing. Imagine, for example, that the organization hypothesizes that repression was directed against certain kinds of political activists. To address this hypothesis, the interview questionnaire includes a section on the political activities of victims. Some interviewers, however, feel uncomfortable asking about political affiliation, and so they decide not to ask these questions. When people later in the information management system analyze the questionnaires, the blank political activities sections -- those left empty by the uncomfortable interviewers -- imply that these victims had no affiliations. Note that the organization has no idea whether or not these victims have any political affiliation. However, because the relevant section of the questionnaire is blank, it will seem as if they have no affiliation. The interviews done by the interviewers who asked all the questions are not directly comparable with the interviews done by interviewers who failed to ask everything.

Note that this problem may also emerge in non-statistical uses of the data. If people simply wanted to read the stories of victims with political affiliation, then the stories of people interviewed by the uncomfortable interviewers would be lost. If we further imagine that the uncomfortable interviewers worked in one section of the country (province A) while the more thorough interviewers worked elsewhere (province B), it will seem from the questionnaires that there was more repression against political activists in province B than in province A. Again, this finding reflects differences in interviewing technique, not differences in patterns of repression.

The second key challenge in designing the information collection process is maximizing completeness. That is, the organization must think about exactly what information they will need to report. What do they need to know in order to argue cases, make social scientific analyses, or present materials to the international community? Then they must make sure that they collect that information. Determining exactly what information will be needed can be an arduous process. Information collection is about "how you are going to get what you need." However, it is very difficult to know how to build tools to collect information until you know exactly what information it is that you need. Using the example above, if the organization's leadership knows that to make their argument they must be able to show that repression was directed against particular organizations (e.g., against a particular peasant or student group), then it is critical that the information collection process focus attention on those particular groups. Interviewers therefore need to find out if victims themselves were members of particular organizations, as well as find out if their family members or friends were affiliates of a particular organization, etc. If information is not collected, then it cannot be analyzed in a report.

Organizations designing their information collection process can work toward unbiased and complete findings by creating a detailed description of how information should be collected and recorded. This detailed description of the collection and representation process is called the protocol. For example, the questionnaire and the interview training manual together are the interview protocol. Each of the three sections below discusses how a protocol might be organized for the three kinds of human rights data collection listed at the beginning of the chapter.