|Emergency Information Management and Telecommunications - 1st Edition (Department of Humanitarian Affairs/United Nations Disaster Relief Office - Disaster Management Training Programme - United Nations Development Programme , 1997, 62 p.)|
|Part 1: Information management systems|
|Data analysis and information production|
Emergency managers generally obtain data from a wide range of sources, including the findings submitted by assessment or monitoring teams, situation reports from field staff, minutes from meetings with colleagues and counterpart organizations, and media reports. With increasing access to the Internet and the World Wide Web, second-hand sources are today practically without limit; every disaster seems to spawn a new set of Web sites, on-line documents, and Internet-based organizations.
This proliferation of sources has increased the need for data cross-checking and verification. Staff responsible for data analysis must quickly but systematically collate and cross-check data against reports received from other sources. Where discrepancies occur, these should be followed up and reconciled. To be sure, where lives are at risk, decision makers cannot and should not wait for perfectly accurate information before acting. The rule of thumb when lives are threatened is: get the whole picture half right. Once the situation stabilizes (ie, once lives are no longer at risk), there will be time to perform needed cross-checking and verification in order to fine-tune the response.
Analysts should be prepared to rate the informations accuracy according to a scale acceptable to the particular organization: eg: confirmed by other sources; probable; doubtful; improbable; cannot be judged.
Verification of registration data in Liberia
A food relief agency decided to register the beneficiaries in its Monrovia distribution program. Household heads were asked for the names of family members, and household lists were then generated by neighborhood. Soon after, teams were sent house to house to verify the names on the lists. Household heads were asked once again to provide the names of family members. Where responses differed from the original registration data, the names of these imagined beneficiaries were stricken from the distribution lists.