All indicators presented as part of this research initiative are estimates based on microdata from sample based household surveys. They are not IDB official statistics, are bound to have an implicit sampling error, and are not strictly comparable with estimates from other data sources (census and administrative records). Official indicators should be obtained directly from each country's statistical agency.

Estimates have been subject to a high level of quality control to increase the level of cross-country comparability and to provide information about their statistical accuracy. The calculations are based on the United Nations standard definitions for the MDG indicators, and use the original microdata from the countries' household surveys. In order to enhance the quality of the data the complete estimation process is as transparent as possible, the definitions for all indicators are included in each table and graph, and the STATA do-files used for the computations are available for downloading. Moreover, in order to indicate the margin of error several statistics of significance are provided for all indicators in each period and country.

Comments are welcomed at eqxis-bid@iadb.org

The indicators presented in were calculated from household surveys and thus have a sampling error associated to them.

The statistical standard error for each indicator has been calculated considering the specific sample design of each survey (not by assuming simple random sampling). For this purpose, specific statistical software that considers clustering and stratification was used. However, for some surveys in which variables and/or detailed sample design information were not available, a plausible approximation was made with the microdata on hand.

Within the Tables of , the following statistical indicators of significance are presented:

• standard error,
• 95% confidence intervals,
• variation coefficient, and
• number of observations.

Charts and tables in automatically do not display data if :

• The number of sample observations is less than 30, and/or
• The standard error exceeds 20% of the mean value (i.e.: variation coefficient is greater than 20%).

Average values for indicators presented in , could differ from those calculated with sources other than household surveys. Estimations were calculated trying to make them comparable along time within each country. However, some estimations within countries, could differ from year to year due to changes in survey's questionnaires.

Within , metadata is included on methodological issues on definitions, methods of calculation, relevant facts about sources of information and complete access to the programs for calculating the included indicators.

The definitions used for estimating social indicators in are taken from the publication of the United Nations (2003): “Indicators for Monitoring the Millennium Development Goals. Definitions, Rationale, Concepts and Sources”

is a product developed by the Poverty and Inequality Unit of the Sustainable Development Department, with the collaboration of other Departments and Divisions of the Inter-American Development Bank.

Please contact us at: eqxis-bid@iadb.org