Education data quality in the Third World: A five country study

This study examined the extent to which ministry level decision makers in five developing countries had confidence in the quality of the national level education data available to them, the extent that respondents' judgments of data quality were consistent with more objective analyses of the accuracy of selected national level data, and the reasons to which ministry officials and headmasters attributed inaccuracies in the data. Results indicate that government and school officials assign great importance to numeric data, but that they believe their national data to show error rates ranging from 16 to 40 per cent. There was low correspondence between respondents' perceptions of data quality and the more objective measures of data quality available in three of the countries. The source of error was widely thought to be failure at the school level to make accurate records or returns. Three implications are brought out: the need to provide evidence of quality of data in order to overcome unwillingness to base decisions on data perceived as flawed; the need for a realistic assessment of the level of accuracy required for reasonable decisions to be taken; and the need for training at the school level, together with some incentive for headteachers to make accurate returns.