Estimating state level indicators from ICT household surveys in Brazil

There is increasing demand from governments and civil society for more data and indicators for smaller territories and/or with higher frequency to help formulate and evaluate policies. At the same time, there is increasing availability of new data sources created by technologies such as Big Data. Some indicators are regularly produced by official and public statistics organizations, while others are demanded but currently not available, leading to substantial data gaps for many topics of interest. In some cases, existing data may lack the quality needed for its use in decision-making. All this coincides with the trend of diminishing resources available to collect data via traditional surveys, limiting sample sizes and increasing difficulties in producing the desired information. Aiming to meet part of this demand, this article compares four small area estimation methods to produce state-level estimates for two selected indicators using two editions (2014 and 2015) of the Survey on the Use of Information and Communication Technologies in Brazilian Households: ICT Households. The empirical study results indicate that using both temporal aggregation by pooling together samples of two consecutive years, and a composite estimator which borrows strength across larger regions leads to improved estimation of the target state-level ICT indicators.