The five V’s, seven virtues and ten rules of big data engagement for official statistics

摘要

Faced with increasing demands from users of official statistics, declining budgets, increasing data collection costs and coupled with the increasing supply of new data sources, national statistical offices (NSOs) are attracted to using these newer sources for the production of official statistics. However, given the under or over-coverage biases and measurement errors inherent in these sources and the high standard of accuracy required for official statistics, what are the rules of engagement that NSOs would need to take into account before deciding to use these sources? In this paper, against the background of the 5 V’s and 7 virtues of big data that are discussed in the literature, we suggest ten rules of engagement, seven of which are considered as non-negotiables and three as essential for official statistics. We also provide three use cases to illustrate the ten rules.