|
Many government statistical agencies, and other large-scale statistical organization, are moving from a paradigm based primarily on sample surveys, to one centered on integration of data from multiple sources, including surveys; administrative and commercial records; and passive data collection. For that work, practical development, implementation and evaluation of high-quality methodology will require in-depth assessment of a wide range of issues related to transparency, reproducibility and replicability (TRR). This paper explores some of these issues, with emphasis on four main topics:
(1) practical evaluation of the profiles of quality, risk, cost and stakeholder value attributed to a given set of statistical information products and services;
(2) development of an operational definition of TRR aligned with the profiles identified in (1);
(3) use of TRR to enhance realistic decisions about integration of multiple data sources within resource-constrained environments; and
(4) broader contributions of TRR to strengthen public confidence in, and institutional credibility of, statistical organizations.
|
