Premium
An embedded experiment for targeted non‐response follow‐up in establishment surveys
Author(s) -
Kaputa Stephen J.,
Thompson Katherine Jenny,
Beck Jennifer L.
Publication year - 2019
Publication title -
journal of the royal statistical society: series a (statistics in society)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.103
H-Index - 84
eISSN - 1467-985X
pISSN - 0964-1998
DOI - 10.1111/rssa.12448
Subject(s) - respondent , certification , computer science , unit (ring theory) , quality (philosophy) , computerized adaptive testing , sample (material) , statistics , operations management , operations research , econometrics , psychology , economics , engineering , mathematics , chromatography , political science , law , philosophy , chemistry , mathematics education , management , epistemology
Summary The US Census Bureau is investigating adaptive non‐response follow‐up strategies for single‐unit businesses in the 2017 Economic Census. The paper describes an embedded split‐panel field experiment in the 2015 Annual Survey of Manufactures that tests two adaptive non‐response follow‐up designs. With the first design, non‐responding establishments in the experimental group received a reminder letter either by certified mail (expensive) or standard mail (inexpensive) based on an optimal allocation that assigns a higher proportion of the certified letters to domains that initially have low unit response rates. This targeted allocation procedure ensures that all units receive some form of non‐response follow‐up but saves cost over the current procedure that sends a certified letter to all non‐responding units. The second studied adaptive design restricts the non‐response follow‐up for the probability subsample of non‐respondents selected for the targeted allocation. We compare the quality effects of the two studied adaptive non‐response follow‐up designs, examining effects on response, respondent sample balance and collected data quality.