(4)

Follow-up activities

 

 

Send enumerators to follow up addresses for which no questionnaire is returned

Obtain accurately completed questionnaire (or designation of address as vacant or nonresidential) from every address not responding in a timely manner

Recruit more enumerators than are expected to be needed so that follow-up can be completed in a timely manner; pay higher area-adjusted wages to attract and keep good people

More timely completion of follow-up could increase data quality and facilitate coverage evaluation (benefit)

Push to complete enumeration could increase duplications and other errors (risk)

 

Send smaller, second wave of enumerators to recheck addresses designated as vacant or delete, enumerate late additions to MAF, check “Be Counted” and telephone forms, and enumerate lost or blank forms, so as to improve coverage

Carry out coverage improvement field checks accurately and on a timely basis

Limit rechecks of vacant units to those not classified as such in two previous operations; drop several 1990 operations: local review of preliminary housing counts, field reinterview of questionnaires with minimal information, and parolee and probationer coverage program

Targeted coverage operations could reduce erroneous enumerations (benefit) at small risk of increased omissions

Reduced follow-up for incomplete questionnaires could increase use of computerized statistical imputation techniques (neutral or risk)

(5)

Process the questionnaires through data capture, coverage edit, unduplication, imputation for missing data, and tabulation

Carry out all data processing operations with a high degree of accuracy in a timely manner

Hire contractors to capture the data, using optical character recognition; identify forms with insufficient information about household members for telephone follow-up; unduplicate people and addresses [see (1)]; edit answers and impute values for missing answers; release data products in various formats (e.g., Internet)

More centralization of computer processing and greater use of computerized editing could reduce errors and variability (benefit)

Unduplication efforts could be incomplete (risk)

Greater use of imputation could or could not effectively replace field work (neutral or risk)



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement