Activity Number:
|
365
- SPEED: Innovations in Survey Sampling Designs: Administrative Data, Record Linkage, Non-Probability Samples, and More
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, July 31, 2018 : 10:30 AM to 11:15 AM
|
Sponsor:
|
Survey Research Methods Section
|
Abstract #332987
|
|
Title:
|
Samples, Unite! Understanding the Consequences of Combining Probability and Non-Probability Samples When Linking Records Is Difficult
|
Author(s):
|
Benjamin Williams*
|
Companies:
|
Southern Methodist University
|
Keywords:
|
Record Linkage;
Non Probability Sampling;
Sampling;
Combining samples;
Biometrics;
Matching
|
Abstract:
|
In this paper, combining a large non-probability sample with a smaller probability sample when it is difficult to correctly link records from the two samples is examined. To estimate the total number of units in the population a new method is proposed that treats the non-probability sample as its own stratum. This method is compared to current techniques under various settings of sample parameters and confidence in the matching process. I also examine the effect that record linkage has on the variance of the estimators. This work is then tested on a real world example: the estimation of total fish removed from the Gulf of Mexico by recreational anglers. NOAA is implementing a new technique for gathering fishing statistics in the Gulf of Mexico that involves captains voluntarily self-reporting catch statistics via electronic tablets alongside a random dockside intercept sample. In order to combine the two data sources it is necessary to combine a probability sample (dockside intercept) with a non-probability sample (self-reports). Accurately matching anglers who both self-report and appear in the dockside sample is challenging, resulting in a useful test case.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2018 program
|