U.S. flag

An official website of the United States government

Skip Header


Measuring the Bias in Gross Flows in the Presence of Auto-Correlated Response Errors

Written by:
Working Paper Number SEHSD-WP1987-14 or SIPP-WP-34

Frequently, a categorical variable will be observed at two or more points in time. The interior cells of the cross-classification of two observations are commonly referred to as gross flows or gross changes. Gross flow estimates are potentially of tremendous value in understanding processes. However, estimates are subject to very complex nonsampling errors that have discouraged their use. In fact, the concept may be fundamentally unmeasureable in the sense that any attempt to measure gross flows may change the characteristics of the process. The most serious problems usually present are mismatched observations, observations not missing at random, and misclassification in the observations. In this paper, we focus on misclassifications for dichotomous variables. To the best of our knowledge, prior work on the effect of misclassifications has assumed that misclassifications on the two observations are independent. We have developed a technique that takes advantage of the design of the Survey of Income and Program Participation (SIPP) to estimate the effect in the presence of auto-correlated errors. Even though not all requirements for the technique are currently met by SIPP design, we did try applying it.

Page Last Revised - January 11, 2022
Is this page helpful?
Thumbs Up Image Yes Thumbs Down Image No
NO THANKS
255 characters maximum 255 characters maximum reached
Thank you for your feedback.
Comments or suggestions?

Top

Back to Header