This document provides a record of the results of the 2010 Census Coverage Measurement Recall Bias Study. The Recall Bias Study was conducted to obtain initial results to validate the hypothesis about the existence and increase of recall bias with respect to reporting residence around Census Day (April 1, 2010) in operations that happen over the several months following Census Day. This study aims to determine if future research is needed into the influences and reasons for recall bias as it pertains to the timing of the Census Coverage Measurement Person Interview and Person Followup operations. The second part of this report reviews how well administrative data obtained from the United States Postal Service National Change of Address file worked as a tool in identifying households with people who moved around March and April of 2010. Finally, the third part of the report examines some results and characteristics of contacts made using a cellular phone random digit dialing sample compared to a standard landline telephone random digit dialing sample.

The 2010 Census Coverage Measurement was a large, complex survey conducted independently of the census. Two of the five field operations conducted were Person Interview, conducted mid-August to early October 2010, and Person Followup, conducted late January to mid-March 2011. The purpose of the Person Interview was to obtain information about the residents of the sample housing unit at the time of the interview. This included nonmovers and people who had moved into the selected housing unit since Census Day (inmovers). In addition, it collected information about certain persons who moved out of the sample housing unit between Census Day and the time of the interview (outmovers). The late interview in the Person Followup operation collected information to help resolve Census Day residence status, enumeration status, match status, or potential person duplication found when matching census and Person Interview person records.

For the 2010 Census Coverage Measurement operations, both the Person Interview and Person Followup operations were conducted later than in previous post-enumeration surveys. As a review of the 2010 Census Coverage Measurement results and for consideration in planning the timing of the 2020 Census Coverage Measurement operations, we reviewed if this change in timing affected the respondent’s ability to report timing and addresses related to potential moves around Census Day as time between the move event and the interview increases.

The Recall Bias Study collected data in four separate panels in 2010 and 2011. These panels were selected to represent the approximate timing of Census Coverage Measurement operations. The study was conducted by telephone using a Computer Assisted Telephone Interview instrument. The study used the Census Bureau’s 2006 Questionnaire Design Experimental Research Survey telephone data collection questionnaire. This questionnaire collects an independent roster of people currently living or staying in the contacted housing units and asks questions similar to those in the Person Interview and Person Followup operations.

The four independent panels were selected using a dual-frame design. The first frame of the study was a random digit dialing list of telephone numbers containing both landline telephones and cellular phones. The second frame of the study was the Mover Sample and consisted of records that matched between the 2010 Master Address File and an extract of the United States Postal Service National Change of Address file. Each of the four panels consisted of 10,000 housing units. Panel 1 (May 2010) consisted of 10,000 Random Digit Dialing housing units. The 10,000 housing units of Panels 2 (June 2010), 3 (September 2010), and 4 (February 2011) were split between 5,500 Random Digit Dialing cases and 4,500 Mover Sample cases. Comparing the percentage of moves reported for March or April of 2010 in the control panel (Panel 1) to the percentage of March/April 2010 moves reported for the later panels should let us measure the rate of change in recalling and reporting moves as the time of interview moves away from the move date.

The Mover Sample was only implemented in the last three panels. As mentioned previous, the Mover Sample composed of records in the 2010 Census Master Address File that matched an extract of the United States Postal Service National Change of Address file. This extract of the National Change of Address file was dated May 1, 2010 and only contained records that had reported a change of address (moves) in either March or April of 2010 by May 1 2010. While allowing us to do some review of the National Change of Address file, this sample should have also provided greater statistical efficiency on measuring recall bias from a relatively small sample by allowing us to target in sample a larger universe of known movers.