Non-Substantive Change Request -- Qualitative Research on Methods for Unduplicating People Suspected of Being Counted More Than Once During a Census

2010 Census Coverage Followup Telephone Operation

Non-substantive Change Request

Qualitative Research on Methods for Unduplicating People Suspected of

Being Counted More Than Once During a Census

OMB Approval Number 0607-0946

September 20, 2010

The Census Bureau requests a non-substantive change to the 2010 Coverage Followup (CFU) Telephone Operation Information Collection (OMB approval number 0607-0946). This request details a follow-on activity to the 2010 CFU.

Research on Unduplicating People Suspected of Being Counted More Than Once During a Census

As the Census Bureau conducts each decennial census, it endeavors to count every resident of the United States at his or her correct location, which is the place where the person lives or sleeps most of the time. Because of complex living situations existing in the United States coupled with a Decennial Census Operational Plan aimed at producing comprehensive counts of population and housing, the Census Bureau sometimes counts people more than once.

A key decision to capture the name and the date of birth data elements via Optical Character Recognition during the data capture of Census 2000 responses gave the Census Bureau the means to examine its data for possible person duplicates at a much earlier point of the census process than it had ever been able to do so in the past. Determining how best to resolve these potential duplicates continues to be a challenge for the Census Bureau.

The Census Bureau’s strategy for resolving potentially duplicate person records involves conducting a followup interview to determine the correct location at which to count the person in question. The Census Bureau must conduct a followup interview both to confirm that the person identified by computer matching is in fact the same person and to gather additional information to correctly apply the census residence rules and to determine where to count the person.

In preparing for the 2010 Census, research began to reveal the need for a followup interview that is tailored to (or targeted for) the resolution of suspected person duplicates. The Census Bureau’s focus during the years leading up to the 2010 Census was on developing a followup interview (that is, the Coverage Followup or CFU) that could comprehensively address a variety of potential coverage problems. While the CFU could sufficiently resolve certain types of problems (for example, discrepancies between household population count and number of people listed on the form, additional people for a household appearing on administrative records but not on its census form, or affirmative answers to whether members of the household live or stay in college housing or a nursing home), mid-decade testing results showed that it failed to prompt respondents to acknowledge or to recognize living situations that have led to the existence of duplicate person records. When the Census Bureau ultimately had to limit the types of cases in the 2010 CFU universe to align with the budget available for the program, followup on potential person duplicates was one type of case it decided to exclude given testing results showing that the CFU handled other types of potential coverage problems more efficiently.

To further aid in the development of optimum methods for resolving suspected instances of person duplication, the Census Bureau requests permission to conduct 300 interviews with individuals who the Census Bureau believes have been counted more than once in the 2010 Census. This data collection will consist of a mix of in-depth qualitative interviews and cognitive interviews. The motivation for this research is twofold. First, the Census Bureau would like to conduct this research to better understand why people are duplicated in a census, in addition to gaining more insight on why the CFU fails to be a successful means for resolving person duplication. Secondly, the Census Bureau wishes to test an instrument developed for interviewers attempting to resolve cases of potential person duplication via a followup interview tailored specifically for this purpose: the Targeted Coverage Followup (TCFU). The knowledge gained from this research will help census planners to design better methods for resolving potential duplicate person records.

For the qualitative interviews, researchers will follow the protocol found in Attachment A and will attempt to collect the following types of information:

• Reasons why the Census Bureau could count respondents more than once in the census;

• Description of respondent living situation (elicited through ethnographic-style probes);

• Respondent willingness to provide information about other place or places stayed, if asked about living situation during an interview – in addition to example questions that would elicit this type of information;

• Distance respondents travel between places that they stay;

• Respondent perception of whether this type of information is “none of the Census Bureau’s business”; and

• Variation in responses, if posing the same questions to others in the household.

For the cognitive interviews, researchers will use the TCFU questionnaire found in Attachment B and the protocol found in Attachment C to capture the respondent’s living situation during the year 2010 (including all of the places that the respondent may have stayed) and to answer the questions listed below. Interviewers will use situational or expansive retrospective probing and other probing as specified in the protocol.

• Does the respondent show confusion or lack of understanding of specific questions or concepts? Which terms or phrases, if any, do respondents feel are confusing or difficult to understand?

• Does the respondent refuse to respond or is he or she unable to respond to any of the TCFU questions?

• Does the respondent have a negative or a positive reaction to the questions?

• Are there any terms or topics that are offensive or inappropriate? Which terms or topics, if any, do respondents believe are offensive or inappropriate?

• Does the questionnaire elicit privacy concerns? Are there questions that are perceived as too sensitive?

• Does the questionnaire elicit confidentiality concerns? Does the respondent figure out that someone else reported them on their census form? If so, does this upset them?

• Does the TCFU interview capture the entire living situation of the respondent, or are there pieces of the respondent’s story that are missing? Does it uncover all addresses and all dates that the respondent stayed at another place during 2010?

• Are there any comprehension problems within the TCFU questionnaire?

The Census Bureau believes that the implementation of this follow-on data collection is a nonsubstantive change to the currently approved 2010 Census CFU Telephone Operation Information Collection because both data collections share the same overall objective. By clarifying initial responses for 2010 Census returns indicative of potential population coverage errors, both data collections aim to improve the accuracy of within household population counts. Furthermore, the sample for this follow-on data collection is a subsample of cases originally approved by the Office and Management and Budget as part of the 2010 CFU Telephone Operation Information Collection universe (that is, initial census returns for which computer matching determined possible duplicate person links), but that were eventually excluded from it to make the operation feasible given the funding available to the program.

Procedures for Data Collection

A contractor will collect the data for this activity using interview questions developed by the Census Bureau. The contractor will have the opportunity to comment on interview protocols, however, the Census Bureau will make the final determination on any input provided by the contractor.

Using a list of potential participants provided by the Census Bureau (that is, initial census returns for which computer matching determined possible duplicate person links), the contractor will develop a detailed plan to recruit participants for the qualitative interviews and the cognitive interviews. The contractor will recruit participants who represent a range of characteristics as outlined by the Census Bureau. For instance, the Census Bureau will require the contractor to contact some households with children who are duplicated and others with adults who are duplicated.

In general, the recruitment plan has two phases. During the first phase, the contractor will send potential participants a letter that provides to them an overview of the data collection activity and that lets them know to expect a telephone call regarding a request for an interview. Researchers believe that giving potential participants advance notice of an interview request has a positive effect on participation rates. In the second phase, the contractor will contact the same potential participants via telephone to request their participation in this data collection. If interested, the contractor will schedule an in-person interview with the participant. The contractor will conduct the interviews in three to four diverse geographic areas. One of the data collection locations will be in or around Washington, D.C. The contractor will pay respondents a maximum of $40 for participating in the study.

The contractor will conduct a total of 300 in-person interviews comprised of the respondent and one interviewer. There are to be 50 semi-structured qualitative interviews and 250 structured cognitive interviews. For the 50 qualitative interviews, the interviewer will have a typed debriefing protocol to assist in directing the conversation. For the 250 cognitive interviews, the interviewer will use a paper questionnaire followed by a typed debriefing protocol to assist in directing the conversation. For all interviews, the interviewer is to find out who completed the census form for the address (or addresses) that interviewers ask about during the interview. Respondents are to remain unaware that they may have been counted more than once in the census.

The contractor will conduct interviews in two phases. The first phase will occur from October 06, 2010 through December 30, 2010 and will consist of 25 qualitative interviews and 50 cognitive interviews. The second phase of interviewing will occur from February 15, 2011 through June 30, 2011and will consist of 25 qualitative interviews and 200 cognitive interviews.

The contractor plans to audio record all interviews upon receiving consent from the respondent to do so. The contractor will obtain approval from the respondent to audio record the interview using the format found in Attachment D. In addition, the contractor will prepare detailed written summaries for each interview and a final report for the Census Bureau that summarizes the data collection. The Census Bureau will report the findings from this data collection in the 2010 Census Program for Evaluations and Experiments report titled, Effectiveness of Unduplication.  

Estimate of Burden Hours

OMB originally approved 1,335,000 burden hours for the 2010 Census CFU Telephone Operation based upon a ten minute interview for an estimated 8,010,000 housing units. The actual number of CFU cases completed was roughly 4.9 million. Therefore, the Census Bureau has a sufficient number of existing burden hours from the original OMB approval to cover this follow-on data collection. The number of burden hours required for it is 450, as indicated in the table below.

Experiment/Evaluation	Total # of Respondents	Estimated Response Time	Estimated Burden Hours
Effectiveness of Unduplication –Qualitative Interviews/ Cognitive Interviews	300	1 hour and 30 minutes	450

If you need additional information, please contact Jason Machowski, 2020 Research and Coordination Branch, Decennial Management Division at (301) 763-4173, Sarah Heimel, Enumeration Methods and Requirements Branch, Decennial Statistical Studies Division at (301) 763-9297, or Ryan King, Enumeration Methods and Requirements Branch, Decennial Statistical Studies Division at (301) 763- 4774.

Attachments

Attachment A: Targeted Coverage Followup (TCFU) Qualitative Interview Protocol

Attachment B: Targeted Coverage Followup (TCFU) Questionnaire

Attachment C: Targeted Coverage Followup (TCFU) Cognitive Interview Protocol

Attachment D: Consent Form For Decennial Census Research