OMB Memo

American Community Survey (ACS) Respondent Comment/Feedback

Request: The Census Bureau received approval under the generic clearance for questionnaire pretesting research (OMB number 0607-0725; approval received February 22, 2018) to complete two rounds of cognitive pretesting on a new respondent comment and feedback question on the ACS. Round 1 was completed in April 2018 by staff of the Center for Survey Measurement (CSM) and Round 2 will be starting in July 2018.

For Round 2 of testing, we will plan to cognitively pretest two versions of the respondent comment and feedback question. In addition, we intend to cognitively pretest interactions between the Hispanic Origin and Race questions with the Ancestry question. The U.S. Census Bureau has contracted with Westat, a statistical survey research corporation headquartered in Rockville, Maryland, to perform cognitive testing on the interaction of these questions, and this proposed research aims to supplement the cognitive results by Westat (see the 2018 American Community Survey Respondent Burden Testing that was approved May 22, 2018).

Purpose:

ACS Respondent Comment/Feedback

The purpose of this testing is to develop a question where respondents believe they can adequately voice their opinion about the ACS and do not feel they need to call the Census Bureau. Multiple versions of the respondent comment and feedback question were tested in Round 1 and a subset of them will be tested in Round 2. The goal is to add a question at the end of the ACS questionnaire where respondents will be able to provide their comments and feedback.

Ancestry

The Census Bureau has proposed to revise the two separate questions on Hispanic origin and race on the 2020 ACS. The questions will be revised to have write-in fields for White and Black or African American; examples will be added for the categories White, Black or African American, and American Indian or Alaska Native; and revisions will be made to the examples for several categories. These changes are reflected in the “Questions Planned for the 2020 Census and the American Community Survey” (U.S. Census Bureau, 2018).¹ The ACS also asks a question later in the survey about ancestry. The expansion of the race question allows the Census Bureau to obtain detailed origin information for all household members, making it possible to consider whether the information already provided in response to the Hispanic origin and race questions may fulfill the data needs the ancestry question is intended to fill.

The 2016 Content Test included a race question very similar to what is proposed for the 2020 ACS. The 2016 Content Test cognitive testing as well as field testing indicated that some respondents found the ancestry question repetitive.² Additionally, analysis of the 2016 Content Test showed that between 72.5 percent and 74.0 percent of respondents provided the same response to the ancestry question that was reported in the race and Hispanic origin questions.³ The Census Bureau is conducting additional research to learn more about the perceived redundancies in the Hispanic origin/race and ancestry questions. This cognitive testing will help to understand whether or not the ancestry question remains to be perceived as repetitive, how respondents interpret the questions, and how information gathered from the Hispanic origin and race questions compares to information gathered from the ancestry question.

Population of Interest: For Round 2, we will still be interviewing respondents from the general population as originally planned. However, we will now focus on recruiting respondents from various races, ethnicities, and ancestries, as described below in the “Sample and Recruitment” section.

Timeline: The first round of cognitive testing was conducted between March and April 2018, as originally intended. The second round of cognitive testing was originally planned to be conducted from June to July 2018 but will now be conducted in July 2018.

Language: As noted in the previous OMB approval, cognitive testing for this research will be in English only.

Method: As mentioned above, Round 2 of this project will be conducted in conjunction with the Ancestry Questions. The number of interviews for Round 2 will still be 15.

Sample and Recruitment: As originally planned, respondents in Round 2 will be recruited from the general population and all interviews will be conducted in the Washington DC area. Respondents will be recruited by CSM staff using the following methods: advertisement postings on Craigslist and broadcast messages distributed to Census headquarters staff. See Enclosure 7 for the recruitment advertisement. Respondents who are interested in participating in testing will be administered the Census Bureau generic pretesting screener questionnaire. We will be looking for respondents who handle the mail in their household and who live with at least one other adult. For testing the race, Hispanic origin, and ancestry question, we will recruit respondents in specific race and ancestry groups, as follows using the additional screener questions included in Enclosure 6:

• Respondents who are non-Hispanic white from any European ancestry

• Respondents who are black, African American, African, or Caribbean

• Respondents who are multiracial

• Respondents who are any of the following: Lebanese, Iranian, Syrian, Egyptian, Brazilian, Guyanese, Belizean, Jamaican, Haitian, Ethiopian, Nigerian, or other Caribbean or African heritage

Interviews will be held at the Census Bureau and at off-site locations convenient to interviewees such as libraries.

Protocol: All respondents will be asked to complete the ACS paper questionnaire when they come in for the cognitive interview. See Enclosure 4 for a copy of the ACS questionnaire. During the interview, respondents will be asked to think aloud when answering specific sections of the questionnaire. Following the completion of the ACS, they will then examine the feedback items for their thoughts and opinions followed by probes about the Hispanic origin, race, and ancestry questions. See Enclosure 3B for the modified Round 2 Respondent Comment/Feedback and Ancestry Questions protocol. The protocol content for the feedback item was shortened to allow for probes on the Hispanic origin, race, and ancestry. (The original Round 2 Protocol with Respondent Comment/Feedback Questions is included in Enclosure 3A.)

Consent: The consent form has not been modified for purposes of this addendum.

Use of Incentive: The amount of money for the incentive will be the same as we will offer $40 as a token for respondent’s participation and to accommodate for expenses like travel and parking.

List of Materials: Below is a list of the documents submitted with the original OMB package clearance. Enclosures that are modified as a result of this addendum are noted.

1. ORIGINAL Round 1 Protocol – Part 1 (Enclosure 1)

2. ORIGINAL Round 1 Protocol – Part 2 (with Respondent Comment/Feedback Questions) (Enclosure 2)

3. ORIGINAL Round 2 Protocol with Respondent Comment/Feedback Questions (Enclosure 3A)

4. MODIFIED Round 2 Protocol Respondent Comment/Feedback and Ancestry Questions (Enclosure 3B)

5. MODIFIED ACS questionnaire (Enclosure 4)

6. MODIFIED Respondent Feedback and Comment Questions (Enclosure 5)

7. MODIFIED Additional Screener Questions (Enclosure 6)

8. MODIFIED Recruitment advertisement text (Enclosure 7)

9. ORIGINAL Consent form (Enclosure 8)

Length of interview: As originally planned for Round 2, we anticipate that the interview will be 60 minutes per respondent. Although we are including the ancestry questions and probes in the protocol, we reduced the number of questions in the Respondent Comment/Feedback protocol to account for this change. As in the original estimate, the burden for interviewing is estimated to be 15 hours.

As planned in Round 2, respondents for this study will be recruited using the pre-approved generic screening questionnaire that will take 10 minutes per person to complete. Originally, we had planned for 2 minutes per person for the additional screener questions. We now plan with the addition of ancestry questions to the screener that it will take an additional 3 minutes to complete for a total of 5 minutes per respondent. Originally, we had planned for the screening of 45 respondents in Round 2. We now anticipate that we will screen 100 respondents since we are recruiting for specific race and ancestry groups. The burden for screening is estimated to be 9 hours. As a result, the revised sum of the estimated burden will be 24 total hours for Round 2 of testing. The prior OMB clearance already allocated 17 hours for this project so this addendum is requesting an additional 7 hours of burden.

	Original estimate			Modified estimate
Category of Respondent	No. of Respondents	Participation Time	Burden	No. of Respondents	Participation Time	Burden
Screening for Round 2	45	2 minutes	2 hours	100	5 minutes	9 hours
Cognitive Interviews – Round 2	15	60 minutes	15 hours	15	60 minutes	15 hours
Totals			17 hours			24 hours

The contact person for questions regarding data and the design of this research project is listed below:

Jonathan Katz
Center for Survey Measurement
U.S. Census Bureau
Washington, D.C. 20233
(301) 763-5956
[email protected]

1 https://www.census.gov/library/publications/2018/dec/planned-questions-2020-acs.html

2 https://www.census.gov/content/dam/Census/library/working-papers/2016/acs/2016_Westat_02.pdf

3 https://www.census.gov/library/working-papers/2017/acs/2017_Harth_01.html