OMB Memo

Generic Information Collection Request (ICR):
2020 Census Integrated Communications Campaign (ICC) –2020 Census Tracking Survey Implementation

Request: The U.S. Census Bureau plans to conduct additional research under the Generic Clearance for Internet Nonprobability Panel Pretesting (OMB Control Number 0607-0978). The Census Bureau’s Integrated Partnership and Communications (IPC) program plans to conduct a public opinion tracking survey as part of the 2020 Integrated Communications Campaign (ICC). Results from the field test from September through December (under a previously approved collection) formed baseline of measurement for the 2020 Tracking Survey. In this current submission, we are seeking approval to conduct the main portion of the data collection, which will be weekly from January 2020 through June 2020.

Purpose: The goal of the 2020 Census Tracking Survey is twofold:

• To track US public sentiment concerning matters that may bear upon 2020 Census participation;

• To examine how attitudes and perceptions change during the Census.

Populations of Interest: The Tracking Survey aims to reflect the general U.S. population.

Timeline: January – June 2020.

Language: English and Spanish only

Background: From a previous request, the Census Bureau conducted a field test of this questionnaire monthly from September through December. During this time period, researchers were able to 1) gain an estimate of change for key questions (e.g., intent, awareness, use of federal statistics) between a baseline question wording that had been used in the past and the new wording that will be used for the remainder of the collection; 2) start looking at the data from the probability and nonprobability sources related to national benchmarks; and 3) refine processes for data delivery, processing and visualization for use during the 2020 Census.

Minor changes have been made to the questionnaire, including the following:

1. Single versions of each of the items that had two versions have been selected. In all cases, the newer version of the question was selected after gaining insight into how the data are expected to map to the formerly worded item.

2. A few situations where the web wording and the phone wording were not identical. These are revised to be more consistent between modes.

3. Revision of the national benchmark questions. After reviewing the data and consulting with external experts, the team discovered that two of the benchmark items (NHIS general health and NHANES moderate exercise) were subject to severe context effects. These two items are being dropped in favor of two items on employment from the ACS. In addition, the team will examine the items on marital status and language spoken at home as potential benchmarks.

Methods

Probability Survey: Collection of data under the 2020 Census Tracking Survey shall begin no later than January 2, 2020 and continue through June 30, 2020. Weekly, the contractor will collect and deliver data of 1400 completed cases representing a national randomized probability sample and 2100 cases representing a nonprobability sample.

The probability data are produced through a dual-frame Random Digit Dial (RDD) telephone interviews with a probability sample of US adult residents. This is considered the Primary Tracking Survey. The survey is a stand alone survey. The samples permit analysis of opinion in different demographic subgroups of the population, including, but not limited to race/origin (white, black, Hispanic, Asian), age (18-24;25-44;45-64;65+) and geography (e.g. Census Region). The sampling frames for the study cover landline and cellphones (dual-frame) in all fifty states and the District of Columbia.

In 2020, these data shall allow examination of how attitudes and perceptions change before and during the Census measurement period (mid-March through June, 2020). In addition to delivering data to the Census team, the contractor will analyze data collected and identify relevant findings that would be of use for assessing potential campaign optimization efforts covered by this order. These findings will be shared with the Campaign Optimization Team for further action.

The specifications of the data collection are:

• The data shall reflect a cross-sectional national random probability sample of the U.S. population yielding 1400 completed cases (200/day) weekly in the 2020 time period.

• The data shall include interviews conducted in both English and Spanish.

• The sample shall include cases in all 50 states and the District of Columbia. The Contractor may exclude remote areas of Alaska and Maine from calling (because they are enumerated differently than the remainder of the country).

• The data shall be weighted to match demographic targets from the U.S. Census Bureau including age, sex, region, education, ethnicity, nativity and race.

• For the telephone survey, the weights will be computed for each seven-day rolling period between January 2020 and June 2020, and compute base weights daily between January 2020 and June 2020.

• The Contractor shall use benchmark questions on the survey for the purpose of performing nonresponse bias analyses by comparing weighted and unweighted distributions to the benchmarks. To the extent possible, these analyses will focus on survey estimates and not just on demographic variables.

• The Contractor shall maintain responsibility for respondent privacy and confidentiality protections. The contractor shall provide sample data at the level that the contractor maintains respondent confidentiality. Language shall be added to the interview to notify respondents that their information will not be made available in any way that would personally identify them.

• The Contractor shall provide the information on calculation of the response rate in accord with the Standard Definitions, published by the American Association for Public Opinion Research.

Nonprobability Supplement: Similar to the probability survey described above, the Census Bureau requires the delivery of supplemental nonprobability sampled data which will focus on representativeness of the hard-to-survey populations as needed for campaign optimization. These data shall be collected parallel to the 2020 weekly probability-based data collections for comparison purposes.

The nonprobability supplement shall include the following characteristics:

• Using the same questions as the probability survey, though modified for self-administration and prescreening to apply quotas

• Fully cross-sectional, meaning that no sample participant is re-interviewed throughout the course of the data collection

• Quotas that reflect nationally representative populations across the following characteristics:

  • Age/gender

  • Race and Hispanic Origin

  • Census Region

  • Education

  • Language (English and Spanish)

• Weighting will be based on 3-day rolling average using the following variables:

  • Age/gender

  • Education/gender

  • Census Region

  • Age/Education

  • Race/Education

  • Race/Age

  • Rent/Own

  • Race and Hispanic origin

  • Density

  • Telephone status

• Sample sizes

  • 2100 per week (300/day) in the 2020 time period

  • Reflecting a composition of approximately 15% Hispanic, 12% Black or African American, 11% Asian, 62% White

• Data Quality controls including:

  • Digital Fingerprinting – checking IP address for multiple responses from the same respondent

  • Checks for speeding, straight-lining, and other relevant quality checks and edits (some additional questions may be added into the online survey for the purpose of verifying that the respondent is a real person and not a “bot.”)

• Data may be reported as 3-day rolling averages

Additional Survey Questions: The Contractor shall have the capability of adding up to 3 open-ended and up to 5 close-ended into the survey for one week with 24-hour’s notice. These are included in the burden estimate. OMB will receive the request for additional questions, documented in an excel spreadsheet as an addendum to this request not requiring additional burden hours.

Incentives: There will be no financial incentives for the RDD survey.

For the non-probability web supplement, the contractor will enlist several nonprobability panel providers to recruit and distribute invitations to participate. Panels incentivize their panel members to complete studies via a points-based incentive system. Panel members redeem points earned over time for a variety of prizes or cash incentives. The average range for the points-based incentive value is $0.75-$3.00 for a general population sample, and up to $3-$5 for the hard-to-count sample. All incentives are distributed upon completion via virtual currency in the form of a correlated point value.

Average Length of Interview: 6 minutes on the web, 13 minutes on the phone, based on field test data.

Burden:

	N per month	Minutes per interview
RDD Phone	1400* 26 weeks = 36,400	13
Non-Prob Web	2100* 26 weeks = 54,600	6
Total	91,000 interviews	13,347 hours

Enclosures:

Attachment A: RDD Questionnaire

Attachment B: Web Questionnaire

Attachment C: Rotating Questions Spreadsheet

The contact person for questions regarding data collection and study design is:

Jennifer Hunter Childs

Center for Survey Measurement

U.S. Census Bureau

Washington, D.C. 20233

202-603-4827

[email protected]