1220-0044 Supporting_Statement_Part_A 2018-2020 Final

Telephone Point-of-Purchase Survey (TPOPS)

OMB Number 1220-0044

November 2017

8

SUPPORTING STATEMENT

A. Justification

1. Necessity of the Information Collected

The purpose of this request is to obtain OMB clearance for the 2018-2020 Telephone Point-of-Purchase Survey (TPOPS). The information to be collected represents a continuation of a currently approved collection, OMB control number 1220-0044, which expires on March 31, 2018.

The Census Bureau conducts the ongoing TPOPS for the Bureau of Labor Statistics (BLS) as part of the Consumer Price Index (CPI) program. The purpose of the survey is to develop and maintain a timely list of retail, wholesale, and service establishments at which urban consumers shop for specified items. The list of establishments produced from the survey serves as BLS’s sampling frame used in pricing goods and services in the CPI. The information collected in the survey also provides BLS with basic expenditure estimates that are used to weight unique items that are priced. Without this information, the BLS would not have a statistically accurate list of current establishments visited by consumers, and therefore, could neither collect prices as needed for the CPI nor weight specific items properly without significant measurement error and bias.

The BLS is authorized to collect these data under Title 29, Section 2, of the United States Code (see Attachment B). On September 30, 1976, Congress enacted Public Law 94-439 (H.R. 14232) to appropriate funds for revisions to the CPI. Within this appropriation is financial support for an ongoing Point-of-Purchase Survey. The Census Bureau collects information in the TPOPS for the BLS under the authority of Title 13, United States Code Section 8(b), which allows the Census Bureau to undertake surveys for other agencies (see Attachment C).

The current TPOPS has been conducted via computer-assisted telephone interviewing (CATI) since 1997. Advance letters are sent to every household for which an address can be obtained. Respondents are contacted within a few days of receiving the letter. During the first interview, the respondents’ addresses are collected so that advance letters/postcards¹ can be sent before subsequent scheduled interviews. Follow up letters are also sent to respondents initially unwilling to participate during the 5 week interviewing period. Eligible respondents who did not complete a survey in the first interview receive a postcard. Hard refusals receive a letter during the interviewing period to encourage participation, with an average conversion rate of 12.2%. For the set of letters that may be sent to respondents, see Attachment D.

Generally speaking, the TPOPS interview is divided into three parts:

 The front portion of the interview is designed to identify eligible sample units and to screen-out ineligible units (e.g., businesses). Respondents using mobile phones are asked to verify they are not driving at the time of the interview.

 The middle portion of the interview contains questions about purchases of selected consumer items made by members of the sample unit. If the sample unit incurred expenses for the specific item in question during the specified recall period, then (a) the amount of each expenditure and (b) the name and address of each outlet where the item was purchased is collected.

 The back portion of the instrument is designed to collect demographic characteristics and contains administrative questions for scheduling subsequent interviews.

The instrument specifications, which contain the actual questions asked, are provided in Attachment E.

To address potential coverage bias in the frame resulting from the omission of households in the U.S. exclusively using cellular phone service, a cell phone number frame was added to TPOPS in April 2012. The target distribution of total interviews between the two frames was set at 75% landline and 25% cell phone, which closely reflected the telephone usage of U.S. households at that time. ²¹

BLS has continued to evaluate alternative sources of outlet frame data. BLS continues to find no single source of outlet data exists that meets all requirements of the CPI:

• The sampling frame must target where urban consumers living in the primary sampling units defined in the CPI survey shop.

• In order for the source of outlet information to be useful, the lag between its collection and use in the CPI should be minimized.

• The CPI Pricing Survey requires identifiable independent frames that match its structure of 180 goods and services.

• The source should be unbiased and accurate.

• The data should be safeguarded and comply with BLS confidentiality guidelines and security measures.

In 2012, the BLS evaluated the Census of Retail Trade, Knowledge Networks, A.C. Nielson, and the Center for Medicare and Medicaid Services Survey as potential sources of retail store information for use by the CPI. None of these met the criteria outlined above. ³

Currently, BLS is evaluating using the database of establishments generated by the Quarterly Census of Employment and Wages (QCEW). This is a business register created from state unemployment insurance tax files, and is a listing of nine million businesses, their NAICS code, address information, and data on employment and wages. QCEW data lack two required data elements to use as a frame source for most CPI items: information about what items sold at the establishments (i.e. a concordance between NAICS and the CPI item structure), and information about where consumers who shop at the establishments reside. BLS is investigating possible uses of QCEW data to supplement household survey data and reduce overall respondent burden. For example, the business register might be sufficient as a frame for a small subset of item categories, or could be used to refine TPOPS address data during its processing.

The BLS is also conducting tests to evaluate the inclusion of outlet questions in the Consumer Expenditure Surveys (CE). With the 2018 geographic revision, all but one of CE’s areas are included in the CPI, overcoming the biggest hurdle for its use in outlet frames. BLS plans to source its CPI outlet frame requirements from CE, thereby eliminating TPOPS.

.

2. Needs and Uses

The BLS will use information collected in TPOPS to select establishments for pricing market basket items needed for CPI calculations. Information received from the currently approved TPOPS collection has been used to select new establishments in geographic areas for the transition to the 2018 Revision CPI Sample. Pricing at these establishments was initiated in February 2017 as part of the ongoing Consumer Price Index Commodities and Services Survey, OMB control number 1220-0039. Information received from future TPOPS collection will continue to be used to select new establishments in geographic areas from the 2018 Revision CPI Sample. The new geographic areas are being brought in over the course of several years. The full implementation will be completed in FY2023.

3. Use of Information Technology

The collection of information for TPOPS is fully automated. Under the CATI environment, interviewing in all CPI geographic sampling areas or primary sampling units (PSUs) occurs from three central locations. Specifically, the Census Bureau operates telephone interviewing facilities in Hagerstown, MD; Jeffersonville, IN; and Tucson, AZ. Interviewers dial a telephone number obtained from a queue of randomly selected telephone numbers for each geographic area, and then conduct the interview over the telephone. Responses are keyed directly via the computer and stored electronically by the CATI instrument. Sample units are identified and interviewed once per quarter, over eight consecutive quarters. The split-panel design minimizes the total number of calls that must be made to obtain a sufficient sample size. Increasing the number of interviews per sample unit minimizes the number of sample units needed to obtain a sufficient sample size.

4. Efforts to Identify Duplication

To our knowledge, no other Federal agency is collecting this information.

5. Minimizing Burden to Small Entities

Not applicable. The collection of information in TPOPS involves individuals, not small businesses.

6. Consequences of Less Frequent Collection

The continuous, quarterly collection of these data enables BLS to reflect current consumer purchasing behavior in its sampling algorithms for the CPI. If the BLS did not conduct TPOPS, BLS would not be able to update and replenish the sample of establishments where pricing agents are sent to verify the price of goods in the market basket. The CPI sample would become out of date and unreflective of current prices faced by urban consumers. The consequence to both the Federal and private sectors, which rely on the CPI as the primary indicator of inflation, would be far-reaching and have serious repercussions on Federal government policy and institutions. For example, Federal fiscal and monetary policies would be hampered due to the lack of accurate information on price changes.

If collection were conducted less frequently, the timeliness and accuracy of the CPI would be significantly compromised.

7. Special Circumstances

There are no special circumstances. The Census Bureau will collect these data in a manner consistent with the guidelines in 5 CFR 1320.5. ⁴

8. Federal Register Notice/Consultation Outside the Agency

The BLS received a letter of support from the Bureau of Economic Analysis in response to the Federal Register notice published on August 31, 2017 in 82 FR 41431. No other comments were received.

During the past 2 years, the BLS has consulted with the following persons and continues to do so as specific problems arise:

Ms. Jennifer Epps

Demographic Surveys Division

Bureau of the Census

(301) 763-4281

Mr. Stephen Ash

Statistical Methods Division

Bureau of the Census

(301) 763-4294

Due to the CPI’s high visibility, the data collection methodology used for its construction is under constant scrutiny by individuals and organizations within and outside the U.S. Government. Members of the CPI staff in Washington, D.C. have participated in various economic association meetings to discuss CPI surveys, including methodological and procedural aspects of the data collection process. The BLS Commissioner and Associate Commissioners report on a monthly basis to the Congressional Joint Economic Committee. Past criticisms of the CPI included its inability to incorporate new products and new outlets into the sample in a timely manner and the belief that the outlets in sample were too old. The current design of TPOPS allows the flexibility to add new products and to select outlets on a continuous basis in all sampling areas in a timely manner.

9. Paying Respondents

Cooperation by the respondents to supply data for TPOPS is voluntary and no remuneration, payment, or gift is provided.

10. Assurance of Confidentiality

The Confidential Information Protection and Statistical Efficiency Act of 2002 (CIPSEA) safeguards the confidentiality of individually identifiable information acquired under a pledge of confidentiality for exclusively statistical purposes by controlling access to, and uses made of, such information. CIPSEA includes fines and penalties for any knowing and willful disclosure of individually identifiable information by an officer, employee, or agent of the BLS.

BLS policy on the confidential nature of respondent identifiable information (RII) states that “RII acquired or maintained by the BLS for exclusively statistical purposes and under a pledge of confidentiality shall be treated in a manner that ensures the information will be used only for statistical purposes and will be accessible only to authorized individuals with a need-to-know.”

The Census Bureau performs this work under the authority of 13 USC Section 8 (b).

Respondents for whom an address can be obtained receive an advance letter that assures confidentiality. If an address is unavailable at the beginning of the telephone interview, the interviewer reads the following statement to respondents who did not receive a letter: “This survey is conducted quarterly. It will take about 10 to 20 minutes to complete and is voluntary. Without Office of Management and Budget approval under project 1220-0044, we could not conduct this survey. The BLS and the Census Bureau will use the information you provide for statistical purposes only. We will protect the confidentiality of the information in accordance with the Privacy Act of 1974 and the Confidential Information Protection and Statistical Efficiency Act of 2002. Title 29, Section 2 and Title 13 of the United States Code authorize us to conduct this survey.”

11. Justification for Sensitive Questions

TPOPS does not include any questions of a sensitive nature.

12. Estimate of Respondent Burden

The collection of information for TPOPS occurs in 75 geographic areas, or primary sampling units (PSUs) under the 2018 geographic design.

TPOPS has a quarterly rotating panel design. Once a household has been selected and identified as an eligible unit, it remains in the sample for eight consecutive quarters. The total sample in each PSU is divided into eight panels. During any given quarter, one panel is administered their first interview, one panel is administered their second interview, etc., through their eighth interview.

The sample design outlined above was used to estimate burden hours for the collection of information in the 2018 through 2020 TPOPS surveys. The BLS estimates that it will take approximately 12.65 minutes on average to interview eligible respondents during the 2018-2020 timeframe. An estimate of average interview time is based on collection results from the second quarter of 2015 through the first quarter of 2017; Table 1 reports respondent burden hours by year and a dollar estimate of respondents’ opportunity costs.

Table 1

Estimates of Annual Response Burden 2018-2020
Number of respondents	10,183
Average number of responses	2.94
Total number of responses	29,938
Minutes per response	12.65
Total minutes	378,716
Total hours	6,312
Mean average hourly wage2	$23.86
Total	$150,603

1 – Estimates based on average of 2015Q2-2017Q1, adjusted for new PSUs in 2019Q2.

2 – Based on mean hourly wage from BLS’ Occupational Employment Statistics May 2016.

13. Estimate of Cost Burden

The total annual cost burden to respondents resulting from the collection of information from the landline frame is $0. The cost associated with the cell phone frame is measured by the utilization of minutes and is assumed to be negligible. There are no capital, start-up, operational, maintenance, or service costs required of respondents.

14. Cost to the Federal Government

The annual cost to the Federal Government of collecting, processing, and reviewing the data collected in TPOPS is expected to be about $4.5 million in fiscal year 2018. This figure includes about $4.0 million in costs incurred by the Census Bureau for collecting and processing the production data and other operational costs associated with maintaining the survey. It also includes about $500 thousand in personnel and computer-related costs appropriated by BLS to manage the survey and process the data upon receipt from the Census Bureau.

In the TPOPS, sub-sampling is done to panel one phone numbers which are found to have the lowest response rates in subsequent panels. Starting in FY2013 sub-sampling was expanded to include unknown eligible cases from the first panel.

• Phone numbers where respondent eligibility could not be confirmed (AAPOR codes 300-314 and 320-365) are sub-sampled at a rate of 25% for panel 2.⁵

• Phone numbers called during panel one which never results in an answer (AAPOR codes 380 and 385) are sub-sampled at a rate of 10% for panel 2.

The eight-panel design was implemented to effect a reduction in costs. Extending the number of interviews for survey responders to eight reduces the number of new responders that must be obtained every quarter. Since obtaining new responders is more costly than re-interviewing previous responders, the expectation was that the eight-panel design would help mitigate the cost increases when fully implemented in the first quarter of 2017. Any savings from the eight panel design, however, has not been realized because of decreasing response rates and higher costs per interview.

In response to rising costs to conduct TPOPS, the target number of total interviews was reduced in Q154 from 14,320 per quarter to 12,000 per quarter. This 16% reduction in completed interviews was distributed across PSUs such that disproportionately more interviews are removed from PSUs that on average exceed the outlet requirements for CPI sampling needs. Beginning in Q163 the target number of completed interviews was again reduced to 10,000 per quarter. Outlet frames for the CPI survey are necessarily smaller as a result.

The transition from the 1998 to the 2018 geographic design was designed to be cost neutral. When fully implemented in the third quarter of 2023, the reduction in the number of PSUs from 87 to 75 would result in cost savings if response rates do not decline further. The transition schedule:

Q174: Regular interviewing resumes in the first set of six 2018 PSUs that are new and will continue with regular rotation. Interviewing occurs in 66 PSUs total.

Q182: Interviewing occurs in 60 PSUs, total. Interviewing is suspended in the third set of three PSUs that are new in the 2018 design.

Q192: Interviewing begins in the remaining three PSUs that are new in the 2018 design. Interviewing will resume in six quarters. Interviewing occurs in 63 PSUs, total.

Q194: Interviewing resumes in the second set of six PSUs that are new in the 2018 design and will continue with regular rotation. Interviewing occurs in 69 PSUs, total.

Q204: Interviewing resumes in the third set of three PSUs that are new in the 2018 design and will continue with regular rotation. Interviewing occurs in 72 PSUs, total.

Q214 Interviewing resumes in the fourth set of three PSUs that are new in the 2018 design and will continue with regular rotation. Interviewing occurs in 75 PSUs, total.

Costs of administering the survey will continue to be monitored. Depending on the results of cost savings from sample design changes, the target number of total completed interviews will be increased or decreased appropriately.

15. Changes in Respondent Burden

The three year average annual burden for 2018 to 2020 is 6,312 hours. The burden estimates for this clearance package are based on average interview time and the number of responses per respondent from the second quarter of 2015 through the first quarter of 2017. The TPOPS is currently approved for 11,818 burden hours. The estimate of 6,312 hours for this package is a decrease of 5,506 hours from the current burden.

This decrease is caused by changes in the number of panel-1 respondents and the estimated average number of responses. First, the BLS decreased the target number of interviews about 30 percent since early 2015 because of increasing nonresponse. Second, the BLS projected the number of responses per respondent to increase to 3.25 when the eight panel design was fully implemented. The number of responses per respondent for this clearance package is lower (2.94) for two reasons. This clearance package estimates include the four panel design for the final three new PSUs for the CPI geographic revision that will be collected in the second quarter of 2019 through the first quarter of 2020. Because of the four panel design, the average number of responses per respondent for new PSUs is smaller than the eight panel design. Additionally, the eight panel design was implemented in the second quarter of 2015, and so the average number of responses per respondent is based on a small sample that was able to be tracked across all eight panels.

For more information about response rates that show the increasing number of non-contacts, please see attachment F, TPOPS Non-Response Bias Analysis.

16. Time Schedule for Information Collection and Publication

Results from TPOPS will not be published. Data will be used as the outlet universe for the Commodities and Services Survey for the CPI. Data are delivered to BLS from the Census Bureau approximately seven weeks after the end of each interviewing period. BLS processes the data and selects establishments for pricing during the following 10 to 12 months. The initiation of pricing activities in outlets reported in TPOPS begins approximately 12 to 16 months after the original data collection in TPOPS. This schedule operates continuously, with new interviews beginning each calendar quarter.

17. Request to Not Display Expiration Date

The 2018-2020 TPOPS survey will be conducted via computer-assisted telephone interviewing, there are no paper questionnaires. Instead, respondents are read statements and questions over the telephone. Currently, the OMB clearance number is read to the respondent during the introductory screens (see Attachment E). However, the BLS does not indicate the expiration date of the collection. Research has suggested that long, superfluous introductory statements are not only burdensome, but are likely to result in a refusal in a CATI environment. In an effort to minimize the likelihood of losing an interview, the BLS would like to keep the introductory statement as short as possible. The expiration date of OMB approval will be provided to a respondent upon inquisition during the interview.

The BLS currently sends advance letters to households for which an address can be obtained. Copies of the advance letters are attached (see Attachment D). The letters include a space to indicate how many more times the household is expected to be contacted. This field is updated each quarter for each household. The BLS requests that the expiration date is not printed on our advance letters. This will allow copies of old letters to be retained and used instead of discarded when an expiration date is met.

18. Exceptions to the Certification

There are no exceptions to the “Certificate for Paperwork Reduction Act Submissions” statement in Appendix A.

1 Postcards are sent as reminders to respondents without the survey name to maintain confidentiality.

2 Blumberg, Stephen J., Ph.D., and Luke, Julian V. Wireless Substitution: Early Release of Estimates from the National Health Interview Survey: Division of Health Interview Statistics, National Center for Health Statistics, January-June 2009.

3 Stockburger, Anya. Memo to OMB: Update on the Telephone Point of Purchase Survey cell phone frame planning. June 6, 2012.

4 Electronic Code of Federal Regulations Title 5: Administrative Personnel – Part 1320 Controlling Paperwork Burdens on The Public. 3July 2013. GPO.< http://www.ecfr.gov/cgi-bin/text-idx?c=ecfr&rgn=div5&view=text&node=5:3.0.2.3.9&idno=5>.

5 The American Association for Public Opinion Research. 2011. Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys. 7^th edition. AAPOR