OMB Supporting Statement Part A 01-29-09

National Survey of WIC Participants II

OMB Supporting Statement (Part A)

Sheku Kamara

FNS Project Manager

USDA, Food and Nutrition Service

3101 Park Center Drive

Alexandria, VA 22302

Submitted by:

11785 Beltsville Drive

Calverton, MD 20705

Daniel Geller, Ph.D.

Project Director

January 29, 2009

Table of Contents

A. JUSTIFICATION................................................................................................................ 1

1. Circumstances that make the collection of information necessary .............................. 1

2. Purpose and use of the information............................................................................... 2

3. Use of information technology and burden reduction................................................... 2

4. Efforts to identify duplication and use of similar information...................................... 3

5. Impact on small businesses or other small entities........................................................ 3

6. Consequence of collecting the information less frequently.......................................... 4

7. Special circumstances relating to the Guidelines of 5 CFR 1320.5.............................. 4

8. Comments in response to the Federal Register Notice …………................................. 4

9. Explanation of any payment or gift to respondents....................................................... 6

10. Assurance of confidentiality provided to respondents…............................................... 6

11. Justification for sensitive questions............................................................................... 6

12. Estimates of the hour burden including annualized hourly costs.................................. 7

13. Estimate of other total annual cost burden to respondents or record keepers.............. 8

14. Annualized costs to Federal Government..................................................................... 9

15. Explanation for program changes or adjustments......................................................... 9

16. Plans for tabulation and publication and project time schedule.................................... 9

17. Reason(s) display of OMB Expiration Date is inappropriate....................................... 10

18. Exception to certification for Paperwork Reduction Act Submissions......................... 10

B. COLLECTIONS OF INFORMATION EMPLOYING STATISTICAL METHODS ...... 11

1. Respondent universe and any sampling methods.......................................................... 11

2. Procedures for the collection of information ................................................................ 17

3. Methods to maximize response rates and to deal with non-response............................ 21

4. Tests of procedures or methods to be undertaken. ....................................................... 21

5. Individuals consulted on statistical aspects of design ................................................ 22

Appendix A-1 — Macro Confidentiality Pledge

Appendix A-2 — 2009 Household Interview Consent Form

Appendix A-3 — Federal Register Notice

Appendix B — Project-Specific Information Cover Letters & Forms

Appendix C — Data Collection Instrument for State WIC Agencies

Appendix D — Data Collection Instrument for Local WIC Agencies

Appendix E — Data Collection Instrument for WIC Participants

Appendix F — Data Collection Instrument for WIC Denied Applicants

Appendix G — Data Collection Instrument for WIC Terminations (Denied Recertificants)

Appendix H — Callback Script for Non-Agency Respondents

PART A. JUSTIFICATION.

A.1 Circumstances that make the collection of information necessary.

The Improper Payments Information Act (IPIA) of 2002 (Public Law 107-300) requires the Food and Nutrition Service (FNS), USDA, to provide estimates of erroneous payments in the WIC Program, and to identify and report corrective actions the agency is taking to reduce them. These measures are necessary in order to enhance the accuracy and integrity of Federal payments in the WIC Program. To achieve this objective, the IPIA provides an initial framework for Federal agencies to identify the causes and ways of reducing improper payments. Consequent to the IPIA law, guidance issued by the Office of Management and Budget (OMB) in May of 2003 (Memorandum 03-13) requires agencies to: (i) review every Federal program, activity, and dollar to assess risk of significant improper payments; (ii) develop a statistically valid methodology for measuring the extent of improper payments in risk-susceptible Federal programs; (iii) initiate process and internal control improvements to enhance the accuracy and integrity of payments; and (iv) report and assess progress on an annual basis. The William F. Goodling Child Nutrition Reauthorization Act of 1998 (Public Law 105-336), which became effective in 2000, requires WIC agencies to obtain income documentation for participants in order to ensure that the right clients with the appropriate income levels are receiving WIC benefits.

FNS is complying with the WIC IPIA legislative requirement using two methods. The first method measures erroneous payments to vendors through a series of WIC vendor management studies that track vendor overcharges, undercharges, and total improper payment to vendors. The second method measures erroneous payments on the participant side by tracking the costs of erroneous certifications and denials through a series of studies entitled National Survey of WIC Participants. These studies measure improper payments in the WIC Program due to certification errors in terms of both magnitude (dollars) and the rate (% of erroneous certifications).

FNS conducted the first National Survey of WIC Participants in 1998 (OMB Control #: 0584-0484; expiration date: 10/31/2000). The study provided national estimates of case error rate and dollar error for the WIC program in 1998. These national surveys are huge and expensive, and are therefore conducted only once every ten years. To continue meeting the requirements of the IPIA, pending OMBs’ approval, FNS will conduct a second survey, the National Survey of WIC Participants II (NSWP II) in 2008. Because of the large burden and cost associated with these national surveys, FNS will determine subsequent annual error rates and estimates using a model and estimation methods without undertaking a large scale, costly, and burdensome effort each year. As a result, NSWP II includes a provision for the development of a model for updating estimates of improper payments due to certification errors in the WIC program that will allow the agency to calculate error on an annual basis for the next 10 years.

A.2 Purpose and use of the information.

FNS will use the information to identify the amount and source of error and to develop corrective measures to be taken to reduce the amount of error. FNS needs information concerning the type and severity of errors that are occurring in order to develop targeted and effective remedies. If these data are not collected, no assessment of the amount and type of errors can be made nor can corrective actions be developed and implemented.

The 1998 National Survey of WIC Participants (NSWP I, 2001), completed before the requirements of the IPIA were implemented and before WIC agencies were required to obtain income documentation, found various payment errors, as follows:

• The WIC case error rate – the percentage of enrollees certified to receive, but not eligible for, WIC benefits – was measured. Analysis of WIC certification case error yielded an estimate of the error rate of 4.5 percent, with 95 percent confidence that the national WIC income-eligibility error rate falls between the interval of 2.8 and 6.2 percent.

• The dollar error is the amount of total WIC food funds spent on ineligible participants. The estimate of the case error rate yielded estimated annual dollar error of $120 million, out of a total of $2.6 billion in WIC food benefits disbursed in FY98. The 95 percent confidence interval extends from $68.9 to $171.9 million dollars.

• The scope of NSWP I did not include determining sources of error for participants incorrectly denied benefits due to sample limitations. The NSWP II study includes this in the study design.

• The NSWP I study found that the requirement to obtain documentation might be effective as the case error rate at agencies requiring income documentation was 2.9%, compared to 6.4% for agencies not requiring such documentation. However, national estimates of the effectiveness of documentation were not possible.

A.3 Use of information technology and burden reduction.

Automation of Participant Data Collection. In compliance with E-Government Act 2002, to reduce burden to the respondent and improve data quality FNS contractors Macro International will implement the use of computer-assisted data collection technology to gather nearly all of the data from project files, participants and denied applicants. Our field staff will use laptops for abstracting participant file data at local WIC agencies and interviewing participants and applicants. In addition, automated tracking and data monitoring systems will be used to ensure that the data and supporting paper documents are collected and accurate. This approach offers the following advantages:

• More objective data collection. Field staff will apply a consistent set of procedures, questions, and probes. Branching and skip patterns applied by the system will prevent field staff from mistakenly skipping sections, omitting questions, or asking the wrong questions during the participant interview.

• Onsite editing of abstraction and interview data. The computer-assisted data collection process will apply logic, consistency checks, and computational checks on all information provided.

• Monitoring of field staff’s productivity and accuracy. Field data, uploaded daily using secure methods, will be monitored by field supervisors for accuracy to assure high-quality data at a reasonable cost.

Automation of Local WIC Agency Surveys. Macro International will encourage local agencies to use the Internet for completing the survey online. The web survey will be programmed and an access URL established once OMB approval is obtained. This offers the following advantages:

• More accurate and complete data collection. Branching and skip patterns applied by the system will prevent respondents from mistakenly skipping sections or omitting questions. Participants will also have the ability to save their responses and continue the survey at another time. Assigned user IDs and password protection will ensure the integrity of the survey and prevent unauthorized use.

• Online editing. The online data collection process will apply logic, consistency checks, and computational checks on all information provided.

• Monitoring completion. We will know whether or not an agency has started or completed the survey, allowing us to target reminders more effectively.

A.4 Efforts to identify duplication and use of similar information.

There is no duplication of the data to be collected in this study. There is no similar information already available.

A.5 Impact on small businesses or other small entities.

FNS has determined that the requirements for this information collection do not adversely impact small businesses or other small entities. Procedures for completing the surveys and assisting field staff in onsite data abstraction at local WIC agencies have been designed to minimize burden as much as possible. Since the data must be collected in a consistent manner, no special procedures are possible for smaller as opposed to larger entities.

The intent of using computer-assisted technology to gather data is to create simplified data collection process. We have used plain language in the questions and provided detailed explanations to guide the applicant to report accurate information; the end result is improving data quality with fewer errors upfront. This also minimizes the administrative burden placed on all respondents who participate in the study, including small businesses and small entities.

A.6 Consequence of collecting the information less frequently.

If this data collection did not occur, FNS will be in violation of the Improper Payments Information Act (IPIA) of 2002 (Public Law 107-300). The central purpose of the IPIA is to enhance the accuracy and integrity of Federal payments. To achieve this objective, the IPIA provides an initial framework for Federal agencies to identify the causes of, and solutions to, reducing improper payments. In turn, guidance issued by the OMB in May of 2003 (Memorandum 03-13) delineates specific requirements to accomplish this. Because of the agreement with OMB that FNS may develop a model to estimate subsequent annual error, this is the only opportunity to collect the data required by the IPIA.

A.7 Special circumstances relating to the Guidelines of 5 CFR 1320.5.

We will request that State and local WIC agencies submit survey responses and/or information on how data are maintained and stored within a 30-day period. This is necessary because data from these two groups will aid in the collection of data at the next stages (that is, for Participants and Denied Applicants). While these agencies are asked to reply within a 30-day period, they are not required to do so.

A.8 Comments in response to the Federal Register Notice and efforts to consult outside agency.

FNS published a notice April 18, 2008 in the Federal Register Volume 73, Number 76, pages 21096-21097 and provided a 60-day period for public comments.

Three public comments were received regarding this notice. Two of these comments, from different entities within the same state, summarized findings and methods used in two surveys on food security in that state conducted in 2002 and 2007. Both recommended keeping the literacy level of participants in mind when conducting the survey, providing the option to complete the survey in languages other than English, attempting to reach participants at various times of day, and considering in-person surveys for those who cannot be reached by telephone. Foreign language survey instruments will be developed once the survey has been finalized and local agencies sampled. The latter will allow Macro to identify which three languages are most prevalent in the communities selected. Based on WIC demographics, one of the languages is certain to be Spanish.

All of these factors had been included in the survey approach since its inception, and so have been duly addressed. One of the comments also suggested use of on-line surveys because of increasing computer access even among low-income users. Since the survey is designed for oral administration, not written, on-line administration is not applicable. The third comment was from an organization representing a group of food manufacturers promoting a specific food product that is interested in questions addressing participant preferences for specific foods, including those using their product, in the WIC package. These comments did not appear to be relevant to the study.

Macro consulted via telephone with representatives of State WIC agencies in some states on the availability of selected participant data, recordkeeping, data elements, and data formats. All of these are non-sampled States except Michigan, which was consulted before the sample was drawn.

Table A8: State WIC Agency Staff Consulted

State	Name	Phone #
Idaho	Mary Kerbs Jill McGregor	208-332-7205 208-334-5948
Michigan	Diane Revitte, Director of Nutrition Program	517-335-8924
Oklahoma	Sean Brown, Vendor Manager	405-271-4676
Wisconsin	Jerilyn Malliet, Project Manager	608-267-3638
New Mexico	Sid Golden, State Director Marsha Padilla	505-476-8801 505-476-8832

In addition, FNS had shared drafts of the instruments with the National WIC Association which solicited input from three member state and local agency representatives: Patricia Waag (Allegheny County, Pennsylvania Health Department, 412-350-4059); Linnea Sallack (California Department of Health Services, 916-928-8806) and Fran Hawkins (Kentucky Department of Human Services, 502-564-3827). Their comments on (insert what was discussed i.e., survey design and format, questions) were incorporated into the final versions of the instruments.

A.9 Explanation of any payment or gift to respondents.

WIC participants who will be interviewed in their homes after completion of an initial telephone survey will receive $20 each for their time. There will be 1,200 such participants and an estimated 640 denied applicants. A similar incentive level ($20) was offered in NSWP I to achieve a 76.5% response rate. Macro used the $20 amount in its NSWP II pretests and found that the amount remains sufficient to elicit participation in the study. (We note this application is a reinstatement with change, of a previously approved collection for which approval has expired.) 

A.10 Assurance of confidentiality provided to respondents.

All WIC participants and applicants participating in the study are given assurance that information provided is confidential to the extent required by law and since participation is voluntary, there will have no impact on their participation in the WIC program. All Macro International staff will be required to sign a data confidentiality pledge associated specifically with this study. A copy of this pledge is in Appendix A-1. Appendix A-2 contains a copy of the consent form for WIC participants, which also addresses confidentiality of the data. A similar statement is read to telephone participants.

Furthermore, Macro International’s Institutional Review Board will approve the study and review all assurances of confidentiality provided and plans for data and document storage to make sure that the requirements of such assurances will be met. The IRB is registered with the Office of Human Research Protection within the U.S. Department of Health and Human Services (IRB Registration No.: IRB00000954, expiration 07/06/2010) and has been granted a Federal-wide Assurance (FWA00000845, expiration 07/20/2010). The IRB package has not been developed or submitted, although Macro plans to seek approval by April 2009. Once the package is developed, Macro is happy to submit it if so requested.

A.11 Justification for sensitive questions.

Many of the questions in the WIC participant/applicant questionnaire concern household income and certain characteristics of household members, all of which could be considered to be sensitive areas. Because the purpose of the study is to measure error in eligibility determinations, and such determinations are based on household income and certain characteristics (e.g., household size, number of dependents), those questions are absolutely necessary to determine error. Other sensitive questions related to breastfeeding, pregnancies, number of children, and personal nutrition are necessary to assess eligibility and the extent to which participants find the information and support (primarily for breastfeeding and nutrition) helpful.

Before the in-home participant interview, the field interviewer will read aloud the consent form shown on Appendix A-2.

A.12 Estimates of the hour burden including annualized hourly costs.

Five phases of data collection activities with four separate groups of respondents are associated with this study.

Phase I State WIC Agency Survey. Ninety agencies will complete a paper survey. State agencies may choose one or more persons to complete various sections of the survey. The estimated burden is 66 minutes x 90 respondents = total burden of 5,940 minutes or 99 hours. The 23 States selected for sampling of WIC participants will provide data on food issuances and redemptions from existing record systems by WIC subgroups for a defined period of time. This will take 60 minutes for participant data by local agencies, 420 minutes for food issuance data for all participants in a 2 month period, and 210 minutes for redemption data for these same participants for a total of 690 minutes per agency. The estimated burden is 690 minutes x 23 respondents = total burden of 15,870 minutes or 264.5 hours.

Phase II Local WIC Agency Survey. The second respondent group is that of local WIC agencies. For 500 randomly selected local WIC agencies completing a survey of practices and policies the estimated burden is 40 minutes x 500 respondents = total burden of 20,000 minutes or 333.33 hours. A separately sampled group of 80 local agencies selected from the 23 states (there may be overlap between these samples but the amount cannot be determined in advance) will collect daily logs of denied applicant data. The logs will take 30 minutes for each of 20 working days in 1 month plus 75 minutes for set up and returning the data to Macro. The estimated burden is 675 minutes x 80 respondents = total burden of 54,000 minutes or 900 hours.

Phase III Data Collection Participant File Abstraction. Phase III of the data collection has no respondent burden. This phase entails a file abstraction task that the data collector executes on site at the local agencies. The agency is not asked to conduct any recordkeeping or provide any information that is outside the realm of what would normally be accomplished in administering the WIC program.

Phase IV Data Collection Participant Telephone Interview and In-Home Interview. The telephone interview will be conducted with one household member from each of the 2,400 sampled participants on program participation, satisfaction, and participant characteristics. The interview will take 24 minutes. The estimated burden is 24 minutes x 2,400 respondents = total burden of 57,600 minutes or 960 hours. Half of these participants (1,200) will be selected for an in-home audit on their household composition and income and will be asked to locate verification information on their responses. The estimated burden is 36 minutes x 1,200 respondents = total burden of 43,200 minutes or 720 hours. A copy of a paper representation of the computer-assisted interviews for both of these segments is located in Appendix E. Both the telephone interview and in-home interview also will be offered in Spanish. Spanish translations are in progress and will be submitted upon completion.

Phase V Data Collection Denied Applicant/Denied Recertificant Telephone Interview. The telephone interview will be conducted with one household member from each of approximately 400 denied applicants and 400 denied recertificants to determine if they were correctly denied benefits. The interview will take 6 minutes. The estimated burden is 6 minutes x 800 respondents = total burden of 4,800 minutes or 80 hours. A copy of the Project Staff Questionnaire is located in Appendix F. Table A12.1 summarizes the anticipated burden of each data collection component.

Table A12.1. Respondent Burden and Cost Estimate by Data Collection Phase

Phase	Respondent Group	Data Collection Instrument	Estimated Number of Survey Respondents	Frequency of Response	Minutes Per Respondent	Respondent Burden Hours	Estimated Cost1 (Burden hours x hourly rate)
I	State WIC Agency	Paper Survey	90	1	66	99	$4,379
		Provide Data Files	23	1	690	264.5	$9,157
II	Local WIC Agency	Web Survey	500	1	40	333.3	$7,629
		Denied Applicant Data	80	1	675	900	$13,275
III	Participant File Abstraction			No Burden
IV	Participants	Telephone Survey Only	1200	1	24	480	$7,666
		Telephone & In-Home Interview	1200	1	60 (24 phone + 36 in-home)	1200	$5,750
V	Denied Applicants/ Recertifi-cants	Telephone Survey	8002	1	6	80	$639
	TOTALS	--	3070	1	--	3357	$48,495.00

A.13 Estimate of other total annual cost burden to respondents or record keepers.

There are no capital, operating or maintenance costs associated with this information collection.

A.14 Annualized costs to the Federal government.

The data collection and associated project costs for the Erroneous Payments study component cover a period of approximately three years; two thirds of these costs will be incurred in the second year. The values provided in Table A.14.1 are the total costs to execute the study and include: (1) developing instruments, correspondence and administrative forms, (2) development of the sampling plan and sample selection, (3) review and documentation of existing WIC policy and its applicability to the study, (4) development of the evaluation, data collection and analysis plans, (5) systems programming of the data collection software and tracking systems, (6) study pretest, (7) field interviewer training, (8) data collection, (9) data cleaning and processing, (10) data tabulation and analyses, (11) report writing, and (12) overall project management. These costs were estimated by calculating the number of person-hours required to conduct study tasks and adding the other associated direct costs. The costs for developing a model that can be used to predict future error, including training on its use and transfer to FNS will be incurred in the third year of the study.

Table A14.1. Costs by Major Study Components

Component	Cost
Erroneous Payments	$2,366,667.00
Model	$207,485.00
TOTAL COST	$2,573,822.00

A.15 Explanation for program changes or adjustments.

This is a reinstatement of a previously approved information collection that has expired. The program change is a result of 3,361 burden hours.

A.16 Plans for tabulation and publication and project time schedule.

The primary purpose of the study is to determine the type, severity, and cost of errors associated with eligibility determination. This study will produce national estimates of error and be published in a final report. The analysis plan is explained in Part B. The schedule for the entire study showing sample selection, beginning and ending dates of collection of information, participant interviewing, completion of reports, and publication dates is shown on Table A.16.1.

Table A16.1. Data Collection, Reporting, and Model Schedule
Activity	Time schedule
State Sample	January 2008
State WIC Agency (Request for Local Agency List and Participant Figures)	October-November 2008
Local WIC Agency Sample	January 2009
State WIC Agency Survey and Initial Data Request: participation data	May-June 2009
Local WIC Agency Denied Applicant Data	June 2009
Local WIC Agency Survey	May-June 2009
Individual Sample Selection	May-July 2009
Participant Telephone and In-Home Interviews	July-September 2009
Denied Applicant Interviews	July-September 2009
State WIC Agency Second Data Request: redemption data	August 2009
Data Cleaning and Preliminary Analysis	October 2009-January 2010
First Draft/ Final Reports State and Local Agency Erroneous Payments Participant Characteristics Technical	March 2010/July 2010 March 2010/July 2010 March 2010/July 2010 April 2010/August 2010
Estimation Model	February 2010-February 2011

A.17 Reason(s) display of OMB Expiration Date is inappropriate.

Approval for non-display is not being sought.

A.18 Exceptions to certification for Paperwork Reduction Act Submissions.

There are no exceptions.

1 Costs to State and Local WIC Agencies based on Bureau of Labor Statistics occupational data for the following job categories: Managers, All Other ($44.23); Computer Programmers ($34.62); First-Line Supervisors/Managers of Office & Administrative Support Workers ($22.89); and Office & Administrative Support Workers, All Other ($14.75);. Costs to Participants and Denied Applicants based on 2004 WIC Participant Survey data, with an increase of 3%/year for four years to reflect salary increases. (www.bls.gov/oes/current/oes_nat.htm#b00-0000)

2 Sample size is based on the ratio of these groups to participants; therefore the number is subject to change. This represents our best estimate.