Office of the Assistant Secretary for Planning and Evaluation; Linking State Administrative Data

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Office of the Secretary

Office of the Assistant Secretary for Planning and Evaluation;

Linking State Administrative Data

AGENCY: Office of the Assistant Secretary for Planning and Evaluation,

HHS.

ACTION: Request for applications for grants to support State efforts to

link case-level administrative data across multiple low-income

assistance programs.

-----------------------------------------------------------------------

SUMMARY: Recent state efforts to link longitudinal, administrative data

across programs have proven extremely successful. Linked databases have

provided a more thorough understanding of many aspects of both program

participation and the characteristics of individuals who receive

benefits from multiple anti-poverty programs. State-supported efforts

have also provided valuable insight into both inter- and intra-state

variations in program participation. Much of this information would not

have been accessible through national panel data.

While the efforts of individual states have been extremely

valuable, they have been limited to relatively few states. Factors such

as prohibitive cost, lack of necessary staff expertise, and

insufficient time and computational resources have precluded many

interested states from linking their administrative data. Total funding

of up to $200,000 is available to provide one to two interested states

with resources needed to successfully link administrative data and use

it for program management, research and scholarly analysis. It is not

expected that the funding available in this grant will be sufficient

for any state to complete a project that links micro-level

administrative data. Rather, this grant is intended to assist those

states which are interested in linking their administrative data, but

currently lack the resources to successfully complete the project on

their own.

Part I. Linking State Administrative Data

A. Background:

In the last five years, several states have begun assembling

administrative data from income-maintenance and other programs targeted

toward low-income individuals and families for use in policy research

and program evaluation. Most notably, administrative data that has been

linked from a variety of anti-poverty programs has been used to study

characteristics of program participation, multi-service usage, and

caseload dynamics. The results from many of these research initiatives

have provided an extremely useful insight into the characteristics of

program participants, the patterns of multi-service utilization, and

the interactions between multiple programs that provide assistance to

low-income families.

Administrative data also offer more possibilities for in-depth

analysis than do other forms of data, such as national panel data. Many

national studies do not give reliable state-level estimates,

particularly in smaller states with relatively few sampling points. As

a result, it is generally quite difficult to estimate the state-level

effects of national anti-poverty programs. State administrative data

offer the opportunity to study inter- and intra-state comparisons of

government programs, and to examine the extent to which variations in

state anti-poverty programs are successful in serving various client

populations.

Usefulness of Project

The research that has been conducted to date has illustrated the

efficacy in using linked administrative data for research and

evaluation. In the vast majority of states, however, the use of linked

data still remains either untouched or far below what is technically

possible. A study funded by the Department surveyed fifteen states and

determined that, for the majority of the states surveyed, linked

administrative data is a potentially rich source of information about

programs targeted toward low-income populations.

Despite the potential of state administrative data, the

Department's previous findings indicate that linked, state-level

program data still remains a vastly under-utilized source of

information. Many states have both the interest and raw administrative

data necessary to produce longitudinally-linked files at the case or

client level. However, as the process of linking data across programs

and over time is an expensive, iterative process that requires

significant time and expertise, many states lack the capacity to link

their data. Some states lack the computer hardware, software, disk

space, and memory necessary to actually perform the process of linking

data. Other states lack the expertise and staff-time to devote energy

to a research project. Many states face both of these obstacles.

This grant will help the selected states overcome the obstacles

that hinder the process of linking administrative data. For example,

states with limited data- [[Page 28420]] linking experience and

capacity could add the hardware and software needed to link and store

data. States with more experience (such as those which currently

operate linked, research data bases) could use the funds to add

administrative data from additional anti-poverty programs.

Part II. Awardee Responsibilities

Due to the substantial variation among states in the level of

experience and expertise in working with linked administrative data, we

fully expect a wide range of proposals to be submitted. Proposals from

states which currently have linked administrative databases will

obviously differ dramatically from proposals submitted by states with

which have never worked with linked data. Given this, the specific

responsibilities of the awardees may vary. Each state will, however, be

expected to follow the following guidelines:

1. Each applicant must develop the computer systems and technical

capacity necessary to produce longitudinal, linked administrative

micro-level data. The focus of the data may be on cases, households,

clients, filing units, etc., or any combination thereof. For those

applicants which currently have linked data bases, it is expected that

this grant will provide the resources necessary to significantly

enhance their current data systems.

2. Each applicant must link administrative data from at least two

programs that primarily benefit low-income individuals or families. The

states that currently operate linked, administrative research databases

began by focusing on data from the AFDC, Medicaid and Food Stamps

computer systems, largely because these data operating systems for

these programs were fairly compatible due to the interactions between

the programs. States in the early stages of data linking may choose to

focus on these programs, but links between other programs are also

strongly encouraged. Other administrative data that states may choose

to link include: Child welfare and foster care, child support

enforcement, unemployment insurance, vital statistics, disability, SSI

and income tax data. Linkages between these programs are especially

encouraged, as they will likely provide fresh insight into the

interactions over time among these programs.

3. Each applicant must develop the capacity and knowledge necessary

to prepare and standardize data for program management and scholarly

analysis. The data resulting from this grant should be able to support

policy research and program evaluation, and should provide insight into

a variety of policy relevant concerns. Data-sets should support

research into questions concerning (but not limited to) multi-program

participation and usage, interactions between various anti-poverty

programs, caseload dynamics, recidivism, fraud and abuse, and the

demographic, economic and social characteristics of multi-program

participants.

4. In addition to preparing the data in a manner suitable for

program administration and scholarly research, applicants must

demonstrate an ability to actually utilize the data analytically.

Linked administrative data allow for a great variety of analysis. For

example, files linked longitudinally can be studied with event-history

and survivor analysis, methods which are used to understand caseload

dynamics and determine how the sequence of service events affects a

client's outcomes. Additionally, since administrative data typically

have more complete and detailed information than panel data,

administrative data analysis can more accurately assess the demographic

and social characteristics of multi-service users. Administrative data

can also be used to do detailed geographic analysis, which is helpful

in studying whether there are significant variations in service usage

across different administrative regions or across neighborhoods.

It is necessary for applicants to detail exactly how their linked

data can be used for scholarly analysis. States with larger social

service departments may have researchers on staff who possess the

skills necessary to fully explore the data. Other states may wish to

combine their efforts with an academic or policy research organization

with expertise in data analysis. Both of these alternatives, as well as

others, would be acceptable. It is not our intent to limit the

analytical choices of applicants, but rather to ensure that the data

sets created under this grant are used to their full potential.

5. Applicants must obtain written agreements with all state or

county social service departments that will supply the source data. The

agreement should clearly indicate the responsibilities of both the

applicants and the state or county agency, and the willingness of the

parties to work cooperatively. Applicants must also include a plan

which ensures that the resulting linked Data ensure client

confidentiality.

6. Applicants must demonstrate an ongoing commitment to the

project. A principal use of these data is to study current policy

relevant questions about programs for low-income populations. Data for

answering current questions are most useful when they capture current

effects of such programs. Consistent with their on-going commitment to

data linking and analysis, applicants must ensure that both recent

historical data and new case data will be added after the Federal

funding for this project expires.

Part III. Prerequisites, Content of Application, Review Process,

and Evaluation Criteria

A. Prerequisites

Who may apply? We will only accept applications from state

agencies, large urban county agencies, or universities working with

them. This announcement is aimed primarily at states that can link

statewide data bases. Applications will also be considered from large

urban county governments that can clearly demonstrate the ability to

link administrative databases in a way that could provide data of

national policy relevance. University-based research teams that are

working with state agencies to develop linked data bases may apply but

must provide assurances from the state that they are intimately

involved in developing and utilizing the data base for policy purposes.

What data bases? Applicants must clearly demonstrate the ability to

link at least two micro (person, family, or case) files and at least be

in the midst of analyzing data for policy research or evaluation

purposes. Examples of files that have been linked in other situations

are: AFDC, Medicaid, Child Welfare and Foster care, Unemployment

Insurance, Child Support, Individual Income Tax, Vital Statistics, and

Juvenile Courts. At a minimum, linked data bases must allow for at

least three years of longitudinal analysis.

On-going commitment? The state agency responsible for establishing

the linked data system must provide evidence of an on-going commitment

to developing the data base and using it to understand poverty, program

utilization, caseload dynamics, program effectiveness, and other

important aspects of administration of anti-poverty, employment, and

welfare programs. Applicants that do not provide assurances that all

three of these prerequisites will be fulfilled will be unacceptable.

B. Applicant Content

The application shall include the following elements:

1. Abstract

A one page abstract of the project and its

objectives. [[Page 28421]]

2. Goals and Justification for Project

This section will discuss why the agency wishes to undertake the

project and what the short and long-term goals of the project are. The

applicant should discuss the background of what it has been doing to

support linking administrative data bases, the current status of data

base development, and what it expects to accomplish with this project.

It should discuss what analysis will be completed given completion of

the project, the analytical report that will be produced, and what

policy relevance it will have. States should also present their plans,

if any, to produce a public use dataset as a result of this project.

Linking two or more administrative data bases for analytical

purposes is a complicated and difficult endeavor. It often can take

several iterations of refinement to produce a data base that supports

analysis of more than simple descriptive statistics about the caseload.

This section should discuss where the agency is in the evolution of the

linking and embedding policy analysis in the administrative management

of the programs involved. It should contain a discussion of how the

agency will carry on after this funding is exhausted. For applicants

who are not currently linking databases, they should clearly

demonstrate their knowledge of the process, as well as their plans to

obtain the necessary expertise to successfully carry out their proposed

project.

3. Project Design and Approach

In this section, the applicant will discuss what, if any, data are

currently linked, what will be added through this grant, and how it

will be accomplished. This section should describe what variables are

available and will be added, what length of time period is covered,

what kind of data analysis currently can be done, and what analytical

capability will be added by this project. The discussion should make it

clear to the reader what is the structure of the data, what are the

building blocks (individuals, families, households, cases, filing

units, etc.), the universe of state population covered, the types of

variables (demographic, program participation, program dynamics, costs,

etc.) that can be used for analysis. The applicant should also clearly

specify how the micro-level data will be linked and how the

retrospective case files will be assembled. Does a unique identifier

exist that will allow data to be easily linked across programs? If not,

what variable or record-matching technique will be employed? It also

should make clear what information is not available, and the

limitations this poses for policy-relevant analysis.

If applicants are not currently linking any administrative data,

then they should assure reviewers that they have adequate access to at

least three years of recent historical administrative data. Applicants

should also convince reviewers that they have the expertise needed to

complete the project, and also have the commitment to continue linking

administrative data for research, analysis, and program management

purposes.

The treatment of confidentiality and proper disclosure is a very

important issue related to linking data and analyzing it. The

applicants will discuss how they will protect data from improper

disclosure, and how they will facilitate analytical use of sensitive

data. This section will discuss the time table to accomplish this

project. Who will do what, when, and how? It also will discuss what

will be the end product of this project. What sort of report will be

produced? What policy relevance will it have to the state and to DHHS?

4. Organization and Staffing

The application will describe the organization applying for the

grant. If the applicant is a state agency, where does it fit in the

state organization? What are its responsibilities? What are its

capabilities and limitations? How can it assure that this project will

be embedded in the state's policy analysis system?

The applicant will discuss the staffing for the project. Who will

be the project leader? What are the qualifications of the staff and who

will be involved? What are their time commitments to the project and

what other time commitments do they have that might interfere with

successful completion of the project? Personal vita and job

descriptions should be attached as an appendix to the application.

If a university group is involved in the project, the application

will clearly delineate what the responsibilities of the group will be

and how the state agency will exercise control over their work. It will

describe the mechanism (subcontract, etc.) used to procure the

university group services.

5. Budget

This section will include a budget summary and narrative which

describes how the budget supports the research plan. It should show the

financial contribution made or expected by other funding sources, and

the share of total project costs covered by ASPE's grant. It will

discuss how the overall funding level and federal contribution relate

to the successful completion of the project. The actual budget will be

presented on the forms and in accordance with the requirements

discussed in the section entitled ``Components of a Complete

Application.''

6. Commitment of State

Applicants should use this section to completely describe the

resources the state has already committed to the project. If the state

has not yet established support for the project, then applicants should

discuss any future involvement expected of the state. Resources

contributed by the state could include any financial assistance (and

whether it is an outright cash grant or is targeted for a specific

purchase such as computing equipment), allocation of staff or computing

time, technical assistance, and any other relevant contribution.

C. Review Process and Evaluation Criteria

A technical panel of at least three people will review and score

those applications which are submitted by the deadline, and which meet

the screening and prerequisite requirements. The review will be based

on the criteria listed below. The review of the technical proposal and

budget will be used by the Assistant Secretary in making funding

decisions. ASPE reserves the option to discuss the application and the

state agency record of performance with other agencies, Regional Office

staff, and experts who may have information that could assist the

selection process.

The evaluation criteria correspond to the outline for the

development of the Program Narrative Statement of the application.

Although not mandatory, it is strongly recommended that applications be

prepared with the format indicated by this outline.

Selection of the successful applicant(s) will be based on the

technical and financial criteria laid out in this announcement.

Reviewers will determine the strengths and weaknesses of each

application in terms of the evaluation criteria listed below, provide

comments and assign numerical scores. The review panel will prepare a

summary of all applicant scores and strengths/weaknesses and

recommendations and submit it to the Assistant Secretary for Planning

and Evaluation for final decisions on the award.

The point value following each criterion heading indicates the

maximum numerical weight that each section will be given in the review

[[Page 28422]] process. An unacceptable rating on any individual

criterion may render the application unacceptable. Consequently,

applicants should take care to ensure that all criteria are fully

addressed in the applications. Applications will be reviewed as

follows:

(a) Quality of the goals and Project Justification. (See Part B,

Type of Application Requested, Section 2.) (15 points) Applications

will be judged on whether they provide a thoughtful and coherent

discussion of the need for the project and what it will accomplish.

Reviewers will judge applicant's past, current, and future commitment

to linking administrative data for policy analysis, research, and

evaluation. Particular attention will be given to the agency's

commitment to scholarly, policy-relevant work, and their commitment to

producing a public use dataset as a result of this project.

(b) Quality of the project design and approach. (See Part B,

Section 3.) (35 points) Reviewers will judge this section on the basis

of whether the research agenda is scientifically sound and policy

relevant. They will also consider whether the applicant is likely to

make a significant contribution to understanding such important issues

as program utilization and effectiveness, caseload dynamics, types of

clients, and multiple program participation. Applications will be rated

on their plans to conduct policy relevant research and interact with

various levels of government to research and evaluate significant

government initiatives and policies.

Reviewers will assess the completeness of the data bases linked,

population coverage, and the extensiveness of the variables in the data

base. A proposal with more data bases linked will be rated higher than

one with only two program databases, all other factors being constant.

Evidence of data quality control and validity is also extremely

important. Ratings will consider the thoroughness of the discussion of

the database strengths and weaknesses. Reviewers will assess whether

there is appropriate use and protection of sensitive or confidential

data. The type and quality of end product anticipated from this project

will be considered and rated. Finally, reviewers will rate the

feasibility of the workplan and time schedule.

(c) Quality of the staffing proposal and proposed organizational

arrangements. (See Part B, Section 4.) (35 points) Reviewers will judge

applicant's staff on research experience, demonstrated research skills,

public administration experience, and relevant policy-research and

policy-making skills. Ratings may consider references on prior research

projects. Staff time commitments to the project also will be a factor

in the evaluation. Furthermore, reviewers will rate the applicant's

pledge and ability to produce a database capable of supporting policy-

relevant analysis.

Reviewers will evaluate the track record of the lead agency ability

to support scholarly, policy relevant research that can meet the

demands of the academic, research, and policy communities.

If a university group is involved in the project, raters will judge

the administrative relationships between the group and the state agency

and whether the administrative arrangements can assure quality data and

analysis.

(d) Appropriateness of the budget to carry out the planned staffing

and activities. (See Part B, Section 5.) (15 points) Ratings will

consider whether: (a) The budget assures an efficient and effective

allocation of funds to achieve the objectives of this solicitation and

(2) the applicant has appropriate financial commitment from the state

and the university, if one is involved.

State Single Point of Contact (E.O. No. 12372): The Department of

Health and Human Services has determined that this program is not

subject to Executive Order No. 12372, Intergovernmental Review of

Federal Programs, because it is a program that is national in scope and

the only impact on State and local governments would be through

subgrants. Applicants are not required to seek intergovernmental review

of their applications within the constraints of E.O. No. 12372.

Deadline for Submission of Applications: The closing date for

submission of applications under this announcement is July 31, 1995.

Applications must be postmarked or hand-delivered to the application

receipt point no later than 4:30 p.m. on July 31, 1995.

Hand-delivered applications will be accepted Monday through Friday

prior to and on July 31, 1995 during the hours of 9 a.m. to 4:30 p.m.

in the lobby of the Hubert H. Humphrey building located at 200

Independence Avenue, SW., in Washington, DC. When hand-delivering an

application, call 690-8794 from the lobby for pick-up. A staff person

will be available to receive applications.

An application will be considered as meeting the deadline if it is

either: (1) Received at, or hand-delivered to, the mailing address on

or before July 31, 1995, or (2) postmarked before midnight five days

prior to the deadline date July 31, 1995, and received in time to be

considered during the competitive review process (within two weeks of

the deadline date).

When mailing application packages, applicants are strongly advised

to obtain a legibly dated receipt from a commercial carrier (such as

UPS, Federal Express, etc.), or from the U.S. Postal Service as proof

of mailing by the deadline date. If there is a question as to when an

application was mailed, applicants will be asked to provide proof of

mailing by the deadline date. When proof is not provided, an

application will not be considered for funding. Private metered

postmarks are not acceptable as proof of timely mailing.

Applications which do not meet the July 31, 1995, deadline are

considered late applications and will not be considered or reviewed in

the current competition. HHS will send a letter to this effect to each

late applicant.

HHS reserves the right to extend the deadline for all applications

due to acts of God, such as floods, hurricanes or earthquakes; due to

acts of war; if there is widespread disruption of the mail; or if HHS

determines a deadline extension to be in the best of the Government.

However, HHS will not waive or extend the deadline for any applicant

unless the deadline is waived or extended for all applicants.

Applications forms. See section entitled ``Components of a Complete

Application.'' All of these documents must accompany the application

package.

Length of Application. Applications should be brief and concise as

possible, but assure successful communication of the applicant's

proposal to the reviewers. In no case shall an applicant (excluding the

resume appendix and other appropriate attachments) be longer than 25

double-spaced pages; it should neither be unduly elaborate not contain

voluminous supporting documentation.

Disposition of Applications.

1. Approval, disapproval, or deferral. On the basis of the review

of an application, the ASPE will either (a) approve the application in

whole, as revised, or in part for such amount of funds and subject to

such conditions as are deemed necessary or desirable for the initiation

and operation of the data linking project; (b) disapprove the

application; or (c) defer action on the application for such reasons as

lack of funds or a need for further review.

2. Notification of disposition. The ASPE will notify the applicants

of the disposition of their application. A signed notification of award

will be [[Page 28423]] issued to notify the applicant of the approved

application.

Components of a Complete Application. A complete application

consists of the following items in this order:

1. Application for Federal Assistance (Standard Form 424, Revised

4-88);

2. Budget Information--Non-construction Programs (Standard Form

424A, Revised 4-88);

3. Assurances--Non-construction Programs (Standard Form 424B,

Revised 4-88);

4. Table of Contents;

5. Budget Justification for Section B--Budget Categories;

6. Proof of non-profit status, if appropriate;

7. Copy of the applicant's approved indirect cost rate agreement if

necessary;

8. Project Narrative Statement, organized in five sections

addressing the following topics:

(a) Understanding of the Effort,

(b) Project Approach,

(c) Staffing Utilization, Staff Background, and Experience,

(d) Organizational Experience, and

(e) Budget Narrative;

9. Any appendices/attachments;

10. Certification Regarding Drug-Free Work place;

11. Certification Regarding Debarment, Suspension and Other

Responsibility Matters; and

12. Certification and, if necessary, Disclosure Regarding Lobbying;

13. Supplement to Section II--Key Personnel; and

14. Application for Federal Assistance Checklist.

Dated: May 22, 1995.

David T. Ellwood,

Assistant Secretary for Planning and Evaluation.

[FR Doc. 95-13220 Filed 5-30-95; 8:45 am]

BILLING CODE 4151-04-M