[House Hearing, 115 Congress]
[From the U.S. Government Publishing Office]


    RECOMMENDATIONS OF THE COMMISSION ON EVIDENCE BASED POLICYMAKING

=======================================================================

                                 HEARING

                               BEFORE THE

                         COMMITTEE ON OVERSIGHT
                         AND GOVERNMENT REFORM
                        HOUSE OF REPRESENTATIVES

                     ONE HUNDRED FIFTEENTH CONGRESS

                             FIRST SESSION

                               __________

                           SEPTEMBER 26, 2017

                               __________

                           Serial No. 115-48

                               __________

Printed for the use of the Committee on Oversight and Government Reform


[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]


         Available via the World Wide Web: http://www.fdsys.gov
                       http://oversight.house.gov
                       
                       
                              __________
                               

                    U.S. GOVERNMENT PUBLISHING OFFICE                    
28-506 PDF                  WASHINGTON : 2018                     
          
----------------------------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Publishing Office, 
http://bookstore.gpo.gov. For more information, contact the GPO Customer Contact Center, 
U.S. Government Publishing Office. Phone 202-512-1800, or 866-512-1800 (toll-free). 
E-mail, gpo@custhelp.com.                        
                       
                       
              Committee on Oversight and Government Reform

                  Trey Gowdy, South Carolina, Chairman
John J. Duncan, Jr., Tennessee       Elijah E. Cummings, Maryland, 
Darrell E. Issa, California              Ranking Minority Member
Jim Jordan, Ohio                     Carolyn B. Maloney, New York
Mark Sanford, South Carolina         Eleanor Holmes Norton, District of 
Justin Amash, Michigan                   Columbia
Paul A. Gosar, Arizona               Wm. Lacy Clay, Missouri
Scott DesJarlais, Tennessee          Stephen F. Lynch, Massachusetts
Trey Gowdy, South Carolina           Jim Cooper, Tennessee
Blake Farenthold, Texas              Gerald E. Connolly, Virginia
Virginia Foxx, North Carolina        Robin L. Kelly, Illinois
Thomas Massie, Kentucky              Brenda L. Lawrence, Michigan
Mark Meadows, North Carolina         Bonnie Watson Coleman, New Jersey
Ron DeSantis, Florida                Stacey E. Plaskett, Virgin Islands
Dennis A. Ross, Florida              Val Butler Demings, Florida
Mark Walker, North Carolina          Raja Krishnamoorthi, Illinois
Rod Blum, Iowa                       Jamie Raskin, Maryland
Jody B. Hice, Georgia                Peter Welch, Vermont
Steve Russell, Oklahoma              Matt Cartwright, Pennsylvania
Glenn Grothman, Wisconsin            Mark DeSaulnier, California
Will Hurd, Texas                     Jimmy Gomez, Maryland
Gary J. Palmer, Alabama
James Comer, Kentucky
Paul Mitchell, Michigan
Greg Gianforte, Montana

                     Sheria Clarke, Staff Director
                    William McKenna, General Counsel
   Christina Aizcorbe, Intergovernmental Affairs Subcommittee Staff 
                                Director
                      Katy Rother, Senior Counsel
                       Patrick Hartobey, Counsel
                    Sharon Casey, Deputy Chief Clerk
                 David Rapallo, Minority Staff Director
                            
                            
                            C O N T E N T S

                              ----------                              
                                                                   Page
Hearing held on September 26, 2017...............................     1

                               WITNESSES

Ron Haskins, Ph.D., Co-Chair, Commission on Evidence-Based Policy 
  Making
    Oral Statement...............................................     4
    Written Statement............................................     7
Katharine G. Abraham, Ph.D., Chair, Commission on Evidence-Based 
  Policy Making
    Oral Statement...............................................    10
    Written Statement............................................    12
Latanya Sweeney, Ph.D., Commissioner, Commission on Evidence-
  Based Policy Making
    Oral Statement...............................................    14
    Written Statement............................................    18
Robert Shea, Esq. Commissioner, Commission on Evidence-Based 
  Policy Making
    Oral Statement...............................................    21
    Written Statement............................................    23

                                APPENDIX

Letter of September 26, 2017, to the White House submitted by Mr. 
  Cummings.......................................................    54
Letter of September 26, 2017, to the Department of Health and 
  Human Services submitted by Mr. Cummings.......................    57
September 2017 CEP report titled The Promise of Evidence-Based 
  Policymaking: Report of the Commission on Evidence-Based 
  Policymaking submitted by Mr. Haskins, can be accessed at: 
  https://www.cep.gov/content/dam/cep/report/cep-final-report.pdf    60
Response from Dr. Abraham, Commission on Evidence-Based Policy 
  Making, to Questions for the Record............................    61
Response from Dr. Sweeney, Commission on Evidence-Based Policy 
  Making, to Questions for the Record............................    68

 
    RECOMMENDATIONS OF THE COMMISSION ON EVIDENCE-BASED POLICYMAKING

                              ----------                              


                      Tuesday, September 26, 2017

                  House of Representatives,
              Committee on Oversight and Government Reform,
                                                   Washington, D.C.
    The committee met, pursuant to call, at 10:02 a.m., in Room 
2157, Rayburn House Office Building, Hon. Steve Russell 
presiding.
    Present: Representatives Duncan, Jordan, Sanford, Gosar, 
DesJarlais, Farenthold, Foxx, Massie, Meadows, Ross, Walker, 
Blum, Hice, Russell, Grothman, Hurd, Palmer, Comer, Mitchell, 
Gianforte, Cummings, Maloney, Norton, Lynch, Connolly, Kelly, 
Demings, Krishnamoorthi, Welch, DeSaulnier, and Gomez.
    Mr. Russell. The committee will come to order. Without 
objection, the chair is authorized to declare a recess at any 
time.
    I would like to begin by thanking our Speaker, Paul Ryan, 
for his commitment to improving the way the Federal Government 
uses data to solve problems and improve lives. His bipartisan 
work with Democratic Senator Patty Murray is the reason we are 
able to hold this hearing today.
    According to a 2013 General Accounting report, only 37 
percent of managers who oversee 1,500 different Federal 
programs say their programs had been evaluated in the previous 
5 years. Without sufficient data and analytics from programs 
like these, government agencies are unable to fully assess the 
benefits or impacts of their decisions and programs.
    Although many administrative records and surveys collected 
by government agencies provide significant sources of data, 
these records are typically not shared from one agency to 
another. Oftentimes, agencies either do not have access to or 
are not aware of data that could contribute to evidence 
building.
    To begin addressing data gaps in policymaking and 
administration of taxpayer-funded programs, Speaker Ryan and 
Senator Murray introduced bipartisan, bicameral legislation in 
2016 to create a commission to evaluate the current use and 
availability of data and make recommendations as to how to 
better improve the process.
    On September 7, 2017, the Commission on Evidence-Based 
Policymaking released its report containing 22 recommendations 
for improving the use of data to inform government programs and 
policies. The Commission found the Federal Government lags 
behind the private sector when it comes to managing and 
documenting data that could be used for evidence building.
    Taxpayers expect Federal policymakers to base their 
decisions on well-founded evidence and reason. But according to 
the Commission, there is not enough evidence being produced to 
adequately inform Federal decisionmakers.
    Seeking to balance privacy interests while simultaneously 
creating a process where data could be used and shared for 
statistical purposes, the Commission's recommendations fell 
into four categories: establishing a new Federal entity to 
facilitate data linkages, increasing access to data, 
modernizing privacy protections, increasing and strengthening 
Federal evidence-building capacity.
    Government information management, data sharing, and 
interagency coordination are at the core of this committee's 
legislative jurisdiction. The Commission's recommendations 
touch on the committee's oversight of the implementation of 
FITARA and FOIA and other recent data policy reforms, many of 
which we have worked on in this committee. The recommendations 
also touch on areas of the committee's jurisdiction that are on 
our to-do list, such as the Paperwork Reduction Act, the E-
Government Act, the Privacy Act, and others.
    I want to thank not only the commissioners who are here 
with us today, but all members of the Commission for the time 
and effort devoted to creating this report in a relatively 
short period of time. The information produced in this report 
will help us strengthen our laws and fulfill our desire for a 
more efficient, effective, transparent, and accountable 
government. We look forward to hearing from you today on how we 
can gain better access to the evidence we need to create 
informed and effective policies.
    It's now my honor to recognize the ranking member of the 
committee, Mr. Cummings, for his opening statement.
    Sir.
    Mr. Cummings. Thank you, very much, Mr. Chairman.
    Today we will hear from four members of the Commission on 
Evidence-Based Policymaking. Let me start by saying this.
    Frankly, when most people think of evidence-based 
policymaking, they don't think of the current administration or 
recent actions by Congress. But that is why today's hearing is 
so very, very important. Too often, the American people see 
firsthand how policies that Congress puts in place are 
completely unrelated to the facts.
    Take voter fraud, for example. President Trump claimed that 
3 to 5 million people voted fraudulently in the last election. 
He had no evidence for his claim, none. To their credit, some 
Republicans have even pointed this out, that is that there is 
no fraud. Yet now the American people are being forced to spend 
tax dollars on a new Presidential commission that is trying to 
hunt for evidence to back up the President's unsubstantiated 
claim.
    Take healthcare. The nonpartisan Congressional Budget 
Office says that tens of millions of people will lose their 
health insurance if Republican plans to repeal the Affordable 
Care Act go through. Again, to their credit, a handful of 
Republicans have refused to go along with this. Yet, the 
majority continues to push repeal with no viable alternative 
for the millions of families who depend upon it.
    Take Planned Parenthood. Our committee conducted an 
investigation of allegations against this group. To his credit, 
our former chairman, Representative Chaffetz, reported on 
national television that we found absolutely no evidence to 
support those allegations. Yet, despite our findings, the 
Republican leadership in the House spent millions of taxpayer 
dollars on a new special committee to continue harassing 
Planned Parenthood in search of evidence that never, ever 
existed. Something is awfully wrong with that picture.
    Take immigration. President Trump and congressional 
Republicans have argued that we need to limit the number of 
refugees we accept into our country because they supposedly 
utilize too many government benefits. Yet, the White House 
reportedly squelched an internal report based on economic data 
showing just the opposite, that refugees provide a net economic 
benefit to our Nation of more than $63 billion.
    However, instead of changing their policy to reflect these 
facts, the White House reportedly ordered the report to be 
stripped of all references to the benefits that refugees 
provide. Just took them out, throw them away. That left only a 
biased, inaccurate picture that happens to match their 
political narrative.
    On this topic, we are sending letters this morning to the 
White House and to the Department of Health and Human Services 
seeking documents about their actions on this report.
    I ask unanimous consent that our letters, Mr. Chairman, be 
included in the record for today's hearing.
    Mr. Russell. Without objection.
    Mr. Cummings. Finally, take the most significant long-term 
challenge of our Nation and the world. We face, for decades to 
come, climate change. We have just seen massive devastation 
caused by Hurricanes Maria, Irma, and Harvey which is projected 
to cost hundreds of billions of dollars.
    We've heard for the past 6 years from the nonpartisan 
Government Accountability Office that climate change is one of 
the top risks to the financial and national security of our 
great country. This finding is based on overwhelming scientific 
evidence.
    A few weeks ago, the Miami Herald asked the Republican 
mayor of Miami about Hurricane Irma. And this is how he 
responded, and I quote: This is the time to talk about climate 
change. This is the time that the President and the EPA and 
whoever makes decisions needs to talk about climate change. 
This is a truly, truly, truly poster child for what is to come.
    Of course, he is saying what we all already know and we 
know in our hearts: We need to anchor our public policy on 
sound evidence, not baseless ideology. We cannot pick and 
choose which evidence we recognize and which evidence we ignore 
because of politics.
    Today's hearing goes right to the core mission of our 
committee. Collecting evidence and making policy 
recommendations based on that evidence is exactly what the 
Committee on Oversight and Government Reform was designed to 
do. When we do our jobs correctly, it leads to important 
reforms.
    For example, this committee recently approved a bill, by 
Representative Farenthold and I, introduced to provide Federal 
employees who blow the whistle the right to appeal their cases 
in courts other than the Federal circuit. This important reform 
responds to evidence that the Federal circuit is unfavorable to 
whistleblowers and that expanding the right to appeal to other 
courts will not result in a flood of appeals.
    So I thank our witnesses who are here today. I look forward 
to your testimony.
    And with that, Mr. Chairman, I yield back.
    Mr. Russell. And I thank the ranking member.
    Before we introduce our panel of witnesses this morning, 
it's my honor to also note the presence of a multiparty 
delegation from the Government of Sri Lanka, led by the Speaker 
of Parliament, His Excellency Karu Jayasuriya. We welcome him 
and his distinguished guests.
    Thank you, sir, for your presence with us here today.
    I'm pleased to introduce our witnesses, the following 
commissioners on the Council on Evidence-Based Policymaking: 
Dr. Ron Haskins, the Co-Chair of the Commission; Dr. Katherine 
Abraham, the Chair of the Commission; Dr. Latanya Sweeney; and 
Mr. Robert Shea.
    Welcome to you all.
    Pursuant to committee rules, all witnesses will be sworn in 
before they testify. Please rise and raise your right hand.
    Do you solemnly swear or affirm the testimony you are about 
to give is the truth, the whole truth, and nothing but the 
truth, so help you God?
    The record will reflect all witnesses answered in the 
affirmative.
    Please be seated.
    In order to allow time for discussion, please limit your 
testimony to 5 minutes. Your entire written statement will be 
made part of the record. And I would also remind you about the 
microphones, and get them close, and also remember to turn on 
the button.
    Thank you, sir. Please, Dr. Haskins.

                       WITNESS STATEMENTS

                    STATEMENT OF RON HASKINS

    Mr. Haskins. Thank you, Mr. Russell and Ranking Member 
Cummings and other members of the committee. I'm very pleased 
to be here today in my role as the co-chair of the Commission 
on Evidence-Based Policymaking. I'm a senior fellow at 
Brookings, and it made participating in the Commission much 
better than our members from California who had to keep coming 
back and forth.
    For most of my professional career I have been focused on 
the importance of generating evidence, especially program 
evaluation evidence, to support policymaking. So when Speaker 
Ryan appointed me to serve as co-chair of the Commission, I 
seized the opportunity.
    There are many issues in our country today that generate 
conflicting views. Some have already been expressed this 
morning. So we are very pleased that the need for more and 
better evidence--and the Commission strategy for getting there 
in a privacy protective way--was unanimously approved by the 
full Commission. All 15 members voted for the final report and 
all 22 recommendations that we're discussing before the 
committee today.
    I request that the final report of the Commission be 
entered into the record.
    Mr. Russell. Without objection.
    Mr. Haskins. The Commission was established by the 
bipartisan Evidence-Based Policymaking Commission Act of 2016, 
which was jointly sponsored by Speaker Ryan and Senator Patty 
Murray and signed into law in March 2016. The legislation 
directed the appointment of 15 commissioners with a broad range 
of expertise, including academic researchers, data experts, 
administrators, and experts in computer science, data privacy, 
and privacy law. The Commission was provided just over 1 year 
to study and develop a strategy for strengthening government's 
evidence-building and policymaking efforts.
    Decisionmakers depend on having reliable and timely 
information to guide their examination of how current programs 
and policies are working and how they could be improved. In 
establishing the Commission, Congress rightly acknowledged that 
today too little evidence is produced to meet this need.
    Most of the Nation's social programs produce modest or no 
impacts. Let me say that again. Most of the Nation's social 
programs produce modest or no impacts on the problems they were 
meant to address. Wisely, in recent years Congress has been 
asking for and paying for careful evaluation of some programs. 
But we still don't know enough about the effectiveness of the 
many of the Nation's programs.
    To help address this gap, the Commission was charged with 
developing a strategy for increasing the availability and the 
use of data to build evidence about government programs while 
also protecting privacy and confidentiality. We took both of 
these charges seriously.
    We wanted to make sure that our recommendations were rooted 
in, well, evidence. So we conducted an exhaustive fact-finding 
effort before launching into serious discussion of how to 
improve evidence building. This fact-finding phase extended for 
8 months and included seven public meetings, three public 
hearings, a request for comments through the Federal Register, 
and a survey of Federal offices that generate or use evidence. 
In all, we received input from more than 500 individuals or 
organizations.
    Our report includes 22 recommendations designed to address 
the barriers to having more evidence available. The 
recommendations fall into three broad categories. First, 
improving data access, data access for evidence-building 
projects. Second, modernizing and strengthening the privacy 
protections for data used in evidence building. And, third, 
strengthening the Federal Government's capacity for evidence 
building.
    I'm delighted to be joined on the panel today by my fellow 
commissioners, who will provide a brief overview of the 
recommendations that we developed under each of these three 
categories. First, Katharine Abraham, the chair of the 
Commission, will highlight recommendations that relate to 
streamlining and improving data access for evidence building. 
Second, Latanya Sweeney will describe the kinds of increased 
privacy protections and transparency that the implementation of 
our recommendations would yield. And, third, Robert Shea will 
discuss the set of recommendations related to strengthening the 
Federal Government's capacity for evidence building.
    Their testimony will show that the members of the 
Commission share a vision in which rigorous evidence is created 
efficiently, as a routine part of government operations, and 
used to construct effective policy.
    Finally, we hope our recommendations will be implemented as 
quickly as possible, and we look forward to partnering with the 
Congress, this committee included, and the administration to 
advance the recommendations of the Commission and to achieve a 
future built on evidence-based policymaking.
    Thank you, Mr. Chairman.
    [Prepared statement of Mr. Haskins follows:]
    [GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Russell. Thank you, Dr. Haskins.
    And the chair is now privileged to recognize Dr. Abraham 
for 5 minutes.

               STATEMENT OF KATHARINE G. ABRAHAM

    Ms. Abraham. Thank you, Mr. Russell, Ranking Member 
Cummings, and honored members of the committee. I very much 
appreciate the opportunity to speak with you today about the 
Commission's report.
    I will focus on our recommendations related to secure 
access to confidential data for evidence building. This is an 
area in which I have some personal experience, having served 
two 4-year terms as the Commissioner of Labor Statistics in the 
Department of Labor.
    I know that this committee has taken a particular interest 
in making sure that open data from the Federal Government are 
available to the public. The Commission, on the other hand, 
focused mainly on data that are not readily accessible for 
evidence building because of confidentiality concerns and legal 
restrictions.
    The open data initiative has done a great deal to make 
government data more available for evidence building, and I 
know all of us applaud the work this committee has done to make 
that possible. Our charge was somewhat different, to figure out 
how to harness the power of data that can't be made publicly 
available.
    One of the central charges in the statute that created the 
Commission was for us to evaluate if and how to create a 
clearinghouse for program and survey data to support Federal 
program evaluation and policy. We had to interpret what the 
statute meant by data clearinghouse. We understood that term to 
mean a data warehouse where large amounts of data would be 
brought together, linked, and retained to be available for 
evidence-building purposes.
    That was an idea that we rejected. We were concerned that 
such a clearinghouse would create substantial risks for 
privacy. Instead, what we have recommended is the creation of a 
service to facilitate access to data and linking of data for 
specific projects. The data would be brought in, linked for the 
specific project, identifiers would be removed. Researchers, 
analysts, would work with the data. When the project was done, 
the data set would be destroyed.
    This service, which we have notionally named the National 
Secure Data Service, would build on existing resources within 
government to facilitate secure access to data for analysts 
inside and outside of government for evidence-building 
purposes, especially to data sets created by linking 
information from multiple agencies.
    During its fact-finding phase, the Commission heard about 
several examples of exciting research done using confidential 
data that has generated valuable information for designing and 
carrying out programs and policies, such as the path-breaking 
research of Stanford University Professor Raj Chetty and his 
colleagues on social mobility.
    Too often, however, we found legal and bureaucratic 
barriers to accessing data have prevented researchers from 
studying important policy questions. Surmounting these barriers 
is especially difficult when the researcher seeks to access 
data from multiple jurisdictions or agencies.
    In its review of applicable laws, the Commission found 
considerable variation in provisions governing data 
confidentiality and permissible uses of data. The laws that 
authorize statistical agencies, for example, include varying 
restrictions on who can access data that has been collected and 
for what purposes.
    Many program agencies' authorizing statutes do not address 
data confidentiality and the use of data for evidence building 
at all. Other program agencies' laws establish narrow standards 
for the acceptable use of administrative data. For example, 
Title 26 of the U.S. Code generally limits the use of tax data 
to projects that would improve tax administration, precluding 
the use of these data under controlled circumstances and 
conditions for other evidence-building purposes.
    To provide clarity about permissible statistical uses, the 
Commission recommends that Congress build on the legal 
framework for data protection already established under the 
Confidential Information Protection and Statistical Efficiency 
Act. We propose to extend that framework to cover the National 
Secure Data Service and enable it to acquire and combine survey 
and administrative data.
    The Commission also proposes that the Congress review and, 
where appropriate, revise relevant statutes that place limits 
on the use of administrative data for statistical purposes as 
well as making more of the data collected by the States in the 
course of operating federally supported programs available for 
evidence building.
    The Federal Government's principal statistical agencies 
already play an important role in generating and providing 
access to data. The Commission has a set of recommendations 
that broaden the role that the principal statistical agencies 
would play.
    Taken together with the Commission's recommendations 
related to privacy, which Dr. Sweeney will discuss, we believe 
that our recommendations on data access will allow data that 
the government has already collected to be safely harnessed to 
produce the evidence that is needed to make government work 
better.
    [Prepared statement of Ms. Abraham follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]

    Mr. Russell. Thank you, Dr. Abraham.
    The chair is now privileged to recognize Dr. Sweeney for 5 
minutes.

                  STATEMENT OF LATANYA SWEENEY

    Ms. Sweeney. Thank you.
    Congressman Russell, Ranking Member Cummings, and members 
of the committee, my name is Latanya Sweeney, and my career 
mission has been to create and use technology to assess and 
solve social, political, and governance problems. I am a 
computer scientist, a data scientist, a professor at Harvard, 
and the director of the Privacy Lab at Harvard. I served as the 
Chief Technology Officer for the U.S. Federal Trade Commission. 
And I really thank you for this opportunity to speak before you 
today about the importance of the Commission's recommendations 
in protecting the American people's confidential data.
    I think it's most directly relevant for you to know that my 
research and my research team have spent many years showing how 
data that should be anonymous isn't. That is, how many 
different ways confidential data sets on health, criminal 
justice, and income records, and other areas could be re-
identified easily using other data sources in today's 
technological society.
    So let me be very clear: We do have a major problem 
protecting the privacy of confidential data today, and it 
stems, primarily, because of a mismatch between our historical 
way of thinking about privacy and privacy protections and 
today's society being so empowered by technology.
    The Commission believes we can securely increase access to 
confidential data for evidence building based in great part on 
new advances made in how we think about privacy, in particular 
data privacy, and new techniques that use the same technology 
that challenges privacy to help provide protections. If we 
don't take these actions, we risk exposing confidential data 
about Americans widely, which leaves us vulnerable to many 
problems.
    The Federal Government collects a lot of information about 
individuals and businesses during the course of its daily 
operations, and much of that information is and should be open 
data. That is, it should be publicly accessible government 
information. And things like weather forecasts and train 
timetables do not carry the same privacy burden.
    The government says it will keep some of that information 
confidential, like names and dates of birth of Social Security 
recipients. And when the government pledges to keep data 
confidential, the data should have strong protections, and data 
used should generally be made known to the American public. 
Versions of the data that can be rendered sufficiently de-
identified should also be made publicly available.
    I'm here today to tell you about why the Commission's 
privacy and transparency recommendations are critical to 
protecting the government's confidential information.
    First, and of utmost importance, there's a great variation 
in how Federal agencies go about protecting confidential data 
today. Instead, this process really needs to be consistent, it 
needs to be rigorous, and it needs to be able to evolve with 
new technological advances.
    Second, protecting the privacy of the American people means 
being transparent and open about decisionmaking and processes, 
and clear about how confidential information is being used, and 
giving opportunities for feedback and improvement.
    So what happens now when a Federal agency wants to release 
a public use version of some confidential data that it's 
collected? Well, the 13 principal statistical agencies 
routinely apply rigorous methods of data masking, and they seek 
to review and get approval from experts on the way they're 
going to disclose the data before they actually release public 
use files.
    That's the current best practice we have, and we've had it 
for a long time and we accept it as pretty sufficient. But the 
context of public use data releases has changed because the 
amount of information about Americans and individuals that is 
publicly available has grown tremendously over the last few 
years.
    In addition, the technology to permit unauthorized re-
identification has improved. Financial incentives exist in many 
data analytic companies to take disparate pieces of data and 
put them together to build profiles on individuals. Within the 
Federal Government alone the open data initiative made over 
150,000 data sets accessible through a single website, 
including many administrative data sets never before released 
to the public.
    While releasing these data can generate tremendous value, 
enabling entrepreneurs to produce better products and 
departments to understand their work better, it's important to 
consider how publicly available data could compromise 
confidentiality if we don't take appropriate actions in 
adhering privacy.
    This is not about whether or not the data is released or 
not. It's not a binary decision. That's how we've historically 
looked at it. But today's technology allows us the opportunity 
to say we can, in fact, provide public versions of data. The 
question is which version and the techniques used to render 
that version.
    Government agencies follow their own applicable laws and 
regulations in providing access to their confidential data. 
These agencies are not necessarily coordinated in the 
decisionmaking they make. And the fact that they have different 
policies and different procedures about what it means to be 
identifiable creates a lot of problems. Sometimes two different 
agencies releasing the same data make different decisions, and 
the two pieces can be put together.
    Some program agencies use the same best practices that we 
just described earlier about the principal statistical agencies 
to assess the risk of re-identification, an ongoing process. 
The Department of Education even set up a dedicated disclosure 
review board for its program agencies.
    But some program agencies do little more than just remove 
the explicit identifiers, like name and address, leaving lots 
of other pieces of information out there available that can be 
linked to other data to re-identify individuals. Confidential 
government data collected by program agencies are often subject 
to FOIA with minimal redaction and also with inconsistent 
coordination.
    The problem is that there are so many sources of data out 
there today that can be matched to insufficiently de-identify 
confidential data to re-identify individuals or businesses. And 
in my resume you will see many collections at both the State 
and Federal level for where we've demonstrated this.
    In fact, we just released a study showing how data on air 
and dust samples from 50 homes in two communities in California 
could be combined with data released under the Safe Harbor 
provision of the Health Information Portability and 
Accountability Act to uniquely and correctly identify 8 of 32, 
or 25 percent, by name.
    Many people--many of us--many of our policies look at the 
protections available in HIPAA as being sufficient and Safe 
Harbor as being sufficiently strong. The fact that we were able 
to re-identify uniquely and correctly 25 percent of the records 
is alarming. It, again, speaks to the nature of data that's out 
there and the old way that we think about privacy.
    So I can tell you by name health information about eight 
people in one community from which that data was released, de-
identified. And if those eight people lived in your district 
and they learned that a Federal agency had just released their 
data with insufficient privacy protections, you would be likely 
hearing about it also.
    This is what my colleagues and I discover every day with 
many different types of data from many different venues. And 
this is a real problem. And it's important that any legislation 
implementing the Commission's recommendations address the 
seamless improvements of privacy head-on.
    Many programs have released the identified public use data 
files for decades without being required to ever formally 
assess its risk. They also often include provisions that make 
it impossible to actually inform them about the risk that we 
find.
    The Commission's recommendations, 3-1 in particular, will 
make sure that Federal agencies planning to release de-
identified confidential data use state-of-the-art methods to 
protect individuals and businesses from privacy harm. Instead 
of being a static way of thinking of privacy in a binary 
decisionmaking, it's a continuum. And as we get better and 
better about how to think about privacy and technical tools to 
release privacy, it enables those new techniques to go right 
into government use.
    Next, I'd like to explain why transparency is so important 
to privacy. Privacy does not mean secrecy, and there's often a 
lot of confusion that the idea of privacy is to hide it and not 
let anyone know about it. But, in fact, the Commission believes 
that advancing beyond the status quo and achieving unparalleled 
transparency means first telling the public about how 
government data are used for evidence building, and, second, 
regularly auditing whether the government is doing what it said 
it would do to protect privacy when allowing access to 
government data for evidence building. Further, transparency 
means how was the data redacted and learning new ways and 
encouraging the use of new techniques.
    As a first step, the government needs to make clear its 
decisions about which data are open data and which data are 
nonpublic confidential data. The Commission calls for OMB to 
develop a public inventory of data available for evidence 
building, including a determination of the sensitivity level of 
the data. Based on the data's sensitivity, we are 
recommending----
    Mr. Russell. Dr. Sweeney, we----
    Ms. Sweeney. Oh, I'm so sorry.
    Mr. Russell. Yeah.
    Ms. Sweeney. And I'll stop right there.
    [Prepared statement of Ms. Sweeney follows:]
 [GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Russell. I've tried to be generous from the chair, but 
we've given you 10 minutes instead of 5. And, hopefully, we can 
get to some of these other points during the questioning.
    Ms. Sweeney. No problem.
    Mr. Russell. We want to be respectful, also, to Mr. Shea, 
who it is my privilege to recognize for 5 minutes.

                    STATEMENT OF ROBERT SHEA

    Mr. Shea. I assure you the committee members benefit far 
more from hearing from Latanya than they will from me. So I 
apologize in advance.
    Mr. Chairman, Ranking Member Cummings, and members of the 
committee, I'm Robert Shea, a proud former staff member of this 
committee, and I'm pleased to be here today with my fine fellow 
commissioners testifying on behalf of the Commission on 
Evidence-Based Policymaking.
    You've already heard the testimonies of my colleagues that 
highlight the Commission's recommendations related to data 
access and privacy protections. I'd like to focus on those 
recommendations that pertain to the strengthening of the 
Federal Government's capacity for evidence building.
    Today, evidence building by government occurs unevenly. 
Some departments have robust approaches for routinely 
generating and using evidence, but these are the exception. 
We've taken to calling those who generate, manage, and analyze 
data, those who transform information into evidence, and those 
who support those functions through the routine processes of 
government, members of the evidence community.
    Principal statistical agencies and other statistical 
programs, program evaluation and policy research offices, 
program administrators, performance management offices, policy 
analysis offices, and privacy offices, all play important roles 
in evidence building. But many shared with the Commission that 
administrative barriers hamper the efficient production and the 
use of this evidence.
    To achieve the Commission's vision--a future in which 
rigorous evidence is created efficiently as a routine part of 
government operation and used to construct effective public 
policy--Federal agencies must have the capacity to support the 
full range of analytic approaches required for evidence 
building.
    To grow the program evaluation function across agencies, we 
recommend Federal departments anoint a chief evaluation officer 
who'd be charged with establishing department-wide evaluation 
research policies, coordinating technical expertise for 
evaluation, identifying priorities for departmental program 
evaluation, and adopting human capital strategies that expand 
the department's program evaluation capacity.
    We also recommend agencies develop multiyear learning 
agendas to support the generation and use of evidence. A 
learning agenda is essentially a strategic plan for evidence 
building, identifying important policy questions relevant to 
the department's mission. The learning agenda can be used by 
leadership and by Congress to prioritize research investments.
    Several of the evidence-building examples referenced in my 
colleagues' testimony involve linking data sources administered 
by different Federal departments. Officially implementing 
evidence-building activities across government requires strong 
coordination. This committee has already vested a great deal of 
responsible in OMB, but we firmly believe it's the right 
institution to help coordinate these activities. We recommend 
OMB facilitate cross-government coordination and consider 
whether consolidation or reorganization of evidence-based 
policymaking functions at OMB would accelerate adoption of the 
Commission's recommendations.
    The commissioners also identified actions related to 
procurement and the review and approval processes for new data 
collections that would require little cost but offer 
substantial benefits and savings while making it easier to 
produce evidence. These are small but important reforms.
    The Commission acknowledges and appreciates the role that 
this committee in particular plays in reforming and overseeing 
the operations of the Federal Government. Though some of our 
recommendations will require legislation, others simply require 
administration action. You can help us ensure OMB and Federal 
departments use their existing authorities to begin to increase 
access, enhance privacy, and expand capacity, and ultimately 
create a future in which rigorous evidence is created 
efficiently, as a routine part of government operations, and 
used to construct effective policy.
    No pressure, but we're all counting on you, and we're here 
to help in any way we can.
    [Prepared statement of Mr. Shea follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]

    Mr. Russell. Thank you, Mr. Shea.
    And thank all of you for your testimony.
    I will now recognize myself for 5 minutes as we move to the 
questioning period.
    Dr. Haskins, you made an interesting comment here in your 
testimony, and also in your written statements, that most of 
the Nation's social programs produce modest or no impacts on 
the problems that they were meant to address. I think most of 
America would not be shocked by that statement, many of the 
things that we see. But can you give us some examples of that 
and, in your discovery, why that is?
    Mr. Haskins. One example that people often mention is Head 
Start, which most people, if you do a survey of Americans, they 
think the program is immensely successful. But research shows 
that it's not necessarily successful.
    Let me be clear on this. There are some--and this often 
happens in programs implemented around the country. There are 
individual sites that are very successful. But if you average 
up all the sites and see if they perform better than, say, a 
control group that did not have the same experience, they 
usually fail. There are all kinds of evidence that 80 to 90 
percent of programs in medicine, in social science, and in 
business, fail.
    That's the main reason that we need to have more evidence. 
We need to develop these programs. I have a feeling we may have 
a chance to talk about some of these programs later in this 
hearing or it would be appropriate for you to look at these 
programs later. But you would find, if you did this, that most 
programs, like Head Start, often have a good reputation, but 
when you look at what they do in the country as a whole they 
are not successful.
    Mr. Russell. And that's one of the key things that this 
committee does, is to have oversight for the American people so 
that we can make sure that we're spending dollars wisely 
instead of just creating new bureaucracies that really don't 
address the problems that they were designed to do.
    Mr. Haskins. Mr. Chairman, can I add one very quick thing?
    Mr. Russell. Sure can.
    Mr. Haskins. Getting from here to there is going to be 
extremely difficult. We're in the middle of this now with teen 
pregnancy prevention, home visiting, and several other programs 
that were initiated 5, 6, 7 years ago. We need a strategy that 
we can gradually build up these programs. We are not going to 
go overnight from 10 or 15 percent success to 80 percent. 
That's just not going to happen.
    So we need better strategies for finding the right people 
at the local level, giving them the right resources, having 
Federal agencies that can help them implement their programs. 
It takes all of the above. It will take us years to develop 
this. We need to be patient and do it right.
    Mr. Russell. Thank you.
    Dr. Sweeney, recommendation 3-3 in the report is that each 
Federal department assign a new role of senior agency official 
for data policy. We already have chief information officers and 
senior agency officials for privacy, and yet now we want to 
create new positions when these old positions, when they were 
first recommended and implemented, it was to fix problems. And 
now we're going to fix a problem by having another person to 
fix a problem. What's the difference between these existing 
positions and why?
    Ms. Sweeney. You know, that succession of increasing those 
positions really speaks to the changes in society and our 
operations based on technology. Many of the Federal information 
technology officers are primarily focused on just that, the 
machines themselves, the ITs, the infrastructure of the systems 
on which work is based. The chief privacy officer in most of 
the agencies is based on making sure that the agency is in 
compliance with privacy laws and regulations.
    But what we're talking about is a different issue. It's 
about the data that's on the technology, the data that's being 
provided within the context of existing privacy laws. That is, 
the agency has the right to give out the data or not, or is 
responding to the Privacy Act. That would be in the chief 
privacy officer. But what version of the data is actually being 
given? That's a technical analysis that neither of the other 
two would be able to actually implement.
    Mr. Russell. I see.
    And, Mr. Shea, in your comments and also your written 
report, you talked about agencies using learning agendas. And 
there's been some examples of success on that. I mean, imagine 
having a strategic vision and laying that out. Could you give 
us an example of what you're talking about there and who's 
using it?
    Mr. Shea. Sure. I think there are some agencies, perhaps 
the Department of Labor, Department of Education are ones, that 
identify early in the year, or perhaps over a longer term, what 
are the major questions they want answered. And they contract 
for rigorous independent evaluations of programs that answer a 
lot of those important questions. And then they do a better job 
than most at integrating the evidence that they learned into 
their decisionmaking process.
    This committee has a long history of trying to inject 
outcome-based management in Federal departments and agencies. 
This is a step in that maturity, in our view.
    Mr. Russell. Okay. Thank you.
    And my time has expired. And it's now my honor to recognize 
the ranking member, the gentleman from Maryland, Mr. Cummings.
    Mr. Cummings. Thank you very much.
    Dr. Abraham, one of the primary themes of the Commission's 
report is the need to provide researchers both inside and 
outside the government with better access to government data. 
You noted in your testimony that the kind of data the 
Commission focuses on is data that the government keeps 
confidential because it contains sensitive information. Can you 
give us an example of the kind of data you're talking about?
    Ms. Abraham. So we're talking about data on participation 
in programs and data on the outcomes associated with 
participation in those programs, data on the earnings of 
people, such as unemployment insurance wage record data, 
potentially Federal tax data on earnings that, if made 
available to researchers under strict confidentiality 
protections, could let them do a much better job of 
understanding how well these programs that Ron was talking 
about were actually working.
    Mr. Cummings. Now, the reason this data is not publicly 
released is because it contains sensitive information about 
individuals, such as Social Security numbers.
    Ms. Abraham. Exactly.
    Mr. Cummings. And the Commission recommends that agencies 
provide better access to data, but, obviously, without 
compromising the security of the sensitive information in that 
data. Is that right?
    Ms. Abraham. That's right. We're envisioning this National 
Secure Data Service where data on, say, participation in a 
program would be brought together with data on earnings 
outcomes. The staff there would link the data up. The 
identifiers would be removed. The researchers would be given 
access to the data within that secure enclave to carry out the 
analysis. The only kind of results that would be released from 
such projects would be aggregated information that didn't allow 
anyone to be identified.
    Mr. Cummings. Well, Dr. Abraham, you talked about how 
confusing and inefficient it can be when laws passed by 
Congress impose inconsistent restrictions on how government 
data can be used. For example, the Commission highlighted the 
Higher Education Opportunity Act. That law limits how the data 
the Department of Education collects from colleges and 
universities can be used. That limitation is creating less 
accountability for the performance of those institutions.
    Should Congress amend the Higher Education Opportunity Act 
to increase access to data from colleges and universities?
    Ms. Abraham. The recommendation of the Commission was, 
essentially, that the Congress take a hard look at that. As I 
think is clear from reading our report, that kind of limitation 
on how data can be used does reduce accountability, it does 
reduce our ability to understand what we're getting for the 
Federal dollars that we're spending. And we certainly think 
that another look should be taken at that.
    Mr. Cummings. Well, are there other laws Congress should 
reexamine for reform?
    Ms. Abraham. Well, I mean, there are similar restrictions 
in law on the Federal Government compiling information about 
people participating in workforce training programs supported 
by Federal dollars. That would be another example of something 
we would think the Congress would want to take a look at.
    Mr. Cummings. Now, Dr. Sweeney, you said in your testimony 
that the government should make more information available to 
the public about its data, how the data is being used. Is that 
something that the administration could do now?
    Ms. Sweeney. In some situations. Not all data--they could. 
But there's no incentive in most--there's no one answer because 
at the last I counted we have 2,167 privacy laws and 
regulations in the United States, and they're inconsistent with 
each other. And so there's no one answer, could they just do 
that unilaterally. But, in certain places, that certainly could 
be done. But the incentivizes may not be there, necessarily, to 
give it. And whether an agency chooses to make that decision is 
not transparent.
    Mr. Cummings. Well, I agree that it's important that data 
collected by the government be as transparent and as accessible 
as we possibly can make it. Taxpayer money is spent collecting 
the data, and taxpayers deserve the highest rate of return 
possible.
    Would you agree, Dr. Sweeney?
    Ms. Sweeney. Yes, I would, especially because versions of 
the data can be made free of privacy concerns.
    Mr. Cummings. With that, I yield back, Mr. Chairman.
    Mr. Russell. The chair now recognizes the gentleman from 
Tennessee, Mr. Duncan, for 5 minutes.
    Mr. Duncan. Thank you, Mr. Chairman.
    I noticed that three of your chapters deal with restricting 
access to confidential information, and especially with 
privacy. Most of your recommendations deal with privacy. And 
then there was an article in The Hill newspaper which said that 
your goal is to make sure that Federal tax dollars are spent 
effectively and how we keep up with rapid, advancing technology 
and still make effective public investments.
    I read after the--also, today, just this morning, the 
National Journal Daily that's on all of our doors each morning 
said the Equifax hack is going to lead to sweeping 
cybersecurity legislation.
    About 2 or 3 months ago there were some worldwide cyber 
attacks. And Robert Kuttner, who is the co-founder of The 
American Prospect magazine and a very liberal columnist who I 
wouldn't ordinarily quote, he wrote this. He said: Last week's 
cyber attack could produce the wrong lessons. The immediate 
take-away seems to be that large institutions need much better 
cybersecurity systems.
    But he goes on. He says: Hackers will always be able to 
find ways of getting into network systems. The fantasy of ever-
better cybersecurity is delusional. We could spend half the GDP 
on network security and someone will still find a way to breach 
it.
    And so I guess I have two sort of related questions. Number 
one, do you agree with Kuttner that--I assume you don't agree 
with Kuttner that cybersecurity is delusional or is a mega-
billion-dollar hoax. And I'd like to hear your comments about 
that.
    But, secondly, if the goal of your Commission is more 
effective spending of Federal money--you know, I drive cars 
that are several years old. They're still working real well. 
But yet they say that computers are obsolete the day they're 
taken out of the box.
    So how do we have the state-of-the-art technology that's 
already been mentioned in your testimony here today, yet we get 
effective use--we can't just throw away a computer just because 
the next year they come out with one that's got more bells and 
whistles on it.
    And I sometimes wonder--I know at this committee several 
years ago, we had a business in here that had downloaded 
250,000 Federal tax returns just to show that it could be done. 
So I sometimes wonder if the technology and the internet that 
has done so much for us has also almost completely done away 
with privacy.
    So, Dr. Sweeney?
    Ms. Sweeney. Thank you.
    So first of all, what the Commission is addressing does not 
actually have anything to do with cybersecurity. So the 
relationship between security and privacy is really interesting 
because usually when we think about computer security, it's the 
breaking in of a machine, you know, getting in through illegal 
means.
    The data that we're talking about are data that's given 
away freely, whether it's through open data, through a public 
use file, or chosen not to be given away out of claimed fear or 
something like that.
    So these kind of data decisionmaking is not the same as 
cybersecurity. No one is breaking in. This is data that's 
usually freely given away or decisions made not to do so at 
all.
    Mr. Duncan. Well, if it's given away so freely, why is 
there so much of your report dealing with privacy?
    Ms. Sweeney. Because the errors are happening on both 
sides. We have situations where the data is given away, and it 
leaks private information, sensitive information, health 
information, income information, and so forth, on Americans. 
And at the same time there's data that could be incredibly 
useful for evidence-based policymaking that isn't being given 
out at all because they say they don't know how.
    And so we have problems on both sides, and the reason we're 
having those problems is because the privacy decision making 
system doesn't really use any of the technology.
    So you talked about the wave of the technology curves and 
the speed at which technology changes. What the Commission does 
is it basically says we want privacy decisionmaking to ride the 
wave of the technology and not be left in its 1970s format.
    Ms. Abraham. If I could just jump in with one quick 
comment.
    Mr. Duncan. Sure.
    Ms. Abraham. In terms of the cybersecurity issue, a point 
that I would make is that what we're talking about is better 
use of data that are, in most cases, already collected and 
being held. So in terms of that risk, we're really--we should 
do as well as we can on the cybersecurity issues, but we're not 
increasing that risk in any of our recommendations.
    Mr. Duncan. What about the second part? How do we have 
effective use of tax dollars but we still keep up with the 
rapid advancing technology?
    Ms. Abraham. I think that what Latanya's response was, was 
attempting to answer that, which is we're not so much talking 
about the hardware and having to buy new hardware. We're 
talking about applying the right sort of methods to the way 
that data files that are going to be released get structured.
    Mr. Duncan. All right. My time is up.
    Mr. Russell. The chair now recognizes the gentleman from 
Massachusetts, Mr. Lynch, for 5 minutes.
    Mr. Lynch. Thank you, Mr. Chairman, and I thank the ranking 
member for his good work.
    And thank you to the witnesses for helping the committee 
with this task.
    Dr. Haskins, we do a lot of work on this committee 
regarding veterans. And we spend right now about $14 billion a 
year on educating veterans after their service. Yet, we don't 
collect any data on how efficacious or how much good we are 
doing on behalf of those veterans. The intent is there, but 
there's no followup. And I'll give you an example.
    So we have a GI Bill that we've offered to members of the 
military. And in the past not only have we given it to 
veterans, but we've also said to Afghan and Iraqi veterans that 
if you already have a degree in higher education, you can give 
this to your child, which is a wonderful, wonderful benefit.
    Some of the members of the committee were in Afghanistan 
not too long ago and we met--we were with a Stryker Brigade 
from Washington State. And one of their sons was actually--one 
of the officers there, one of their sons was entering 
University of Washington.
    But when we look at the numbers, we find that only 50 
percent--50 percent--of our veterans are actually using that 
bill. That's unbelievable. With the cost of education, with the 
benefit that it could provide to them and to their children, 
there's only 50 percent uptake of that benefit. And then we 
don't know whether, for those who do choose to go to college, 
we don't know if those veterans are actually benefiting to the 
full extent that they may.
    So there's sort of a vacuum of usable data. I think that 
part of the problem is because of the restrictions we put on 
the use of information regarding--under the Higher Education 
Act--we have put a clamp on some of that. And I think we've 
restricted our own ability to collect and to use that 
information.
    It's unbelievable to me that that benefit would be out 
there, so desperately needed, and left unused at least half of 
the time.
    Do you have any thoughts on that, Doctor?
    Mr. Haskins. Yeah. Several. One thought that I have--first, 
of all, I want you to know that I went to school on the GI Bill 
during the Vietnam era. So I am very appreciative of what 
Congress does for----
    Mr. Lynch. Thank you for your service to your country, 
Doctor.
    Mr. Haskins. Yeah. I was glad to do it and even more 
pleased to go to college and have somebody help me.
    Second point, we can find out almost anything you want to 
know about these programs. We have great research designs for 
almost all education and training programs. Dr. Abraham was the 
head of the BLS and knows as much about this as anybody in the 
country.
    Mr. Lynch. I'm sure.
    Mr. Haskins. So if you approve the money and direct the 
Department of Labor to do careful studies of these veterans 
programs, they can answer any question that you want.
    Now, I'm going to tell you something that's very important. 
It is not the case that you can just take anybody that comes 
out of military service and send them to a good university and 
they'll do well.
    I have looked into this issue several times over the years, 
and I always come away with the same conclusion. There is a 
recent book by Harry Holzer, who is a very well-known labor 
economist, comes to the same conclusion. A lot of kids are not 
ready to go to college. And the GI Bill, they don't wind up 
with a debt. But on Pell Grants and other means and loans, they 
can wind up owing money. And they don't get the degree, they 
owe the money, they're really in a bind.
    So those are the kind of considerations we need to take 
into account. People in the military need to be better prepared 
for college, and many other people do as well. So the idea that 
we're going to send all of our kids to a 4-year college and 
they're going to make $80,000 a year, that's not going to work. 
They need other things. And, fortunately, the Department of 
Labor specializes in those other things as well.
    Mr. Lynch. Dr. Haskins, let me just reclaim my time. I 
concede your point. However, I also have examples in my own 
district where veterans did not know that they could have sent 
their child to college on their GI Bill. So that was a missed 
opportunity. And not only that, but a high number of our 
veterans are coming out of the service with great skills in 
STEM, you know, on science and math, and those are also missing 
the opportunity.
    But I do appreciate the candor and the insightful answer 
that you've provided. Thank you.
    And, Mr. Chairman, I yield back.
    Mr. Russell. The gentleman yields back.
    The chair now recognizes the gentleman from Texas, Mr. 
Farenthold, for 5 minutes.
    Mr. Farenthold. Thank you very much.
    I want to visit with Dr. Sweeney a little bit. You've kind 
of hit on a topic that I'm on three sides of. I'm the sponsor 
of the OPEN Gov Data Act, which makes a lot of government data 
available that was just included in the Senate version of the 
NDAA. So it actually has a pathway to becoming law.
    I also consider myself to be a privacy advocate. And as 
someone who runs a campaign, having a lot of information about 
individual people that's identifiable makes it a whole lot 
easier for me to get them the information that they need to 
make an informed choice to vote for me.
    But I want to talk a little, you mentioned that--you cite 
some examples of how anonymized data is de-anonymized. And I 
wasn't able to find it in any of the material that I have. I 
wanted to ask that you could get that to me. I assume it's more 
complex than just finding two data sets that have a common 
field. But that's probably the easiest way to do it.
    Ms. Sweeney. Are you asking--clarification: Are you asking 
about re-identifying or de-identifying?
    Mr.  Farenthold. Yeah, de-anonymizing or re-identifying 
information. You said some--you had some examples. I'd like to 
see those, and I couldn't find them in the material that I had.
    Ms. Sweeney. Very easy. I have a long list of them. I'll 
give you one very simple one. Washington State releases 
hospitalizations on--all hospitalizations made in the State. 
It's over a million visits. And we were able to just simply 
match the de-identified versions of those data against simple 
blotter stories, the kinds of things, you know, that appeared 
in the newspapers, just simple matching, no statistics or 
anything, for 41 percent of the records----
    Mr. Farenthold. So is that, is the solution to that just 
more education or policy with respect to how that information 
is released, or--I mean, is there a way to solve this, or is 
this just going to an inherent unsolvable problem?
    Ms. Sweeney. So I'll stick to the Washington State example, 
because I think it's a great one.
    So Washington State responded by getting rid of its 1970s 
way of thinking about privacy. That is the old way of saying, 
``Oh, I just removed these fields, and the rest of the data is 
fine.'' And they went instead to a risk-assessment model, 
exactly the thing that we're talking about in the Commission 
report. And they came out the other where you can still get, 
for $50, the 1 million records, but now it has a scientific 
assessment that's done that you can't do the same type of re-
identification.
    If you still need the more sensitive version, they have a 
more aggressive application process that you would have to----
    Mr. Farenthold. So this is solvable?
    Ms. Sweeney. Yeah, it's solvable. And risk assessment is a 
great way to do it, because risk assessment keeps us riding 
what's the latest thing that the technology allows.
    Mr. Farenthold. And I think this is important, because I 
think way too many decisions here in Washington are made on 
anecdotal information and not scientific information.
    Now, Dr. Haskins, you talk about--the report talks about 
creating a whole new government agency. As a conservative who 
wants to shrink the size of government, that really rubs me the 
wrong way.
    Why couldn't this be something that's rolled into--well, 
let's pick the Census Bureau, because they're the first agency 
that comes to mind for dealing with large amounts of data and 
has that experience. Can there be something done without 
setting up a whole new bureaucracy? And could you talk a little 
bit about why you recommend creating a whole new government 
agency?
    Mr. Haskins. Yes. As Dr. Abraham already said, we created 
this--or proposed this new agency called the National Secure 
Data Service. It would first exist within the Department of 
Commerce. And one of our explicit intents was not to create 
some big new agency that would have mountains of data. Rather 
it was to build on things the Census Bureau is already doing 
and expand those gradually over a period of years so that 
primarily we could have a temporary repository for data that is 
needed for good studies that have been approved through an 
elaborate process. And then the data would be sent back to 
wherever it came from, whatever agency it came from.
    And over a period of years, I could imagine that we would 
wind up spending more money on this new National Secure Data 
Service. But to begin with, I think we've been as efficient as 
you could be to create the ability to do this kind of making 
the data available and linking the data. And I think I'm 
confessing it would cost more over a period of years, but I 
think we've done it in an efficient way.
    Mr. Farenthold. And finally, one last question.
    Mr. Haskins. Can I just add one thing? If we ask the Census 
Bureau to do this, they would have to stop doing something 
else. So we're hoping that, at least on a temporary basis, 
certain employees can be borrowed and that they can make a 
minimum of hiring in order to build this agency. And if you 
went out and hired all new people and created all these new 
positions, then it would cost a lot more than under the system 
that we're recommending.
    Mr. Farenthold. Well, you ran the clock out on me.
    Thank you very much. I yield back.
    Mr. Russell. The gentleman yields back.
    The chair now recognizes the gentleman from Virginia, Mr. 
Connolly, for 5 minutes.
    Mr. Connolly. Thank you, Mr. Chairman.
    Dr. Haskins, you've stressed evidence-based policy making, 
and I certainly agree. I think probably everyone does in the 
abstract. But it's almost laughable that public policy is, in 
fact, always consistently based on evidence. In fact, quite the 
opposite.
    Marijuana. Evidence-based policy?
    Mr. Haskins. I don't know.
    Mr. Connolly. You don't know.
    Mr. Haskins. I haven't researched marijuana.
    Mr. Connolly. Okay. Wouldn't you say metrics tell us 
evidence?
    Mr. Haskins. Wouldn't I say what?
    Mr. Connolly. Metrics has something to do with evidence?
    Mr. Haskins. Yes.
    Mr. Connolly. All right. Any metrics on how many marijuana 
overdoses there are every year?
    Mr. Haskins. I believe there are, yes.
    Mr. Connolly. No, I believe--we've had testimony before 
this committee there aren't. And this is classified as the most 
dangerous drug in America.
    How many marijuana users die on the roads every year? Do we 
measure that?
    Mr. Haskins. I don't know.
    Mr. Connolly. No. The answer is no, sir.
    So we have incarcerated millions of people, arrested 
people, and decided marijuana is the most dangerous drug in the 
United States since the era of Richard Nixon. And we have very 
little data, scientific data to justify that. And the damage 
done--enormous.
    Terri Schiavo. Remember that case?
    Mr. Haskins. Yes.
    Mr. Connolly. So a Republican Congress and a Republican 
President, who actually interrupted his vacation, a rare event, 
to come back to Washington to actually sign into law an 
unprecedented intrusion by the United States Congress, imposing 
its judgment, scientific judgment, on the state of the health 
of that young lady so that her husband couldn't make a private 
medical decision. Was that evidence-based, do you know, Dr. 
Haskins?
    Mr. Haskins. I know something about these type of cases. 
There is evidence, but the doctors often disagree about the 
evidence. It's pretty murky.
    Mr. Connolly. Pretty murky?
    Mr. Haskins. Yeah.
    Now, some cases are clearer than others. And my 
understanding, the Schiavo case was quite clear.
    Mr. Connolly. They were quite clear.
    Mr. Haskins. Yes.
    Mr. Connolly. Right. Because when she was autopsied after 
she died, her brain was one-third the size of a normal human 
brain. It had atrophied. That didn't stop Congress from 
overriding scientific data.
    Climate change. Evidence-based? The decision, for example, 
to rip up the Paris Climate Accord, was that evidence-based?
    Mr. Haskins. I would say no.
    Mr. Connolly. No.
    In fact, would you not agree that the overwhelming evidence 
is climate change right now is certainly being--a key variable 
is human activity, and the evidence is pretty overwhelming 
about that in the scientific community. Would you agree with 
that?
    Mr. Haskins. I don't consider myself an expert in this, but 
I've read a fair number of things, and I think that is correct.
    Mr. Connolly. And certainly, given your position, you would 
want Congress to base all--as much as possible--its decisions 
with respect to climate change, not on belief, but on evidence.
    Mr. Haskins. Okay. Let me say this.
    Mr. Connolly. Yeah.
    Mr. Haskins. I do not expect that Congress would make 
decisions exclusively based on evidence. I want evidence to 
have a place at the table. I want Congress to understand what 
the evidence is. But they will use other factors to decide how 
much money they should spend or whether they do anything at 
all.
    Mr. Connolly. Oh, believe me, I know.
    Okay. One of the findings in your report, you say that 
you're worried about data center consolidation and that it 
might hamper the ability of the evidence-building community to 
limit access to confidential data.
    How many data centers are there in the Federal Government? 
Do you know, Dr. Haskins?
    Mr. Haskins. We have 13 agencies now, but it's growing.
    Mr. Connolly. Data centers?
    Mr. Haskins. Yes, special data centers are 13, and they're 
growing.
    Mr. Connolly. Mr. Shea.
    Mr. Shea. I think there's confusion. We're talking about 
statistical agencies who have the mission of collecting data 
for producing statistical information versus the myriad data 
centers that are created for a wide variety of----
    Mr. Connolly. So you're not referring to the latter.
    Mr. Shea. That is correct.
    Mr. Connolly. Okay. Because let me just say, this committee 
has spent years looking at that subject, Mr. Shea.
    Mr. Shea. We think you should divert the savings from data 
center consolidation to evidence-based policymaking.
    Mr. Connolly. We're with you. Because we've got four 
agencies that have saved $2 billion. And Mr. Hurd's not here, 
but he and I and Ms. Kelly and Mr. Meadows have cosponsored 
legislation to allow agencies to reinvest in themselves 
pursuant to the savings from data center consolidation.
    Okay. That's really important, because we were going to get 
in a bit of a tizzy about that.
    Let me just say in closing, if I may, boy, do I agree with 
the premise of your Commission. But what is so troubling, 
frankly, about the era in which we operate is how easily 
dismissed facts and evidence--measured facts--are because of a 
priori beliefs or because of denial. I don't want to accept 
that.
    And it's true in creationism versus evolution. It's true in 
climate change. It's true when conclusions come up from a 
scientific agency that somebody doesn't like or doesn't want to 
accept.
    And Congress is as guilty of that practice as any entity. 
And until we try to move more toward the mean, we take Dr. 
Haskin's caveat seriously, we won't have the best public 
policy.
    Mr. Russell. The gentleman's time has expired.
    Mr. Connolly. I thank the chair.
    Mr. Russell. The chair now recognizes the gentleman from 
Montana, Mr. Gianforte, for 5 minutes.
    Mr. Gianforte. Thank you, Chairman Russell and Ranking 
Member Cummings.
    And I want to thank the Commission for your work on this. 
It's very important that we have efficient and effective 
Federal programs.
    I want to direct my first question to Mr. Shea, if I could. 
You mentioned in your comments about outcome-based objectives. 
And I was just curious, being a business guy, I know that 
evidence is critically important, especially as you apply it 
against the goals of a program.
    What evidence did you find that the 209 government offices 
you surveyed had clear outcome-based goals that they were 
collecting evidence against?
    Mr. Shea. So that's not one of the questions we asked in 
our survey. But having spent my career trying to help Federal 
organizations develop those measures, I can tell you they're 
not on a steady glide path to all have clear outcomes with 
aggressive measures of their performance year over year.
    Mr. Gianforte. Okay. So I'd like to follow up on that. So 
you're saying that these, in your experience, government 
agencies and programs don't have clear goals for what they're 
trying to accomplish?
    Mr. Shea. In many cases they don't. I find them much more 
willing to measure their inputs or perhaps their output. 
Outcomes, things over which they don't have complete control, 
things that the programs Dr. Haskins was talking about don't 
work, they're very reluctant to hold themselves accountable for 
those kinds of goals
    Mr. Gianforte. Kind of hard to hit a target if you don't 
have one.
    Mr. Shea. That's exactly right.
    Mr. Gianforte. Yeah.
    So do you have any--in your experience, what suggestions 
you have along these lines?
    Mr. Shea. Well, you can play a major role in assessing 
agencies' strategic and annual performance planning process. 
Ask them to what degree they have clear outcome-oriented goals. 
And if you don't like what they produce, tell them to refine 
and improve them.
    When I was a staff member on this committee, the Government 
Accountability Office was an enormous help in setting a 
framework with which to judge those outcome measures.
    Mr. Gianforte. Yeah. I can't imagine that there's much 
satisfaction even for our dedicated government workers if 
they're working hard but not knowing what they're trying to 
accomplish.
    Mr. Shea. It's a very important point, because to clarify 
the mission that employees are trying to work to accomplish is 
a major factor in improving their engagement and improving 
recruitment and retention, which, as you know, is a big 
challenge.
    Mr. Gianforte. Okay. Thank you.
    Chairwoman Abraham, just in the work that you've done 
collecting evidence, I know in the private sector big data 
analytics, artificial intelligence have been used for a long 
time, particularly in the financial industry, to uncover fraud 
and inefficiencies. My question is, in your working with these 
209 agencies, what evidence did you find that these government 
agencies were using these common private sector approaches to 
uncover inefficiency?
    Ms. Abraham. That is, again, not something that our survey 
really spoke to directly. I guess I should say, when you think 
about big data, a lot of--there are a lot of data that the 
Federal Government collects that we could be doing more with. 
In terms of these agencies that we surveyed, you know, many of 
them don't necessarily have the capacity to do that.
    Part of what we were attempting to accomplish with the 
recommendations that we were making was making the data that 
they collect more accessible to people who do have those skills 
and who could do some of the kinds of things that you're 
talking about.
    Ms. Sweeney. May I add to that?
    You know, big data and AI are basically statistical 
algorithms. And when you think about where in the government we 
see that, it's primarily in the statistical offices, the very 
data that we're talking about. And so we're also saying that 
kind of innovation use is exactly the kind of thing we want 
used in privacy in rendering the data confidential. And this 
would be among the group who's more likely to be able to use 
it.
    Mr. Gianforte. And I know we all benefit every day when 
credit card transactions are identified that might be 
fraudulent. You know, looking at the volumes of data we have in 
the government and being able to say, you know, this one is not 
like the others, we maybe investigate a little further. I think 
there's opportunities there to uncover inefficiency and fraud.
    Ms. Abraham. I should clarify one thing in response to your 
question. What we are proposing is a facility that would be 
used for statistical analysis, to identify patterns and the 
outcomes associated with programs. We are not envisioning that 
this facility would be used to go in and identify individuals 
who had committed fraud and then come back and go after them. 
We're talking about improving the use of data for statistical 
purposes, not for targeting individuals and taking action 
against them.
    Mr. Gianforte. But it's certainly in our interest to 
uncover fraud where it exists.
    So, Mr. Chairman, I yield back.
    Mr. Russell. The gentleman yields back.
    And the chair now recognizes the gentlelady from New York, 
Mrs. Maloney, for 5 minutes.
    Mrs. Maloney. I thank the chairman and the ranking member 
very much. And I thank my good friend Val for yielding.
    I do want to share that we have a meeting coming up at 
11:30 with the Foreign Minister of South Korea. Everyone has 
been invited. I need to get to that too.
    But I wanted to ask about, really, the largest collector of 
information in our Government is the Census Bureau. And I'd 
like to first ask Dr. Abraham, the Commission's report noted 
that principal statistical agencies, such as the Census Bureau, 
have, and I quote, ``demonstrated responsible stewardship of 
data collection through census and surveys,'' end quote. It 
also noted, and I quote, ``Not surprisingly, public trust in 
the accuracy and validity of statistical data reflect the 
public's trust in the statistical agencies that produce them,'' 
end quote.
    Particularly for the Census Bureau, it seems that 
maintaining the trust of the public is important to encourage 
participation in censuses and surveys. Do you agree?
    Ms. Abraham. Absolutely.
    Mrs. Maloney. Well, then, Dr. Haskins, one of the most 
critical surveys of the Bureau is the annual community survey, 
which is randomly sent to addresses across the country to 
collect data on an ongoing basis.
    How important is the American Community Survey for 
evidence-based policymaking, Dr. Haskins?
    Mr. Haskins. Extremely important.
    Mrs. Maloney. Some of my colleagues have suggested making 
the Community Survey a voluntary rather than a mandatory 
survey. And do you expect that such a change would impact the 
amount of data the government is able to collect and the 
integrity of that data?
    Mr. Haskins. Yes. But the more important point is, if you 
can't have a random sample of the public, we won't know the 
frequency of all sorts of things that we estimate when we get a 
random sample. So if you do it on a voluntary basis, it 
basically ruins the American Community Survey, which is the 
best and most accurate survey of the population we've ever had.
    Mrs. Maloney. Well, thank you. And I know not only 
government, but the business sector relies on it. And I want to 
note that Canada made its long form census voluntary in 2010. 
And like you said, Dr. Haskins, the response rate dropped from 
93 percent in 2006 to just 68 percent in 2011, and government 
agencies were forced to make policy decisions based on old 
data.
    So reducing the volume and quality of data collected by the 
Census Bureau seems to be exactly the opposite of the 
Commission's goal of expanding access to and use of federally 
collected data. Does everyone on the panel agree?
    Mr. Shea. Yes.
    Mrs. Maloney. Okay. Thank you.
    Mr. Haskins. Can I add one thing very important?
    Mrs. Maloney. Sure.
    Mr. Haskins. The decline in the number of responses is not 
the biggest problem. The biggest problem is that the sample is 
no longer random. So you can't draw conclusions from it. It's 
ruined. That's a big point. I mean, the more people you have, 
the more reliable the numbers are. But if they're voluntary 
participants, none of the numbers are any good, they're not 
reliable.
    Mrs. Maloney. Thank you. And I want to note on that that 
the current President's budget proposal does not include enough 
for funding for the Census Bureau to perform the work it does 
for the upcoming 2020 Census. So the point you made earlier, if 
you add something, they have to stop doing something. So the 
appropriations bills currently moving through Congress would 
implement the President's approach. So we don't have enough 
funding right now.
    So do you believe that underfunding the Census could impair 
the quality of the data collected in the 2020 Census?
    Mr. Haskins. You're asking me?
    Mrs. Maloney. Yeah.
    Mr. Haskins. Yes, I do.
    Mrs. Maloney. And the President and Congress must increase 
funding. We need to do that. Starving the Census will result in 
inaccurate data and undercuts the vulnerable populations who 
rely on programs that are funded on this data. And if you don't 
have good data, you don't have good policy.
    I want to thank my colleague Val so much. I've got 4 
seconds left. I'm going to yield it back to you. I know you've 
got another 5 minutes coming up. But if you've got a point to 
make in 4 seconds, I'm sure it's an important one.
    Thank you so much. That allows me to get to my other 
meeting. Thank you so much.
    Mr. Russell. The gentlelady's time has expired. But we will 
get to Mrs. Demings.
    The chair now recognizes the gentlelady from North 
Carolina, Ms. Foxx, for 5 minutes.
    Ms. Foxx. Thank you very much, Mr. Chairman.
    And I want to thank this distinguished panel very, very 
much for the work that you did.
    I feel almost as though I have always been cursed with 
having real objectivity flowing through my veins and always 
being concerned about evidence-based decisionmaking. It may be 
a product of my having grown up extremely poor. You've got to 
make really, really good decisions in your life. But I have 
done this, again, all my life.
    I also think that Congressman Gianforte opened up a very 
important item when he talked about whether there are clear and 
measurable goals and outcomes out there in most Federal 
programs. My experience has been that there are not.
    And I think it was Dr. Haskins who mentioned--someone did--
that we have really, really great research designs out there. 
And maybe I confused that with measurable outcomes. But at some 
point, I'd like to talk, whichever member of the Commission 
talked about that, because I would really like to see where 
those great research designs are in the Federal Government, 
because I'm not aware of them.
    And then the other issue I'd like to follow up on later, 
but not now, on the what you all call workforce training 
programs. I try never to use that ``T'' word when it comes to 
dealing with human beings, because I think we educate people. 
And I don't use that. But because that comes in the other 
committee on which I serve, and, in fact, I chair the Education 
and the Workforce Committee, I would like to follow through on 
that.
    But I want to ask a couple of questions related to the work 
that you've been doing. Again, I think it's extraordinarily 
important work, and I can't thank you enough.
    Did the Commission consider how data quality affects the 
ability to use the data?
    Dr. Haskins, I think so you alluded to this.
    Mr. Haskins. Yes, I think we did.
    Ms. Foxx. Great.
    In the DATA Act pilot report, which was passed out of the 
Congress about 3 years ago, OMB recommended continuing the 
effort to standardize data. Would standardized data collections 
from Federal award recipients improve the evidence-building 
community's ability to use the data?
    Mr. Haskins. I think the answer is yes, but with a caveat 
that different programs have different objectives. So you'd 
have to have measures of those particular objectives, and they 
would differ substantially across projects.
    Ms. Foxx. Again, I'm really well aware of that. And I think 
one of the concerns I have--I was a reader for programs in the 
Department of Education. And to the extent possible, readers 
got information, got evidence about the success or failure of 
programs. And then the readers would--readers would evaluate, 
recommend to the staff. The staff would then ignore the 
recommendations of the readers based on the evidence that we 
were presented, which we thought was pretty good.
    You all have any ideas on how we can make sure that 
decisions within the agencies are made on evidence other than 
utilizing our oversight responsibilities here?
    Mr. Haskins. I have two quick recommendations. One, grill 
the senior officials when they're nominated about their 
attitudes about evidence and whether they plan to use it. And, 
second, when you get reports from Federal agencies that are 
under your jurisdiction, which is huge, as I understand it, 
call them before the committee and grill them on these issues, 
because what you described does happen, I know it happens, and 
the officials that are responsible for it should be called on 
the carpet.
    Ms. Foxx. Right.
    And the last thing I would say to you is, in considering 
the reauthorization of the Higher Education Act we've had 
probably 20 hearings. What we heard from people who complained 
about data they submit to the Federal Government is that we 
have lots and lots of data and little information.
    So I would ask you all to keep that in mind as we push the 
collection of more data, that we try to make the connection 
between--and I think you are--the connection between data for 
data's sake and data which provides us information.
    Thank you, Mr. Chairman. I yield back.
    Mr. Russell. The gentlelady yields back.
    And the chair now recognizes the gentlelady from Florida, 
Ms. Demings, for 5 minutes.
    Mrs. Demings. Thank you so much, Mr. Chairman. Thank you, 
Ranking Member. And thank you to our witnesses for joining us 
today.
    Dr. Sweeney, in your written statement, you stated, and I 
quote, ``Preventing bad actors from breaking into confidential 
data requires consistent and rigorous processes,'' unquote. In 
other words, we basically need to do more to protect 
confidential data.
    The inspector general of the State Department recently 
issued a report finding that 77 percent of the Department's IT 
assets are not in compliance with Federal cybersecurity laws.
    Dr. Sweeney, does that surprise you? Why or why not?
    Ms. Sweeney. No, it doesn't surprise me. But in terms of 
the Commission, even though the language I use sounds like 
security, about breaking into confidential data, I mean 
exploiting the data as it's given to you, which is not the same 
as breaking into a computer and breaching a database. But 
computer security problems are also rampant.
    Mrs. Demings. The State Department until recently was 
making strides in another area of its cybersecurity efforts. 
The State Department's Office of Coordinator for Cyber Issues 
serves as liaison between the State Department and the White 
House, other agencies, and outside stakeholders. The office 
also engages with international partners.
    Dr. Sweeney, do you think it makes sense to have a senior 
official within an agency like the State Department to 
coordinate cybersecurity efforts?
    Ms. Sweeney. So the Commission report isn't on 
cybersecurity. So I'm speaking for myself and not for the 
Commission. But, you know, cybersecurity problems are huge in 
the United States, and they do dovetail constantly with these 
data problems, because they just leave all of our systems 
vulnerable. And so we don't have a rigorous--we haven't come to 
a full circle as to how we address that in a comprehensive way.
    I could imagine having senior people in the way that you've 
provided would actually be incredibly helpful. But we still 
need a comprehensive perspective both on the data side as well 
as on the computer security side.
    We just published a paper showing how data--the data to 
impersonate a voter can actually--could be used to actually 
change voter files by making changes. So that's kind of a 
combination of data that we were talking about in the 
Commission effecting a kind of security outcome.
    Mrs. Demings. Any of the other witnesses would like to?
    Mr. Shea.
    Mr. Shea. Yeah. Are you asking whether the State Department 
should govern the Federal Government-wide cybersecurity 
efforts?
    Mrs. Demings. If the State Department should eliminate or 
create the Office of Coordination.
    Mr. Shea. You know, you can take the boy out of OMB but you 
can't take the OMB out of the boy. So having seen the 
coordinating role OMB can take, whoever has the lead role at an 
agency level, OMB has an important role to take in this effort, 
or at least the White House--something that has the imprimatur 
of the White House.
    The Department of Homeland Security and the Department of 
Defense are also key players in this initiative, National 
Institute of Standards and Technology. All that needs to be 
governed somehow centrally. So that I don't really--I don't 
think the State Department seems to me is the right place for 
that, but there does need to be central governance of the cyber 
apparatus across government.
    Mrs. Demings. Dr. Sweeney, in the wake of the Equifax 
breach, one of the worst breaches of personal privacy in 
history, it is critical for Federal agencies to reevaluate the 
state of their security protections. And I'd like for you to 
begin, but any of the witnesses can talk about, in the limited 
time that we have, what immediate steps can an agency take to 
improve the security of their data?
    Mr. Haskins. Hire Latanya Sweeney to be in charge of their 
data.
    Mrs. Demings. That's why she was asked the question.
    Ms. Sweeney. So data are just valuable. I mean, they're 
worth a lot. There are different kinds of actors who are after 
this data and the mechanisms that they will go through to get 
the data make it constantly a kind of cat-and-mouse kind of 
game.
    And so how we engage--you know, are we using the latest--
are businesses using the latest? And even then, there are all 
these simple vulnerables, like getting someone through email to 
reveal their password and that particular account having 
particular access privileges. And that's a human engineering.
    So it's not quite as--I didn't answer your question, but I 
just bring emphasis on security is multifaceted in its 
complexity. The data problems that we talk about in the 
Commission are certainly a piece of security, because the more 
vulnerable data, whether it's health data on Americans, whether 
it's voter records on Americans, whatever it might be, income 
records, it allows other kinds of systems to be infiltrated by 
actors as well, whether it's our tax return system or--and what 
have you.
    So there is a relationship between security and the data 
things that we talk about. And this problem that we have where 
the privacy of the data is vulnerable on both sides, and what 
we choose to give out is not adhering to a kind of risk 
assessment and improvement model and what we choose not to give 
to let researchers help us learn more, is also a problem.
    Mrs. Demings. Thank you.
    Mr. Chairman, I yield back.
    Mr. Russell. The gentlelady yields back.
    The chair now recognizes the gentleman from Wisconsin, Mr. 
Grothman, for 5 minutes.
    Mr. Grothman. Dr. Haskins, in your testimony you said that 
most of the Nation's social programs produce modest or no 
impacts on the problems they were meant to address. I think 
given the mass of money we've flooded at these programs over 
the last 50 years, that would be obvious.
    But could you--first of all, how do you know they don't 
produce significant results? And then, secondly, could you give 
us some examples?
    Mr. Haskins. There's an interesting study of the first 90 
studies that were--Congress established the Institute of 
Education Sciences, and they launched into--they revolutionized 
research and education. And the first 90 studies, about 85 
percent of them failed to produce a major impact. This is 
entirely consistent with what's in the literature on medical 
research, on business research, and on research in the social 
sciences.
    So there's--I don't think anybody seriously questions that 
most of our programs don't work. If we want them to work, we 
need to continue to evaluate them and we need to improve upon 
what the programs are doing. We can find programs that work for 
almost any problem. And then we need to learn how to implement 
them widely in the country.
    Mr. Grothman. Okay.
    Next general comment for Dr. Abraham. Looking at it, you 
had 16 people on that committee is right?
    Ms. Abraham. Fifteen.
    Mr. Grothman. Fifteen, okay.
    I counted, including the adjunct professor at Notre Dame, 
11 of the 15 right now are professors. Do you think there's a 
diversity problem on the committee, or will that have any 
impact on----
    Mr. Shea. I was very loud.
    Ms. Abraham. No, I don't think there was. Many of the 
people who are currently academics are people who in previous 
lives have had extensive experience doing other things; a lot 
of experience in government. We were asked to tackle a set of 
fairly technical questions. So technical expertise seemed 
called for.
    And we heard from a--we made a big effort to hear from a 
lot of people to get input into our deliberations. We ended up 
hearing from--I want to say over 500 people, including at 
hearings where we invited anyone who wanted to come talk to us 
to do so.
    So I feel like we had a pretty broad set of perspectives 
represented, and we did hear a lot of input from people.
    Mr. Grothman. Okay. One of my concerns on your 
recommendations is it seems that if you talk to the right 
people you can get the conclusion that you want. I mean, we 
just heard from Dr. Haskins say we need to change programs to 
improve them. I think there may be a bias out there to how can 
we improve these programs rather than how can we get rid of 
these programs, which is a much more difficult thing to do.
    But I guess my concern here is that usually there are--
people of, let's say, goodwill can come up with different--
evidence for different results, you know. And we talk about 
global warming, and various smart people can say it's not going 
on, but I think the overwhelming bias in the community is 
saying it is. We talk about--another one that comes around is 
early childhood education, very smart people feeling we're not 
getting a lot of bang for the buck there. But I think the push 
for more, and people who want more government are always going 
to be there pushing their people to the fore, you're going to 
find people to say early childhood education is a positive.
    How do you guard against having what I'm afraid would 
happen any time you set up a permanent committee or permanent 
commission, they are taken over by people who want more, and 
they will find the experts who claim we need more, they will 
twist the data to say we need more?
    Ms. Abraham. One of the things that is emphasized in the 
report is the importance of rigor of the evaluations that are 
undertaken. So what we are envisioning and what the structure 
that we've laid out we believe would promote is a system where 
on an ongoing basis there are rigorous evaluations of what the 
outcomes associated with different programs are. And there are 
scientific standards for determining whether a conclusion from 
a study is valid or not valid. That still isn't going to tell 
you what it makes sense to do.
    Mr. Grothman. Other than, say--my concern is that, despite 
supposed rigor, which we should probably have throughout the 
whole world of academia, the vast sea of academia comes out 
left. So even though they're all supposed to be bound by rigor, 
unless our country, our idea of limited government is wrong, it 
seems--you see where I'm going with this? Is this a concern?
    Ms. Abraham. I think that evidence can tell you that if you 
do A, the outcome appears to be B. That can't tell you what 
your policy should be. So what we're really arguing for--this 
is something I've heard Ron say any number of times--is that 
what we're advocating for is evidence having a seat at the 
table so that everyone can look at it and then, given that and 
other value judgments and other inputs, make decisions about 
policy. But we are just--we're advocating for evidence to have 
a seat at the table.
    Mr. Grothman. You have to trust the people. But thank you.
    Ms. Sweeney. I know you're over time, but I would just add 
one thing. Clearly, you know, as a scientist, clearly, I 
believe in the pursuit of truth. And so having more data is 
better than not having the data.
    So we might argue about whether there were confounders in 
an outcome. But not having the data at all increases the space 
of uncertainty. And the issue here is making more data 
available to get rid of some of the uncertainty on which these 
issues would be based.
    Mr. Russell. The gentleman's time has expired.
    The chair now recognizes the gentlelady from the District 
of Columbia, Ms. Holmes Norton, for 5 minutes.
    Ms. Norton. Mr. Chairman, I want to thank you and the 
ranking member for this hearing.
    And isn't it interesting that, by coincidence, of course, 
it comes at a time when there's an outcry in the country about 
fake news and alternative facts. And here we're discussing the 
Commission's work, for which I thank you, on evidence-based 
data. The uses of data, however, takes into account more than 
the evidence they purport.
    Ms. Foxx on the other side asked this difficult question: 
How can we make sure decisions are made on the basis of data? 
Well, decisions are here made on the basis of many factors, on 
the values of each side, how one construes data, debatable 
data.
    We're in the middle of the Affordable Health Care Act, and 
even the underlying data is being contested. And one side is 
actually telling us that if you have a preexisting condition 
you'll be covered in the same way you're covered by the 
Affordable Health Care Act. You know, my side can only call 
that a lie. But, of course, the Affordable Health Care Act has 
gone down. So somehow or the other, there were enough people on 
the act or being served who were their own evidence.
    But trying not to be cynical, you can see how focusing on 
data alone may not get us the results we want. We just finished 
a big debate, and it comes up every single year, on whether 
Planned Parenthood should be funded. It usually gets funded 
because there's an outcry in the country if you don't fund it 
because of the work it does on so many conditions affecting 
women.
    But there was a whole special committee set up here in the 
Congress to investigate a Planned Parenthood episode. And now 
the so-called data from this committee is being used to try to 
defund Planned Parenthood based on whatever this special 
committee found out, completely contradicted by the entire 
scientific community. So you can see there's a little cynicism 
on my part when I hear talk of data, at least when used in a 
body like the Congress of the United States.
    Mr. Haskins, I have a question for you, because there's a 
current controversy as the administration prepares or asks 
Congress to cut the Teen Pregnancy Prevention Program. In this 
Congress, on both sides of the aisle, we would like to see at 
least fewer abortions and some no abortions at all. So we're 
together on that. But, of course, if teen pregnancy goes up, 
there will be more abortions. And yet you will find that the 
side who is against abortions does not move on teen pregnancy. 
Maybe we'll see what happens in the appropriation.
    But this is a very small investment, a small program. And, 
Mr. Haskins, you recently wrote an op-ed, and it was entitled 
``Trump Team Doesn't Understand Evidence-Based Policies 
Regarding Social Problems.'' And in that, you actually cited 
that 40 percent of the projects funded through the Teen 
Pregnancy Prevention Program produced, you said, at least one 
significant impact.
    Do you believe when a program produces at least one 
significant impact that eliminating 81 percent of the 
organizations that participate in the program is perhaps not 
the wisest idea?
    Mr. Haskins. They actually eliminated 100 percent. It was 
81 programs, and they eliminated all of them.
    Ms. Norton. All 81 of them.
    Mr. Haskins. Yeah.
    No, I don't think it is. And I think in this case it's 
really important, because teen pregnancy prevention in a way is 
the most advanced of the big sets of programs that we're trying 
to improve and develop ways to improve, starting with evidence-
based policy, which, as we all know, doesn't always work well. 
And that's what this study showed.
    But we've talked repeatedly that most social programs don't 
work; 80 to 90 percent is a good estimate. And in this case, 
only 60 percent failed, 40 percent worked. That's progress. And 
I think it's because the agency did a good job implementing a 
program, the people out there in the countryside who are 
learning this are going along. They were in the program for 5 
years.
    Ms. Norton. So you would think that if you were going to 
cut anything, you would say, okay, those didn't work, maybe you 
cut those. But those that did work, showed progress, maybe we 
ought to continue that progress by funding those programs that 
did work in cutting teen pregnancy.
    Mr. Haskins. That would make sense to me.
    Ms. Norton. Thank you, Mr. Chairman.
    Mr. Russell. The gentlelady yields back.
    The chair now recognizes the gentleman from Kentucky, Mr. 
Comer, for 5 minutes.
    Mr. Comer. Thank you, Mr. Chairman.
    My first question is for Mr. Shea. The Commission found 
that bureaucracy caused problems related to unevenness and 
capacity in agencies' evidence-building efforts. Can you 
explain some of the problems that the Commission found as it 
relates to this bureaucracy?
    Mr. Shea. I can give you a couple of examples. One was 
procurement. A lot of the questions about adequate rigor in 
evaluation are solved by ensuring that the people you contract 
for for evaluations bring sufficient independence and rigor and 
experience to conducting the evaluations. The procurement 
process is very cumbersome, as this committee; well knows.
    Likewise, the talent you need to oversee and conduct an 
evidence agenda is really hard to recruit and retain. There's 
an enormous competition for this kind of talent. And the 
personnel system under which we operate today makes it really 
difficult to hire, recruit, retain that workforce.
    Mr. Comer. Why does the Commission suggest that OMB be the 
agency that coordinates Federal evidence-building efforts, Mr. 
Shea?
    Mr. Shea. OMB has that tacit threat of budget impact from 
everything it does. So it's got a lot of juice that other 
similar entities don't have. And this committee, the Congress, 
in general, has vested in OMB a great deal of central 
management responsibility, and we think they can have a real 
important role to play in coordinating this government-wide.
    One of the things we did see is that OMB's efforts are a 
little disjointed. There's a performance and personnel office. 
There's an evidence team. There's the office of information 
regulatory policy, the Chief Statistician at OMB. Those aren't 
necessarily well coordinated at OMB. So OMB itself could do a 
better job coordinating its own investment in this enterprise.
    Mr. Comer. Let me ask you this, and this will be my last 
question to you. The Commission recommends considering 
reorganizing and consolidating aspects of OMB. Can you explain 
what the Commission means by reorganizing OMB? Is the 
Commission suggesting a complete overhaul of the agency, or 
what?
    Mr. Shea. No. Just what I said. These various entities that 
have responsibility for certain aspects--the office of 
information--Office of E-Gov, the office of information and 
regulatory policy, the performance of personnel team, the 
evidence team--all of those four could be better coordinated. 
Consolidation is only one option in the various things you 
could do to improve that coordination.
    Mr. Comer. My next question is for Chair Abraham.
    How did the Commission weigh privacy concerns against the 
need to ensure transparency of government information?
    Ms. Abraham. So as I indicated in my opening comments, we 
are all for making publicly available government data where 
that can be done without violating the confidentiality of 
individual people. In cases where it's not possible to do that, 
we adopted an approach that Latanya has laid out for sort of 
tiered access to data. If you can't just make the data 
available publicly, maybe you can make available a stripped-
down data file that lets people learn things about what the 
Government's doing but doesn't violate confidentiality. And if 
you can't do that, maybe you can make the data available 
through the National Secure Data Service.
    So we are really pushing as hard as we can in our 
recommendations for making data available in the best way that 
you can without violating confidentiality.
    Mr. Comer. In gathering testimony and during deliberations 
over the report, did the Commission solicit feedback from the 
transparency community? If so, who and what did they suggest?
    Mr. Shea. We had an open comment period and a number of 
witnesses, and we received a number of written input from what 
I would call the transparency community, ways to leverage the 
amount of data currently being reported and what 
standardization could do to improve our access to a variety of 
sources of data.
    Ms. Abraham. We do have a list of all of the people that 
provided input, and we would be happy to call that out and let 
you know specifically who we heard from.
    Mr. Comer. Okay.
    Ms. Sweeney. I would also say too that the report itself is 
very much about transparency. It makes more transparent 
arbitrary decisionmaking around what data is being given. It 
also makes more data possible to be given, which adds to the 
notion of transparency in oversight. So in that way it's very 
much, the report itself, very much is a champion for the 
transparency position.
    Mr. Comer. Okay.
    Thank you, Mr. Chairman.
    Mr. Russell. The gentleman yields back.
    And the chair now recognizes the ranking member for any 
followup.
    Mr. Cummings. I just want to--you know, I was just 
wondering, I was thinking about this whole opioid situation 
where so many people are dying from overdoses and the 
President's Commission has, in their preliminary report, has 
said that this should be declared a state of emergency.
    And I'm wondering how we can use the information that our 
fellow Americans are dropping dead everywhere. And there are no 
boundaries with regard to race, no boundaries with regard to 
location, rural, urban. They're dying big time.
    And then we have fentanyl, which is really taking a lot of 
people out, where they say that in some instances if you touch 
it you die.
    It just seems to me, you know, when evidence is in your 
face, how do we best--I mean, you know, we as legislators, I 
think it's important that we act on evidence. I really do 
believe that. I would say you can't make a decent decision 
without evidence.
    But I'm trying to figure out how does information like 
that, how is that--how do you all see that as best handled? Or 
you don't see that within your purview?
    Hello? Somebody talk to me.
    Mr. Shea. I never miss a chance, Mr. Cummings.
    I think what you're talking about is our central 
motivation. The Nation's challenges demand effective solutions. 
And our investments to date, as Ron has said, haven't really 
made much of an impact.
    And so the more and more we can learn about what has the 
greatest impact on solving our biggest problems, we want more 
and more of that evidence. And so what we think our 
recommendations do is create a structure, governance, a cycle 
in which we are developing more and more evidence of what works 
so that more and more people can learn and invest in those more 
proven activities.
    Mr. Cummings. Anybody else?
    Dr. Sweeney?
    Ms. Sweeney. So in Massachusetts we have an opioid crisis. 
And I was really taken by the difficulty of getting the data 
needed to address what was really an acute problem. And, even 
though it's not the Federal Government, it was very similar in 
the sense that what did it take to get these different agencies 
and different groups to share data widely.
    And so in very much the same spirit as the Commission 
report, we had two tools. One is we could try to change a law, 
or the other one was we could argue and explain how it is that 
this version of the data is okay to give out even within the 
structure of your existing rules and regulations. And we were 
very successful in beginning to put together combined data sets 
in Massachusetts under those kinds of models. So to the extent 
that that helps.
    Mr. Cummings. Mr. Haskins?
    Mr. Haskins. I just want to add one thing, quickly.
    Mr. Cummings. Dr. Haskins.
    Mr. Haskins. Pardon?
    Mr. Cummings. I said Doctor.
    Mr. Haskins. Oh, thank you.
    Quickly. There are certain problems that are almost 
impossible to overcome. Addiction is one of them. I come from a 
family of alcoholics. Some of them drank from the time they 
were teenagers. Many of them died as a result of drinking. And 
this is an even bigger problem than opioids. It doesn't have 
quite the flash that opioids do. But the fact is we do not have 
good solutions to get people to stop their addictions. And 
opioids appears to be even more addictive than alcohol.
    So even if we were willing to spend money, if you did all 
the investigation, we had all the evidence and so forth, it 
would still be a very difficult problem and we would have low 
success rates.
    I think they could be improved. There are some improvements 
that are better than the other. I'm going to Kentucky in 2 days 
to see some opioid treatment facilities. And maybe some of them 
are going to be successful and we will be more successful in 
the future. But addiction is always going to be a problem, and 
it'll be very difficult to solve
    Mr. Cummings. I want to thank all of you for being here 
today. It's been extremely helpful.
    And thank you, Mr. Chairman. I yield back.
    Mr. Russell. The gentleman yields back.
    And I'd like to do a couple of followup questions regarding 
the placement of this data service.
    Dr. Abraham, are you aware of the Census Bureau's 2020 
Census efforts which include the creation of administrative 
records verification and access systems, that these were placed 
on the GAO's high-risk list? So if they placed it on a high-
risk list, why would we want to then move immediately to 
something like that?
    Ms. Abraham. So I think that what we're talking about is a 
little different than the Census operations issue that you're 
raising, which I have to say I'm aware of only in quite general 
terms.
    What we are talking about building on is capacity at the 
Census Bureau that has been quite successful at figuring out 
methodologies for bringing administrative data in from various 
sources and linking it for statistical purposes. There's a 
group at Census that has a lot of expertise at doing that and 
has been very good at doing that and we think that that's 
something that would be a good core to build on in establishing 
this National Secure Data Service.
    Mr. Russell. Well, there's no doubt they collect an awful 
lot of information and have since we've been wearing tricorn 
hats. But I guess given the longstanding problems with the 
Census Bureau when it comes to estimating costs of large-scale 
projects, including the massive technological overhaul for the 
2020 Census, I go back to, why would the Commission feel that 
leveraging Census employees and resources would be a best 
option? Because, you know, we've got some costs, we've got some 
security concerns. And this would probably be true of any 
agency.
    But I'm just curious. I mean, you're all highly intelligent 
people. You've certainly done your homework. So given these 
issues and problems, why do we think that we would want to do 
this?
    Ms. Abraham. So the Census Bureau is not--it's a big place. 
There's a lot of different parts of the Census Bureau. Doing 
the Decennial Census is an enormous undertaking. I mean, every 
time the Census comes around, there are issues of one sort and 
another that arise and have to be confronted.
    The Census Bureau is hiring hundreds of thousands of people 
to do the Census. That's not what we're--it's a big operational 
and management challenge. That's not really what we're talking 
about. We're talking about starting out on a smaller scale, 
building the capacity to do data linkage for evidence-building 
purposes.
    And so I guess in my mind it's a very different enterprise. 
And the issues that you're raising don't really, as far as I'm 
aware, exist in that part of what they do.
    Mr. Russell. Well, and along that line, I guess, you know, 
the Commission made a recommendation that using Census Bureau 
systems as models for the data service, and, yet, these 
processes for verifying this are still--they're still trying to 
verify that the systems will work in the 2020 Census.
    So, I guess, in your opinion, before selecting a system as 
a model, would it not be better to wait until the model has 
been fully tested and vetted?
    Ms. Sweeney. Let me say the following. So the operation of 
the Census, the sort of production system of producing the 
Census, is not what got them into the report. What got them 
into the report is, of all of the parts of the government, they 
are, by far, the best----
    Mr. Russell. The best collector.
    Ms. Sweeney. No. They're the best if the world at providing 
a public version of that Census. You can go on to a website, 
and you can get information from that Census. And that 
information that I get on that website doesn't violate the 
privacy. I can't even do it, you know?
    And so this is an amazing feature, that the Census Bureau 
has had this capacity and has been a leader in Federal 
statistics offices on this idea of, how do I render data 
sufficiently de-identified that I can share it publicly? This 
is the skill that got them into the report.
    Mr. Russell. And that's useful. And, you know, we certainly 
see release of data and release of Census information. You 
know, you've got to live a long time before you can reach back 
into a previous Census to start mining the information.
    Mr. Haskins. Mr. Chairman, can I add something?
    Mr. Russell. Dr. Haskins, please.
    Mr. Haskins. Yeah. I can tell you why I supported the 
proposal. I never thought of having a National Secure Data 
Service before I joined the Commission, so I think I would be 
like a Member of Congress that would for the first time 
confront this choice.
    And here's why I decided. The Census Bureau is already 
doing almost everything that we want the National Secure Data 
Service to do. They're selecting the best proposals. They are 
helping people analyze their data. We envision that that will 
be an important thing in the future. They give secure access to 
data that they bring into the Census Bureau from other 
agencies.
    So those are three big functions that will be required of 
this agency and--of this new agency--and the Census Bureau 
already is as good as anybody in government, I would say 
probably best in some of those cases. And there are other 
things as well. They are just so experienced and so competent 
that it makes sense to start with them.
    Mr. Russell. And it's useful to hear each of your lines of 
thinking on why you chose this particular agency.
    I guess my last question, and to close out, would be the 
report suggests that FITARA may hamper agencies' efforts to 
limit access to confidential data. We've made a lot of progress 
on FITARA and those that have implemented its measures have 
actually been pretty successful. It's just trying to get people 
to comply.
    And so I guess it raises some other questions. What 
evidence does the Commission have to support the assertion that 
FITARA may hamper agencies' efforts to limit access to 
confidential data? Whoever would like to take that.
    Mr. Shea. It's not an issue that I can talk very deeply 
about, but I think generally centralizing a lot of authority in 
the CIO may conflict with data stewards or evaluation officers 
implementing the systems getting access to data they need to 
drive a learning agenda.
    Mr. Russell. And in this mining of and, you know, studying 
the issue, did you share these concerns with Federal CIOs? And, 
if so, what was some of the commonality of feedback?
    Mr. Shea. We had brought input, as you know. But I wouldn't 
call what we had a conversation.
    Ms. Sweeney. But I would say that when I was at the Federal 
Trade Commission the role of the CIO--so take the Equifax 
breach. The Equifax breach means that if I want to release the 
data, and I think that Social Security numbers are private, 
they're not so private, right? Especially if copies of the 
Equifax data are available for $500 on the black market, right, 
on the dark net.
    So in that kind of thinking, that's outside of the scope 
that our CIOs think, this idea of how I do think about the 
privacy of the data, the decision I'm making. You know, it's 
not the same as how do I choose a particular technology, or an 
infrastructure, or even the security of my laptop, or----
    Mr. Russell. No, that's useful, and I appreciate it.
    And I'd like to echo the ranking member's comments and to 
thank each of you, not only for the hard work that you've done 
in trying to improve our own government, but also to protect 
all Americans, you know, how can we best protect Americans with 
the data. At the same time, we have a responsibility to the 
Constitution to have an open government and these two halves 
trying to balance.
    So I really want to thank each of, and all of the members 
of the Commission, for the work that you've done.
    The hearing record will remain open for 2 weeks for any 
member to submit a written opening statement or questions for 
the record.
    And if there is no further business, without objection, the 
committee stands adjourned.
    [Whereupon, at 12:07 p.m., the committee was adjourned.]

                               APPENDIX

                              ----------                              


               Material Submitted for the Hearing Record
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]

                                 [all]