http://www.ntia.doc.gov/advisory/broadbanddata/WorkshopTranscript_10302009.txt
This is an unedited transcript of the October 30, 2009,
NTIA Broadband Data Transparency Workshop meeting.
0001
1
2
3
4
5 NTIA Broadband Data Transparency Workshop
6 October 30, 2009
7 U.S. Department of Commerce
8 Herbert C. Hoover Building, Room 4830
9 1401 Constitution Avenue, NW
10 Washington, DC 20230
11 1:00 - 3:20 p.m.
12
13
14
15
16
17
18
19
20
21
22
0002
1 Attending:
2 Government/Presenters:
3 Larry Strickling, NTIA
4 Susan Crawford, NEC
5 Andrew McLaughlin, OSTP
6 Daniel Weitzner (Moderator), NTIA
7 Jennifer Duane, NTIA Staff
8 James McConnaughey, NTIA Staff
9 Anne Neville, NTIA Staff
10 Invited Researchers/Presenters:
11 Dr. Robert Atkinson, Information Technology and
12 Innovation Foundation (via phone)
13 Professor Ken Flamm, University of Texas-Austin
14 (LBJ School)
15 Dr. George Ford, Phoenix Center for Advanced Legal &
16 Economic Public Policy Studies
17 Professor Blanca Gordo, University of California
18 Berkeley (Center for Latino Policy Research)
19 Professor Shane Greenstein, Northwestern University
20 (Kellogg School)
21 Professor Eszter Hargittai, Northwestern University
22 (Communications Studies Department)
0003
1 Dr. Nicol Turner-Lee, Joint Center for Political and
2 Economic Studies
3
4 Department of Commerce:
5 Patricia Buckley, Office of the Secretary
6 Mark Doms, Chief Economist, ESA
7 David Payne, ESA
8 Judith Means, OGC
9 Jeffrey Roberson, OGC
10 Gregory Weyland, Census Bureau (via phone)
11
12 Federal Communications Commission:
13 John Horrigan, BB Team
14 Kirk Burgee, WCB
15 Ellen Burton, WCB
16 Ken Lynch, WCB
17 Steve Rosenberg, BB Team
18
19 Department of Agriculture:
20 Peter Stenberg, ERS
21 Department of Education:
22 Chris Chapman, NCES
0004
1 Department of Labor:
2 Richard Home, Office of Disability Policy
3 Chris Carroll, CEA
4 Cary Hinton and Barney Krucoff, DC PSC
5
6 Non-Government:
7 Robert Roche, CTIA
8 Rick Cimerman, NCTA
9 Patrick Brogan, USTA
10 Cecilia Gomez
11 Kate Williams (via phone), University of Illinois
12 Jenny Toomey
13 Helen Brunner, Ford Foundation
14 Ben Lennet
15 James Losey, New America Foundation
16 Steve Smith, Rainbow Coalition
17 Lawrence Spiwak, Phoenix Center
18 Sharon Strover, University of Texas
19 Raquel Noriega, Connected Nation
20 Barry Goodstadt
21 Jessica King
22
0005
1 MR. WEITZNER: Good afternoon, everyone.
2 Thanks for being here. We're going to jump in and
3 get started.
4 Welcome to the NTIA's first Broadband Data
5 Transparency Workshop.
6 Before I get started, I want to remind
7 everyone that the proceedings are being Webcast real
8 time with a transcript, so we'd ask that all of you
9 who speak, even if you are sitting at the table and
10 have an identifiable tent card, just say your name so
11 that we know who is speaking and the people on the
12 Webcast know who is speaking.
13 We're delighted that you all can be here.
14 I'm going to start with some very, very quick
15 introductions and logistics.
16 We have a brief two-hour period. We have
17 a tight agenda, so we're going to race through all
18 this. Despite the fact that many of you have come
19 from far and wide, We want to make sure to hear from
20 all of you.
21 I'm going to introduce our guests. We have
22 actually rearranged a little bit the order of the
0006
1 agenda already so that our guests can go first and
2 then Assistant Secretary Strickling will follow.
3 So let me introduce, or clean up, as the
4 case may be, as the baseball fans would say -- let me
5 introduce our two guests who are going to kick off
6 the discussion.
7 Susan Crawford, who many of you know,
8 Special Assistant for Science and Technology
9 Innovation Policy. She has been the indispensable
10 leader for the first year of this administration, and
11 none of us would be here and we wouldn't have the
12 progress we have without Susan. So we are thrilled
13 to have Susan. And Andrew McLaughlin is the
14 Technology Officer in the Office of Science and
15 Technology Policy. I think to the extent that this
16 administration gets anything right about the Web, it
17 will be in large part because of Andrew. So it's
18 great to have you.
19 We have asked Susan to introduce comments
20 and kick off this discussion to try to frame some of
21 the questions that they think are important as policy
22 makers to understand and hopefully that you can
0007
1 contribute to, and then we will go to Larry
2 Strickling and on to our presentation. So Susan,
3 please.
4 MS. CRAWFORD: Thank you. And these are
5 very kind words. I cannot overstate the importance
6 and seriousness and priority of this meeting. And
7 also I am also delighted to be in the meeting where
8 academics -- we have a great need for academic input.
9 And also people from the outside focusing on the need
10 for data inputs.
11 Let me list four or five projects that I
12 think would be extremely helpful to this process,
13 Several of which the FCC and companies are looking at
14 as well, so we can do some coordination.
15 First of all, as we have seen in recent
16 weeks, the interaction between investments and
17 regulation is an ongoing, very serious question for
18 policy makers who are grabbing at straws to try to
19 understand the relationship between where we are
20 going as private leaders and whether that governing
21 role will have some effect.
22 A bigger vision piece is where do we want
0008
1 to get? It would be terrific to have researchers
2 help us say we need 25 megabits up or down, if that's
3 where you want to go. And where do we, as a country,
4 want to go with broadband, what services I need in a
5 data point don't give up. So where should we be
6 going and what will we be missing if we don't get
7 there in -- and what policy should be is another huge
8 area of extraordinary importance to this endeavor.
9 Real community comparisons, not just
10 anecdotal, but what difference it makes to a
11 community situated to another community with
12 broadband, and having that in terms again will help
13 policy makers. And finally, a cultural point, either
14 to George and Phoenix Center's report on senior
15 depression, how that is part of their infrastructure.
16 So those are my points. It would be
17 terrific to get the academic community and others
18 involved in producing exactly that kind of input.
19 And it is -- we are closely coordinated with the FCC.
20 They are clearly proceeding along all of these
21 tracks. But outside data gathering will continue to
22 be enormously helpful.
0009
1 I want to pass my microphone over to Larry
2 McLaughlin.
3 MR. McLAUGHLIN: And we will take questions
4 during the round table portion of the discussion. We
5 are going to race through this part of the agenda so
6 we can get to that.
7 Here's my appeal. My appeal is, I would
8 like to hear from people that have got a sense of
9 data sets that the government is sitting on that you
10 can't get access to. We are a limited number of
11 people, and so ideally we are looking for some sense
12 of prioritization from the research community about
13 which data sets held by which actors would be most
14 useful so that we know where to go and start knocking
15 on doors and so forth.
16 The emblem of the White House's commitment
17 to data transparency is data dot govern, which we
18 launched about 46 data feeds, and I think we are up
19 to more than 150,000. So data sets and data feeds
20 that are machine readable in standardized formats,
21 searchable through metadata, and a great resource.
22 And so an area of broadband transparency, as Susan
0010
1 just outlined, is a set of critical problems to try
2 to figure out.
3 We need to understand the economic inputs
4 to go into the broadband and economic system and the
5 economic outputs that come out of it. We need to
6 understand the reality on the ground. One of the
7 great puzzles, I think, of the Internet age is that
8 we have taken these protocols that were built for
9 data sharing, and we now treat it -- properly so, but
10 oddly -- as critical infrastructure.
11 But at the same time, because it is in an
12 environment completely interconnected voluntarily by
13 the network operators themselves, we have very little
14 window into the actual topology and physical
15 infrastructure that constitutes the network itself.
16 So this has many virtues, but it also has some
17 downsides when you are trying to work with data
18 strategy.
19 There are two important data problems that
20 can be solved by two different sets of actors. One
21 data problem is understanding, for example -- I
22 should say one data problem is that which can be
0011
1 solved by nongovernmental actors exclusively. So we
2 are not going to be able to test broadband speeds
3 from individual households around the country for a
4 bunch of practical and also perfectly good sort of
5 structural reasons.
6 There are, however, some interesting
7 academic and private sector efforts underway. We
8 would like to see more of them to provide tools to
9 people that they can use to test their broadband
10 speeds up and down, and help us to build a crowd
11 source data set that assesses actual throughputs
12 against advertised speeds and physical capacity as we
13 understand it.
14 There is another set of problems, the
15 second set which can be addressed by unearthing and
16 making useful data that the government is in fact
17 sitting on. So we've got some vague sense of what
18 some of those data sets might be, but that's where we
19 would very much like to get some input from this
20 group.
21 And so, I will just say, my sort of usual
22 concluding remark which is I'm on LinkedIn. If you
0012
1 want to connect to me, you can send e-mail to -- I am
2 looking for data sets within the federal government,
3 and we are looking for sort of collaborative
4 approaches outside the government that we can get
5 behind and highlight and valuize so you have the best
6 sense of --
7 Thanks, Andrew. Larry, please.
8 MR. STRICKLING: Thanks to Susan and
9 Andrew for coming over this afternoon. Their
10 presence and comments should demonstrate to each of
11 you the commitment this administration has made to
12 engaging with researchers and dealing with data and
13 getting data out to people.
14 So what I want to focus on, in my brief
15 time, was the fact that today, with respect to NTIA,
16 you are going to hear about three sets of data that
17 we either are collecting or are about to collect.
18 The focus today really needs to be on
19 understanding your needs for the data, your wants for
20 data. Some of these, you would hear from the
21 attorneys, have issues in terms of disclosure. We
22 are not going to solve those problems today, and
0013
1 let's not get distracted on those. Let's focus on
2 making the case as to why this information needs to
3 be in the public eye, and we will take it from there.
4 But the one thing I really hope people
5 focus on, and I can't imagine that every researcher
6 who is sitting around the table probably has had the
7 experience, as you have looked at data sets or
8 started to do reports, that you turn your head and
9 say, "Gosh darn it, why didn't they collect
10 information on this? If I only had this piece of
11 data, my research could have been so much better."
12 Or the idea that it would be nice to know
13 what the preexisting conditions really were, so I can
14 really chart and analyze and document the changes,
15 whatever they would have made on a particular
16 opportunity.
17 We are about to drop $7 billion in the U.S.
18 economy to fund broadband projects around this
19 country. The people receiving these dollars are
20 going to be obligated to make public reports every
21 quarter. We are at the stage now of trying to
22 determine what information shall we be collecting
0014
1 from these people as they report to us each quarter.
2 This is a perfect opportunity for those of you who
3 really want to track what is going on and the impacts
4 we have to give us that guidance now, to help us
5 create that data form, so we are collecting the
6 information that is going to be most important to
7 you, as researchers, to evaluate these very important
8 questions that Susan and Andrew have identified, as
9 we try to understand the impact of what works in
10 terms of broadband projects, what the impacts are,
11 both directly on employment and indirectly on the
12 economy in these communities and in the nation.
13 So this is a great opportunity, and we
14 welcome you all here today, and we welcome a
15 continuing dialogue with all of you as we try to sort
16 this out. Because this is a terrific opportunity and
17 we just cannot waste it. We have to take advantage
18 of it and really collect information.
19 There is no issue about this data, but
20 what will be made public, so let's understand what
21 the pieces of information are that we need to be
22 collecting so we have as rich a data set created
0015
1 through this investment in broadband infrastructure.
2 With that I will turn it back to Danny.
3 MR. WEITZNER: Thanks, Larry, very much.
4 So we are now going -- we are now in the
5 part of the agenda where we tell you something about
6 what we have, and what we'll -- what we'll be
7 releasing. And the goal there really is to partly
8 let you know, but you can find that out yourselves by
9 reading the Notice of Funds Availability and
10 documents. The idea is really to spur conversation
11 about in that universe what is useful to you, what
12 could we think of supplementing, and what is not
13 useful to you.
14 Just a couple of procedural points.
15 Number one, when we get to questions, which will be
16 at the end of this slide presentation, there will be
17 a wireless mic, is that correct -- a wireless mic, if
18 you are not on mic. I will remind everyone to speak
19 into the microphone.
20 We have a set of three presenters coming
21 up. We will begin with the discussion of the BTOP
22 program with Laura Duane -- I'm sorry, Jennifer
0016
1 Duane. I know you are not Laura. Laura Pettus was
2 not able to be here, but Jennifer Duane happily is.
3 Then we will have Anne Neville, who will
4 talk about mapping data, and Jim McConnaughey will
5 talk about CPS data, and then we will have a brief
6 discussion with some of the legal counsel about some
7 of the issues involved with data collection and
8 release.
9 We have, as you can see around the table,
10 a tremendous representation from different parts of
11 the government. We have colleagues from the FCC, and
12 we are very glad you could be here; the Department of
13 Labor, the Department of Education, Council of
14 Economic Advisors, other parts of the Commerce
15 Department, and the Economic Statistics
16 Administration.
17 When we get into the discussion, you should
18 feel free -- first of all, I hope those individuals
19 will participate and people should also feel free to
20 direct questions to them as well.
21 So, Jennifer.
22 Can I suggest you actually stand up there
0017
1 so you are next to the slides? And I will remind
2 everyone, somewhat rudely, that we are going to
3 finish this part of the presentation in 20 minutes.
4 So that gives you about 6 1/3. Six because the
5 lawyers get 1. 2, rather.
6 MS. DUANE: As Danny mentioned, I am
7 filling in for Laura Pettus. I learned that moments
8 ago, and so I will try to get organized. I apologize
9 if it seems a bit disjointed.
10 My name is Jennifer Duane, and I am a
11 policy specialist with the BTOP program. I will give
12 you a brief overview of BTOP and focusing on the data
13 collection components of that program.
14 As most of you know, BTOP is the Broadband
15 Technology Opportunities Program, and it is allocated
16 $4.7 billion to NTIA to give out grants for
17 infrastructure-related projects. We have been
18 focusing on last mile and middle mile projects for
19 this round.
20 At least $200 million will be given to
21 public computer centers, and there is $250 million
22 that has been allocated for innovative programs to
0018
1 incur sustainable broadband adoption. There is
2 $350 million available to develop a National
3 Broadband Map.
4 Now, BTOP has five main purposes of the
5 program. It's to provide access in unserved areas of
6 the country, broadband access, that is, And provide
7 improved broadband access in under served areas of
8 the country; to provide broadband education,
9 awareness, training, access, equipment and support;
10 it's to improve access and use by public safety
11 agencies, and finally, its purpose is to stimulate
12 broadband demand, economic growth and job creation.
13 Now, there's a number of reporting
14 responsibilities under BTOP that we have. First, we
15 report to Congress every three months on the status
16 of the program. NTIA is charged with establishing a
17 publicly-available database of grant applications and
18 recipients. We have that up at the moment at
19 www.broadbandandusa.gov. And that will continue to
20 be updated as we get more information as we move
21 through the process of awarding the grants. We will
22 be continually updating that database with more
0019
1 information. The grant recipients are required to
2 report on the Recovery Act measures, and that is at
3 www.FederalReporting.gov. And finally, we require
4 grant recipients to report quarterly on financial
5 information and programmatic progress, whether they
6 are meeting their time lines and other measurements,
7 with how they are progressing in either building
8 infrastructure or installing public computer centers,
9 whatever project they are engaged in.
10 Now, these are the types of recipient data
11 that will be collected. We have some reporting
12 requirements under development. We are still
13 developing the reporting forms that we will be using,
14 and these reports comply with OMB information
15 collection requirements.
16 Some of the examples of the Recovery Act
17 and BTOP data elements are for Recovery Act purposes,
18 the number of jobs created and retained by the
19 recipient project, the cost of the infrastructure
20 investment made by state and local governments.
21 There are also BTOP-specific reporting requirements.
22 We will be getting information on the applicant's
0020
1 progress, reviewing their project goals and meeting
2 their objectives and milestones. Also we want to
3 know how the grant funds have been expended, the
4 amount of the nonfederal investment funds that
5 applicants have added to complete the project as
6 well.
7 Now, in our notebook we listed -- our Joint
8 Notice of Funds Availability we listed a number of
9 reporting elements that we would be seeking from the
10 applicants. We just highlighted a few here. They
11 are -- obviously, there is a more extensive list in
12 the NOFA. For infrastructure projects we are asking
13 for information on availability of broadband
14 offerings, including the technology that is being
15 used, is it wireless, wire-lined, what frequencies,
16 what type of infrastructure is being implemented, the
17 location of that infrastructure, and the area that is
18 being served, the number of households that are being
19 passed. For sustainable adoption projects, we want
20 to know the technology that is being fostered by the
21 project, whether it's wireless or wire-lined. What
22 you are asking these users to adopt. The increase in
0021
1 the number of homes, businesses, community anchor
2 institutions that are subscribing to broadband
3 service as a result of the project. And for public
4 computer centers, we are asking for the number of
5 work stations available to the public, the total
6 hours of operation per week that the public computer
7 center is open, and the speed of the broadband
8 service that is being provided to the public through
9 the public computer center.
10 This really just gives us a broad
11 highlight of the data that we might be collecting
12 from the BTOP program. But as Larry and Danny have
13 mentioned, we really want to know from you what data
14 do you want from BTOP and what data would be useful.
15 To the extent we get a dialogue on that, we will
16 welcome your comments and thoughts.
17 MR. WEITZNER: Write down your answers to
18 those questions. Jennifer, thank you very much.
19 Don't forget the answers to those questions. Don't
20 get distracted by the next two presentations, which
21 are going to be distracting in a good way, of course.
22 I will remind everyone as well, we will
0022
1 post the slides on the workshop Web page. If they
2 are not already up, they will be by the end of the
3 day, so you don't have to take notes on all this.
4 Anne, please go ahead.
5 MS. NEVILLE: This will be a brief
6 presentation on the data we will be collecting on the
7 state broadband data and development grant program.
8 There was one eligible entity per state.
9 So we received applications from all 50 states, five
10 territories and the District of Columbia.
11 We are initially funding two years of
12 broadband mapping efforts and up to $500,000 for
13 broadband planning efforts. The awardees are
14 required to collect and verify broadband availability
15 data for their states, and there is very detailed
16 information in our notice of funds availability and
17 our technical appendix. That was originally released
18 on July 7, and then a clarification was made on
19 August 9. So I will go through in the next slide the
20 specifics of the type of data that we expect to
21 receive from grantees through this program.
22 So for wire-lined services, applicants or
0023
1 grantees, as I guess they will be soon, can submit
2 data to NTIA by address or census block. And in a
3 case where census block is greater than two square
4 miles, they can submit the information by street
5 segment. And included in this data set that gets
6 sent to NTIA will be information such as the provider
7 name, the end-user category, the type of technology,
8 the maximum advertised up- and downstream speeds and
9 the typical up- and downstream feeds.
10 For wireless services, the data will be
11 submitted by state, generally visual shape files to
12 detect areas of which broadband services are
13 available, the type of technology, the spectrum used,
14 and the maximum advertised speeds and typical speeds.
15 Additionally, we will also be collecting
16 information specific to community anchor
17 institutions. That would include schools, libraries,
18 healthcare providers, public safety entities,
19 community colleges, other institutions of and
20 community support organizations and entities, so
21 applicants, in their application, were encouraged to
22 name specific community support entities that they
0024
1 would want to include in this data set. And for
2 those organizations rather than just availability, we
3 will also be looking at what they are currently
4 subscribing to. So what do they currently have
5 coming into those facilities? We will be looking at
6 the type of technology subscribed to, the current
7 advertised up- and downstream, and the current up-
8 and downstream speed subscribed to.
9 The notebook also contemplates collecting
10 pricing data that would also eventually be placed on
11 the National Broadband Map. This is a requirement
12 for applicants or grantees, but it is something that
13 NTIA contemplates collecting and ultimately also
14 displaying on the National Broadband Map.
15 And so with all of this broadband
16 availability data, our big question for you is: What
17 do you want the National Broadband Map to display?
18 Of the data that I just talked about, what
19 is it that is useful to you in terms of being public
20 and why is it useful to you? How could it benefit
21 your research? That was our two minutes on broadband
22 mapping.
0025
1 MR. WEITZNER: Thank you, Anne.
2 While Jim is coming up, let me extend a
3 special thanks to Jim McConnaughey, the chief
4 economist at NTIA.
5 SPEAKER: He's the only economist.
6 MR. WEITZNER: I thought you were the only
7 economist.
8 (Laughter.)
9 MR. WEITZNER: Jim is really the reason we
10 are all here. Jim, as you know, has been working in
11 this area of Internet and broadband data for many
12 years, and we owe a lot to Jim for pulling this
13 together.
14 MR. McCONNAUGHEY: Welcome and thank you
15 coming here. And Happy Halloween. We are here for
16 treats, and we will hold off on the tricks.
17 Our current population survey, I think a
18 lot of people are familiar with this in terms of the
19 unemployment figures that people have heard over the
20 years. It's a census survey that has been sponsored
21 by BLS for 50 years.
22 Supplements. There is a myriad of
0026
1 supplements here. As you can see, it ranges from
2 displaced workers to veterans, to fertility, to
3 school enrollment. The best one of all is Internet
4 use. The CPS is -- I'm a little biased, but I think
5 it's a superb, well-oiled machine. It's a large
6 sample. They routinely try to get 60,000 households.
7 They don't always succeed, but they usually get very
8 close. It's scientifically designed. They have a
9 rotational scheme. They use weightings by states.
10 The census is the gold standard, in my
11 mind. They know what they are doing. Small standard
12 error. You would expect, out of a large sample, to
13 have a larger margin for error, but they hold true to
14 form. High response rate for basic CPS and for the
15 Internet use supplement. Routinely the basic CPS has
16 a 92 percent to 93 percent response rate.
17 When I was a consultant, I used to do
18 these amateur surveys. If I got 30 percent response
19 rate, I would have drinks on the house. They do a
20 great job. The Internet use supplement is very high
21 in terms of a very high response rate.
22 I won't bore you with this. Only the real
0027
1 officionados will care about this. They don't use
2 weights on the entire country. They break it out by
3 53 separate samples. 48 states, the District of
4 Columbia, New York gets special treatment as does
5 California. They hone in on trying to make this a
6 scientifically designed survey.
7 All right. Let's hone in on the one we
8 like, the Internet use supplement. We first engaged
9 the Census Bureau back in 1990, early 1990s, to
10 collect data on a new form of communications called
11 the Internet. Our early surveys, I have to admit,
12 were kind of primitive, looking back. We asked them,
13 do you have a modem attached to your computer?
14 Later we got to thinking, maybe we ought to ask them
15 if they actually use the modem.
16 I was curious, my memory failed me, so I
17 looked for the title -- the first falling through the
18 net report we did, which, for those who remember, was
19 very simple, table in the corner, mainly cables,
20 things like that. It was a very small one. The
21 title of the thing was, "A survey of the haves and
22 have-notes in rural and urban America." In the
0028
1 report itself, we actually asked, or offered, a
2 policy advice, which we didn't do what we did often
3 later on in our various reports. But we said, in
4 effect, that, well, it's too bad more homes aren't
5 connected -- I'm paraphrasing -- but have you thought
6 about community access centers, including schools and
7 libraries and community centers, as a possibility to
8 help bridge the digital divide? It was pretty giddy
9 stuff back in the early '90s.
10 After that, we went through several more
11 reports based on Internet use. There was one called
12 Computer and Internet Use Supplements back then.
13 October 1997, December 1998, August 2000, the term
14 digital divide was coined and policy started growing
15 to address the problem.
16 September 2001, a rosier picture was more
17 like the glass was at least half full. So it was a
18 nation online as opposed to the digital divide.
19 There was a different administration. We had to
20 change the title of the report.
21 October of 2003, we had a nation online.
22 October 2007, a networked nation. That was our
0029
1 broadband report based on the president's goal from
2 March of 2004, and it comprised a very small part of
3 the broadband report. And in October 2009, the
4 Internet Use supplement. They just stopped
5 collecting data yesterday. So this data is wetter
6 than wet paint.
7 Okay. We will look at the two most recent
8 Internet use supplements. Just quickly,
9 October 2007, 54,000 households, 128,000 persons.
10 These are good sizes. Not the decennial, not the
11 ACS, but it's big. The focus of dial-up versus
12 broadband at home versus nonuse. Only four
13 questions. We find out through the census, it makes
14 a difference is what point in time you get on their
15 dance card. If there is no other supplement there,
16 you can ask more questions.
17 2007 we are bootstrapping on to the School
18 Enrollment Supplement. The school enrollment piggy
19 backed on the basic CPS. And we boot strapped on the
20 class enrollment, so there was a limited number of
21 questions we could ask. It's frustrating, but that's
22 the way it is.
0030
1 Basic finding of October 2007: A little
2 over half of the households have broadband, and rural
3 use, true to form, trailed rather badly in 2007.
4 Okay.
5 October 2009, an estimated 54,500
6 households, 135,000 persons. This will be another
7 good size. Similar focus as before. We jacked it up
8 all the way to five questions this time, because,
9 again, we are a supplement to a supplement. It
10 includes my favorite question, the major reason for
11 using broadband at home, Which NTIA pioneered in the
12 1990s. Results: Stay tuned. We hope to get summary
13 data by the end of this year, and the public use
14 file, which will be open to the world, will be toward
15 the end of the first quarter of 2010 or April, one of
16 the two.
17 Just quickly, I know I am droning on here,
18 but the five questions that we posed for 2009: Does
19 anyone in the household use the Internet in any
20 location at all? And then who is that? Who is that
21 in the sense of how many in the household use the
22 Internet. Does anyone in the household connect to
0031
1 the Internet from home itself? Do you currently
2 access it from dial-up or do you use broadband?
3 What is the main reason you don't have high speed
4 Internet access at home for those who answered no to
5 the question.
6 Quickly, we have household and person data.
7 Seven demographic categories: Family, income, age,
8 race, gender, education, household type and
9 employment.
10 Geographic categories. We have a national
11 aggregate figure. We get rural versus urban. I
12 don't want to go off the board there, but if you can
13 read that, it would be states and regions as well.
14 We get it down to states in terms of good, reliable
15 data.
16 Okay. Links to the data. This would be on
17 our Web site for the historical data, and when it's
18 available, the October 2009 data. Basic questions.
19 You are becoming familiar with this one. What data
20 do you, the researchers and the public, want future
21 surveys to capture, and please tell me how that would
22 be useful for what you are doing.
0032
1 MR. WEITZNER: Again, write them down, and
2 don't forget the answers. So we have finally a very
3 short discussion from Jeff Roberson from our general
4 counsel's office, who will give as big picture of
5 some of the issues we will have to think through in
6 releasing data in the disclosures that we do or in
7 response to other kinds of requirements.
8 I will just say here as a caveat, as Larry
9 said, we are not really here to debate what our legal
10 obligations are. I'm sure we will do that in other
11 forums at some point or another, but we did want to
12 give you a picture of the kind of considerations that
13 we will face in looking at a data collection and
14 dissemination. So, please, Jeff.
15 MR. ROBERSON: As previously mentioned
16 lawyers, don't hold that against me. I am here to
17 discuss the legal framework and not here to depress
18 you. But just so you know the legal requirements
19 that our data exists within.
20 The first one is the Paperwork Reduction
21 Act. To the extent you want us to gather more data,
22 it has to go through an OMB process to keep paperwork
0033
1 burden down on the public.
2 Second, to the extent that NTIA would want
3 to proactively release information, it's subject to
4 the Trade Secrets Act. We are not allowed to release
5 proprietary business information.
6 Third, to the extent that you would ask us
7 to release information to you we, are subject to the
8 FOIA, Freedom of Information Act, and there are
9 certain prohibitions that would prevent us from
10 disclosing it to you. The first one is b(4), trade
11 secrets and confidential commercial financial
12 information, again, business and proprietary
13 information. The second is b(5) that allows for
14 certain privileges under the deliberative process.
15 We can choose to not release information that would
16 in some way hurt our deliberative process, so that it
17 would chill speech or make people less likely to
18 participate or cause conclusion to the public.
19 And the third most likely Freedom of
20 Information Act exception is b(6). It protects
21 personal information privacy information, Social
22 Security numbers and other stuff you probably
0034
1 wouldn't want to know about anyway.
2 MR. WEITZNER: Thank you, Jeff. I don't
3 think anyone looks too depressed. You all look okay.
4 So you did your job, Jeff.
5 Finally, we get to hear from you, which is
6 good news for us.
7 Let me just talk about how we are going to
8 proceed with this part of the discussion.
9 I do have one caveat. As you may well
10 know, we are in the middle of making decisions about
11 a number of BTOP grant applications. It should go
12 without saying that we are not here to discuss the
13 specifics of any applications, if anyone asks any
14 questions or offers any information. We are going to
15 have to move very quickly on from that. So we are
16 not here for that purpose.
17 We do have -- I will just remind everyone
18 that we have people watching and listening to the
19 Webcast, so when you come to ask a question, make
20 sure you are at a microphone, either at the table or
21 over in the back where there is a wireless mic.
22 These are some of the questions that we
0035
1 hope to be able to address in this discussion, and
2 what I have done, just a little bit in advance, is
3 asked a couple of the people who were attending to
4 kick off one question after another. We are going to
5 try to kind of stick to one question, and then move
6 on to the next. So, I don't want to give the
7 impression that I have been completely spontaneous.
8 So, I will say we will ask Shane
9 Greenstein to address the first question. Ken Flamm
10 will talk about the second one. Nicoll Turner-Lee
11 the third question, and the fourth question will go
12 to Eszter Hargittai, and if Rob Atkinson is on the
13 phone, I hope, are you there, Rob? Not yet. And
14 finally George Ford will address the last question.
15 What I have asked each of these
16 individuals to do is frame these questions and then
17 we will open it up for discussion.
18 So, Shane, talk to us about the types of
19 broadband data that you think NTIA could be compiling
20 and how you ought to think about how it would be
21 useful for researchers.
22 PROFESSOR GREENSTEIN: Thank you for
0036
1 letting me attend this, and I am pleased to affect
2 data collection at the federal level.
3 I am going to focus particularly in that
4 broad topic on the biggest unanswered question in the
5 research community, and that is: What is the
6 consequence from the adoption and use of the Internet
7 for the productivity, wage and income growth at the
8 household and regional level across the country? We
9 don't know.
10 We don't know, partly, I'm sorry to say,
11 because the CPS supplement is not designed to help us
12 answer that question, and, actually, some small
13 changes could make an enormous difference in our
14 ability to answer that question.
15 The other half of answering that question
16 has to do with broadband mapping, which you are going
17 to undertake, and so I actually don't have much to
18 say about that, other than if you make it available
19 at a fairly small geographic level, that is, at a
20 census block, at a minimum level, prices,
21 availability and capabilities, that will be useful.
22 Let me go to the CPS supplement. On the
0037
1 CPS supplement right now we have good information at
2 the MSA level. That's what is available. However,
3 it does not over sample small cities and rural areas.
4 Consequently, we actually have terrible standard
5 errors when we go to any city underneath about
6 200,000 population.
7 So I would -- my first thing to say is
8 actually over sample less populated areas so we can
9 do a reasonable comparison across -- so we can get,
10 you know, 10, 15, 20 observations in a city with a
11 population of 80,000 to 100,000. That way we can
12 actually do a standard error. That's easy.
13 The next one: We would love finer data.
14 If we had at least at the city level, not the MSA,
15 city, and arguably census block level we could
16 associate those data with availability and do a
17 direct mapping from the availability of broadband to
18 use to consequence for wages, productivity at a very
19 fine level of the economy, and we could actually
20 answer the question all of us want to answer and
21 Paul's community wants to answer.
22 I know you are going to say privacy is a
0038
1 problem and there is a readily solution available,
2 and that is work through the data centers that the
3 census department has already set up. The security
4 is available there. There are 13 around the country.
5 They are windowless offices. You are never allowed
6 to bring anything in. You are never allowed to bring
7 anything out. There is somebody watching everything,
8 and everything has to be cleared before it comes out.
9 If we could do the associations inside the research
10 data centers and walk out with the aggregate
11 statistics, that would be sufficient for most of what
12 we are, most of what all of us are interested in.
13 Finally, not to over use my five minutes,
14 we all recognize that there is a changing boundary of
15 work and leisure inside the home. And we don't know
16 anything about it, other than what we have learned
17 from the John Horrigan's survey at the Pew Center.
18 It would be nice to do this at the level of 60,000
19 households. The basic things you want to know are
20 things like allocation of time between work and home
21 use, between different types of Internet use such as
22 e-mail -- I mean just e-mail, commuting and part-time
0039
1 work. And finally, type of use such as, you know,
2 BlackBerrys or mobile devices. If you wanted a model
3 of how to do this, I would suggest the City of Hong
4 Kong survey as an exemplary example of time. They do
5 time of use devoted to different categories of
6 Internet surveys. It's a marvelous survey. I had an
7 undergraduate do an honors thesis on it. That's how
8 easy it is to use. And so really it's an example for
9 the U.S. government. It's a wonderful survey.
10 I don't think I exceeded five minutes.
11 Okay?
12 MR. WEITZNER: Thank you very much.
13 Questions, comments?
14 Mr. McCONNAUGHEY: Shane, as usual, you cut
15 to the chase. Thank you for doing that.
16 We hope -- one point in time during our
17 old falling through the net glory days, we did
18 applications of the sort you are talking about, how
19 people, specifically what activities they undertook
20 while they were on the Internet. There was no
21 broadband back then, but it could be carried forward.
22 So we hope to resume that in the future, again,
0040
1 Census and budget permitting.
2 MR. WEITZNER: So is everyone besides me
3 clear about what would be needed to actually be able
4 to associate this data, the broadband data that we
5 have with the protected census data?
6 PROFESSOR GREENSTEIN: So what you would
7 have to do -- the key thing at the household level,
8 if you knew where they were -- that's going to
9 violate every privacy norm. And we all know it.
10 But the CPS does originally have that data in it.
11 And what would you want to do as a survey researcher?
12 What you would want to do is have an address and
13 location associated with the mapping data at the
14 census block level. And then what will you do? You
15 are never going to publish that. We all understand
16 that. But what you do is a broad statistical model,
17 and then walk out with the broad statistical
18 association across the whole country. That will pass
19 a clearance test.
20 I mean, I have supervised dissertations
21 using the census data doing this, and that's the
22 model. We know how that model works. It's
0041
1 cumbersome because we have a responsibility to
2 protect everyone's privacy. So the only thing we
3 walk out with are these very broad associations you
4 do at a broad level. And someone's always clearing
5 it and someone's always watching, and that works well
6 enough for the kind of data, for the kind of issues
7 we are all interested in.
8 So if you want to ask the question, which
9 regions of the country have the biggest growth, which
10 ones have the biggest productivity, the way you do it
11 is, you would do these associations, and then you
12 walk out with these very broad statements that said
13 people who tended to work in the following type of
14 industry tended to have the following wage growth or
15 not. That's what you would be able to say. But you
16 can't make such an association until you actually get
17 to do the very very fine geographic matching. Is
18 that a helpful way to --
19 MR. WEITZNER: Yes.
20 PROFESSOR GREENSTEIN: I'm trying to be
21 very direct here.
22 MR. WEITZNER: That's good. Any comments?
0042
1 Do others think this would be an important thing to
2 do, or are you just going to let Shane carry the
3 water on this one? Please. If you could use a mic.
4 MS. GORDO: Blanca Gordo. Because there
5 needs to be research to understand the contextual
6 indications of that use within the home, and because
7 it's a family institutional structure, there are
8 different ways by which that regulates how people get
9 to use it. We don't fully understand what are the
10 roles of its use within the home, what kind of
11 competitor factors are involved. You can say every
12 home has a computer and every household has a
13 computer, but who gets to dictate who gets to use it
14 at what time? Where is it positioned if you want to
15 get at it? Its productivity. We need to explore
16 those types of issues to also observe the ways in
17 which people recognize some of these issues.
18 You know, we use very simple questions, but
19 we have also haven't specified culturally how to
20 capture that. I do agree we definitely need to have
21 more geographic data that is more inclusive, but if
22 we look at a regional analysis, like Los Angeles, it
0043
1 is highly connected, but there is lacking missing
2 data for the pockets of poverty, so you would not be
3 able to capture that kind of inequality.
4 PROFESSOR GREENSTEIN: I agree with that.
5 I was silent on these questions because I am sitting
6 next to Eszter Hargittai, and I am expecting Eszter
7 to talk about it. And I have a little bit of time.
8 I don't disagree with that. There is an
9 interesting question also that in principle the CPS
10 data could address it if it understood multiple users
11 inside the same household and spread between users.
12 We are past the point now where -- you
13 know, two-thirds of the country has adoption. There
14 is an interesting question how first adopters get
15 second adopters. But we are also past the point when
16 that is the only question.
17 The other question is, if you have
18 multiple users and they vary, how do you see the use
19 and transmit across households. That's a totally
20 fascinating area, because we are observing that
21 affecting the wages they walk out with and
22 differences in the productivity of regions.
0044
1 So I totally -- I am focusing on economic
2 support because it's my comparative manner. But I
3 totally agree with you, that that's an interesting
4 thing to examine as well. And we are going to miss
5 some of the things you are saying. That's fine.
6 MR. WEITZNER: Nichol and Richard?
7 Sorry.
8 MS. TURNER-LEE: I was just going to add,
9 having had my foot in the academic community, getting
10 my Ph.D., and coming back into the policy world, I
11 think the other thing that is different from the
12 conversation we have had is frequency, and one of the
13 places that I think we might want to figure out is
14 the speed at which broadband changes, how do we
15 update the processes in which we collect data, I
16 immediately went to my research assistant and asked
17 her what were some of her challenges, and she said
18 the fact that she is still working off of 2007 CPS
19 data, because the new data hasn't been released, and
20 when I look at questions -- we are now immobile. So
21 if we could figure out how to be more creative at
22 NTIA or other places that collect data, there may be
0045
1 a time where we have to figure out also with the
2 investments that NTIA is making in BTOP, are there
3 opportunities to do instant consumer surveys that are
4 a little more or less traditional but allow for real
5 time action?
6 When thinking about this question when I
7 saw it, the possibility of bringing access to an
8 anchor institution and asking people to fill things
9 out or requiring that as part of the data collection
10 process allows us as a country -- going back to
11 Susan's earlier question -- to be on point with how
12 we are progressing. So I wanted to put it out there
13 for frequency so we don't have these lags,
14 particularly in the area of technology, where we
15 can't afford to continue to lag behind.
16 MR. WEITZNER: Richard and then Andrew.
17 Richard.
18 MR. HOME: Good afternoon. I am Richard
19 Home with the U.S. Department of Labor Office of
20 Disability Policy. I'm the one at the table that
21 says what's missing, and usually 99.9 percent of the
22 time we totally overlook the population in this
0046
1 country of people with disabilities, and the majority
2 of those people are not even in the labor force,
3 which makes it even more difficult. So I want you to
4 include the people with disabilities in the
5 categories that you are looking at.
6 It was a ten-year effort to get a set of
7 disability questions in the population survey, so
8 that information is now there, and I encourage you to
9 include that in your research.
10 The second area that I see grave concerns
11 about, for the particular community that I work with
12 is access, accommodation and cost. If the equipment,
13 if the technology and the linkage to the broadband
14 are not accessible to people who use different types
15 of technology, including screen readers, motion
16 detectors, lasers, you are not asking any questions
17 about what are the barriers to access within that
18 household, should they have an individual in the
19 household that has a disability, and there is a
20 likelihood that they will; a child in special
21 education, a young adult in vocational rehabilitation
22 services, or an older American who has developed
0047
1 disability as part of the aging process.
2 Another group that I would ask you to
3 include in your communities are workforce. I am very
4 surprised that workforce development was not listed
5 among the agencies. You are working with schools,
6 but we have thousands of one-stop career centers
7 across the country run by the employment and training
8 administration, that uses broadband very creatively,
9 in terms of looking for work and finding jobs and
10 getting workforce training. And we collect wage data
11 from those populations, so I think you need to figure
12 out how to include my department more proactively and
13 in an approved way, particularly when it comes to
14 research with marginalized populations.
15 MR. WEITZNER: Thanks, Richard. We are
16 glad you are here.
17 MR. HOME: Just a little nudge.
18 One of the things we have been
19 grappling with in data.gov is the longstanding, and I
20 think very admirable, commitment by the federal
21 agencies that generate statistical data to extremely
22 high levels of data quality. And in the data.gov
0048
1 world, the kinds of feeds that we are encouraging
2 people to produce is sometimes seen as -- you know,
3 as a sort of infringement on that commitment, that we
4 are asking people to get data out before it has been
5 fully baked and before it has been fully refined and
6 so forth. And it's both for statutory reasons and
7 also just for the credibility that citizens place in
8 data that comes out of the government, it's important
9 to get this balance right.
10 We do have a tradition, though, of being
11 able to release data before it is completely baked
12 with appropriate disclaimers. We do this with GDP
13 growth, inflation numbers and so forth, which are
14 subject to revision down the line as we get more and
15 better data and work it over.
16 One of the things that is interesting to me
17 is, I would like to see greater frequency of data as
18 well. We've got to come up with a way. I think this
19 is really not just a broadband mapping problem. It
20 applies to a lot of things that are in data.gov to
21 come up with ways to describe the confidence we have
22 in the data and try to cabin its use to fit the
0049
1 qualifiers that we want to attach to it. I don't
2 have any brilliant way to do that other than to sort
3 of experiment with it in the data sets where that
4 seems appropriate.
5 MR. WEITZNER: Thanks. Ken, I'm going to
6 come to you in a minute. Let me ask, I will remind
7 people around the outside of the room, please feel
8 free to get up and grab a mic somewhere. Just
9 because you don't have a tent card doesn't mean we
10 don't want to hear from you. John.
11 MR. HORRIGAN: A quick observation about
12 the types of collection to do. Focusing on minority
13 or special populations, we do have a large sample
14 size, yet there could be culturally specific
15 questions you would want to ask to specific small
16 population groups that might warrant a separate
17 survey. So getting at those kinds of issues with
18 surveys of certain subpopulations in cities would be
19 worth considering.
20 MR. WEITZNER: I have a question. We are
21 going to kind of segue very shortly to the second
22 topic. I am just curious, though, to follow up on
0050
1 Nicol and Andrew's point about more rapid sampling
2 methods -- taking advantage of the fact that we do
3 have the worldwide Web in more and more places and
4 more and more people on it -- I am curious whether we
5 have any experience with how to do that well, other
6 models that we can pick up from anywhere. Obviously
7 the commercial side of the Web does a fair amount of
8 this to different degrees of success. Do we have
9 examples of this? Has the research community started
10 to pick up on this kind of thing? Please.
11 MS. TURNER-LEE: I --
12 MS. CRAWFORD: I want to add a friendly
13 amendment. In addition to friendly ways of gathering
14 things on line, I am concerned about wire lining
15 phones and what that means and how we are thinking
16 about changing that?
17 MS. TURNER-LEE: I want to respond to that,
18 because we were at the joint center -- all of us look
19 at the Pew data as a source of information because
20 it's timely, and it came out when John was there
21 pretty rapidly with regards to particular areas. And
22 one of the things at the joint center that we started
0051
1 to work on that is going to allow for rapid turnover,
2 for example, is on African American broadband
3 adoption and use data. I think it's one of those
4 things where the quality is a concern, but one of the
5 things we talked about and we just talked about it
6 with the FCC, are there are ways to continue to build
7 off of that group. Are there mechanisms that you can
8 put in place that get at the consumer side of
9 adoption, obviously more longitudinal studies where
10 you lock in the data group where you continuously go
11 back to see to see what their experience has been
12 with broadband.
13 The hardest thing I think for many of us
14 at the table to measure is that sense of improvement
15 around quality of life. You know you can see more
16 jobs or you can look at the metrics and see the
17 creation of jobs. You can see improvements in
18 educational disparities, but can you see the
19 improvement in quality of life. That's where quicker
20 consumer based studies, working with nongovernmental
21 actors to determine their ways if their methods focus
22 on niche groups or certain areas might be helpful and
0052
1 creative and innovative to, I think, bring the two
2 paradigms together in terms of how we do research.
3 MR. WEITZNER: Go ahead, Eszter. I will go
4 to the second question.
5 MS. HARGITTAI: I thought I would address
6 the methodological question. There are two issues.
7 How do we get the data that we gather out of
8 traditional methods more quickly, and the other is
9 can we use new ways to gather data.
10 With respect to the latter, there is
11 methodological data on this. It depends on the
12 questions you are asking, and the methodological
13 issue on this doesn't attempt to ask questions about
14 Internet use.
15 I think when you were studying Internet
16 use by gathering data on the Internet, that may be
17 problematic because the selection bias that you have
18 as to who is actually responding to that survey is
19 too directly connected to the dependent variables of
20 interest. I think in that case it may be more
21 problematic, and I don't think the methodological
22 literature has actually focused on those questions.
0053
1 I don't know if we can draw on that in this context.
2 MR. WEITZNER: I asked Ken to just start
3 off the conversation about current sources of data
4 that are available more broadly in the Internet
5 arena, what we should learn from that, and again how
6 we might think of supplementing that or looking at
7 those data sources, moving them in some direction or
8 another to answer the kinds of economic and social
9 questions that we are trying to get guidance on.
10 PROFESSOR FLAMM: Okay. Well, thank you
11 very much, Danny.
12 I would first like to start out by
13 endorsing everything Shane said. But I also want to
14 add on to that the issue of, so Shane kind of assumed
15 we are going to to be having price data to tack on to
16 other things. As an economist looking at those
17 issues it's hard to, in addition to not knowing
18 something about socioeconomics and demographics of
19 households, you may also want to know about some
20 prices they are facing because that is a determiner
21 of economic decisions.
22 In general, data sets are very bad for
0054
1 looking -- the data is terrible, to put it bluntly.
2 If you -- to think about the questions that Susan
3 Crawford was asking before, sort of, you want to know
4 what the impact of your policy is going to be in
5 designing that policy, and if you are considering
6 various subsidies, for example, you want to know how
7 price sensitive those things are. We just don't
8 know. The studies on sensitivity of broadband to
9 price are all over the map. Any result you want.
10 It's elastic. It is inelastic. I can find you a
11 study that will give you that result. The reason for
12 that is because the data are just really bad.
13 Now, coming back to Shane's point on the
14 CPS. The CPS was going down the road of becoming a
15 really good data source around 2001, and then it got
16 chopped off at the knees. It got gutted, or
17 degutted, whatever colorful adjective I can come up
18 with, and it pretty much became useless in terms of
19 looking at these issues. There is an opportunity to
20 make it useful again. This is an opportunity that we
21 should seize.
22 And in terms of the types of data that we
0055
1 were talking about from the BTOP and new programs in
2 organizing, and think about what you were getting out
3 of those programs, the one thought is to try to
4 coordinate it with the CPS to report at levels where
5 it could be combined with CPS data and made into
6 something more useful for that reason.
7 I'm going to shift the focus and talk very
8 briefly about other sources of data. If we are
9 talking about price data, which is sort of the great
10 unknown out there, since no one has really addressed
11 how you are going to do this exactly.
12 There is public government data that is
13 collected by the Bureau of Labor Statistics in the
14 form of the consumer price index and the producer
15 price index that, for various historical and sort of
16 obscure reasons which I'm not entirely familiar with,
17 is reported in pretty much a useless form at the
18 moment.
19 So, for example, the CPI is reported as a
20 public price index on Internet services and
21 electronic information providers, which of course
22 tells you nothing useful about different -- anything
0056
1 you would be interested in. There is no reason, it
2 seems to me -- this is reported
at all U.S. cities
3 national level. There is no reason you couldn't
4 break it out into subaggregates. You may have to
5 increase your sample size perhaps, but nonetheless,
6 that's what you have to do.
7 It seems to me we are putting more effort
8 in measuring cheese prices than we are sort of
9 broadband access prices when you consider what the
10 typical household budget goes for.
11 Another point on that is that this is a
12 fixed price -- a fixed weight price index, and the
13 current weights are December 1997, which is a long
14 time ago. It's a decade ago. That is when they are
15 producing this index. They were using weights that
16 were fixed weights corresponding to December 1997.
17 It is way out of date. I don't need to tell you
18 this. There are well-known methods for indexing
19 price weights and they should be employed.
20 That is a long discussion. There are
21 consumer price index services that are produced by
22 the Bureau of Labor Statistics. They basically
0057
1 report. It's hard to -- I would be hard pressed to
2 come up with a more -- a less useful way of
3 aggregating my price index than this one.
4 They have one price index for dial-up and
5 asymmetric DSL index. All a single index. Okay.
6 One for leased line and symmetric DSL Internet index,
7 again, all lumped together, a single price index, and
8 a third for other Internet access and related
9 services, which presumably includes things like cable
10 modem, the second most common form of Internet
11 access, along with God knows what else.
12 There is no reason in principle why you
13 can't unsquish these things and make them more
14 detailed indexes of things that people are interested
15 in, asymmetric DSL versus cable modem access to start
16 with. This is sort of a -- the same issue of sample
17 weights applies here. This is 2004, but this is
18 changing rapidly. You want to change your weights on
19 this fairly rapidly as well.
20 As far as other public sources of data, I
21 think -- let me turn briefly to the Pew data set,
22 which is a remarkable sort of sociological story here
0058
1 where John constructed, through perseverance and
2 grit, a sample that is typically 2500 households, a
3 dinky, tiny little ant of a sample, okay, but
4 nonetheless this is the current basis for most sort
5 of policy making going on right now, this very tiny
6 sample.
7 The CPS was on the path to becoming a much
8 larger, much more scientific, much better -- not
9 necessarily much better, because it didn't have John
10 Horrigan shoving it in the ribs with a stick -- but
11 was on the path to becoming something like what the
12 Pew sample is, but on a larger industrial scale. And
13 this is the opportunity to get it back on track, to
14 make -- you know, to take the weight off of Pew, make
15 it a big sample, cover the right kind of stuff.
16 Let's do it already. There is no reason not to.
17 OCED collects price, and everybody likes
18 to dump on the OCED the Internet comparisons. I
19 won't go down that road. But the bottom line is,
20 like it or not like it, they are at least making some
21 kind of an effort to collect price data and also
22 penetration data.
0059
1 The FCC form 477 is the other major sort of
2 government data collection instrument. It is being
3 made much better, is my understanding, although I am
4 sort of not directly involved in it. It's going to
5 be available at a much more granular level. But a
6 caveat, the granular level is sort of the half zip
7 code is kind of where it is going in terms of
8 household numbers, which is still pretty up there in
9 terms of -- and the issue now is that if you look at
10 the broadband mapping data which is being collected
11 down to the street address level perhaps, that's sort
12 of a long time, one-off shot, you are going to get
13 something for some period, but after that, we are
14 back to form 477, which is considerably more
15 aggregated than you are talking about with the
16 broadband map. It's not clear where all of that is
17 going.
18 One thing to think about is, there are ways
19 to make form 477 more interesting and more useful.
20 This private proprietary information data constantly
21 gets raised.
22 I would like to point out that the IRS, in
0060
1 an effort to promote -- the Internal Revenue Service,
2 in an effort to promote electronic filing, releases
3 zip code level data on a number of electronic filing
4 companies who are offering services in every single
5 zip code, the number of returns that are filed in
6 every zip code, and as far as I know, no proprietary
7 issues have ever been raised to obstruct the IRS from
8 releasing that data.
9 Why is it considered so proprietary or
10 business confidential or sensitive to release data on
11 the number of broadband providers in a zip code or a
12 census block? Why is it not sensitive for IRS
13 business, but is sensitive for broadband providers?
14 I don't understand the answer to that question, as it
15 has been implemented in current practice. I would
16 like someone to ask it. No, I'm not sure that is
17 correct.
18 Finally, I will want to talk about two more
19 sets of data that are potentially useful. There are
20 private providers of information, things like
21 comScore data, which is essentially a type of data.
22 If comScore can do it, the U.S. statistical apparatus
0061
1 can do it, too. There is no reason that we can't use
2 some of the methods of private survey firms like
3 comScore uses.
4 Finally, there is interest expressed in
5 sort of how broadband use is evolving, how higher
6 speeds become available, what types of new
7 applications, how things are changing. Even so, that
8 one of the issues was that was raised earlier, the
9 policy issues was a tradeoff between investment and
10 regulation. Well, so nobody has a very good idea on
11 how much is actually being invested in the broadband
12 infrastructure, by the way.
13 So the most recent -- was it a 977 report?
14 I get confused on my FCC numbers. From the FCC last
15 year, notes passing about all the information
16 submitted about broadband investment in Internet
17 infrastructure. But if you look at the report, there
18 is no data anywhere in the report on this subject.
19 It would be -- if you are asking the question, what
20 is the interaction between investment and regulation,
21 you have to know something about investment, and
22 there is no data public, or, as far as I know,
0062
1 proprietary data that is useful that is available to
2 examine this question.
3 Finally, broadband providers actually have
4 a great deal of information about traffic
5 characteristics in their networks.
6 A good example of this is, if you go back
7 and look at FCC hearings from the early 2000s, 2003,
8 2004, you will see cable labs, individuals from cable
9 labs who were reporting data on traffic symmetry, for
10 example, that is, what the upstream and downstream
11 loads were for different types of -- were on cable
12 networks.
13 And actually, most broadband providers now
14 have -- make use of apparatus, things like sand vine
15 boxes, which actually sniff out packets, and they
16 have a pretty good idea of different peer-to-peer
17 networks, and they are constantly in real time
18 sniffing and sampling, and in some cases, using it to
19 shape traffic.
20 But there is technology that would allow
21 real time monitoring of traffic at different times of
22 day through different types of customers that could
0063
1 be linked to customer demographics, and ultimately
2 companies actually do have this data in principle.
3 One could imagine some sort of public sniffing post,
4 if that's the right way to put it, which would create
5 an appropriate data set that can be used to examine
6 these questions when, sort of, bandwidth becomes
7 available at lower prices. How is it used? How do
8 patterns of use change?
9 So, in principle, taking the model of
10 cable labs' networks or traffic monitoring projects
11 as a base, and think about how you can create a
12 public version of that that can be used to debate
13 openly and publicly some of the questions that are
14 being asked about the future of broadband in the net
15 and what kinds of traffic is going to be going on
16 that, That might be a useful thing to consider.
17 MR. WEITZNER: Ken, thanks. Comments?
18 We won't get into the privacy issues that you raised.
19 John.
20 MR. HORRIGAN: One observation, at the FCC,
21 pursuant to the Broadband Improvement Act, FCC is
22 charged with doing periodic surveys of people on
0064
1 their Internet use, why they don't subscribe
2 including what they pay per month. That survey
3 currently is in the field, what the periodic will
4 mean over time, I'm not sure is entirely clear. But
5 Ken just touched on the challenges of getting the
6 price data correct. How you ask those questions.
7 In the broadband plan we are looking at
8 that will collect that information, that is in the
9 survey that is in the field currently. We are
10 getting that from other private data sources that we
11 are procuring, but asking that question, getting at
12 that in the right way, especially with services being
13 bundled is, A, very difficult, and then, B, as Ken
14 was suggesting, trying to institutionalize that over
15 time across agencies is doubly difficult.
16 So I think it's something that you have
17 to, we have to watch closely over time. I think it
18 will require a lot of coordination across agencies.
19 But, you know, Ken, Shane and others have been
20 looking at this question. Myself, I have been
21 looking at this question for a while now. So keeping
22 a dialogue open with the people who have been
0065
1 experiencing and living the pitfalls is very
2 important. But it's important also for all of us to
3 work to institutionalize this in various data
4 collection efforts throughout the government.
5 PROFESSOR GREENSTEIN: I agree with
6 everything Ken said, too. We have a mutual
7 admiration society. There is a model of how, on the
8 business side to do this. The European. If you look
9 at what Tony Clayton, who is the head of statistics
10 at the United Kingdom, organized around Europe, they
11 have a perfectly good model of how to take data from
12 the broadband side, aggregate it to the industry use
13 side, and output it in such a way as to be very
14 useful for at least the economic side of this, which
15 is understanding its contribution to productivity and
16 economic growth. They have done the hard work. So
17 that model is out there. You can just borrow it. It
18 would involve other parts of the Department of
19 Commerce. But that's the -- that's for you to figure
20 out. But there is a model how to do this,
21 Obviously, on the business side.
22 PROFESSOR FLAMM: So one other quick
0066
1 comment. It is true that the broadband is being
2 bundled with various other services and it makes it
3 more complicated to measure prices but anything but
4 impossible. So the bottom line is that there is
5 well-developed models of complex products that
6 are -- there are bundles of different components with
7 different characteristics.
8 That theory has been well developed and is
9 employed widely within the government to measure
10 things like prices for computers and prices for
11 various kinds of services, so there is no reason, in
12 principle, why you couldn't use the same kind of
13 methods, including hedonic pricing models, to measure
14 broadband prices in precisely the way it's measured.
15 MALE SPEAKER: Ken, I thought "hedonic"
16 meant "happy."
17 PROFESSOR GREENSTEIN: Ken is happy.
18 MS. CRAWFORD: Thank you, Chris.
19 PROFESSOR FLAMM: It does make me happy,
20 actually. I am using it in a technical sense.
21 So, hedonic pricing indexes are widely
22 used in governments throughout the world for
0067
1 measuring price indexes for products which are
2 differentiated with multiple characteristics. And
3 basically it's a methodology for taking into account
4 the various characteristics of the product in coming
5 up with the price index which accurately measures its
6 movement over time.
7 MS. GORDO: I would like to say --
8 MR. WEITZNER: Let me remind everyone,
9 first of all speakers, please say your name. Second
10 of all, for those of you who are watching the
11 Webcast, you are -- we encourage you, if you have
12 questions, to send questions in by e-mail, and will
13 someone tell me the e-mail address? It's on our Web
14 site, I know. We will find the address. But if you
15 can look on the home page for this workshop, you will
16 find an e-mail address. We will get the e-mail, and
17 I will make sure the question is posed. Blanca?
18 MS. GORDO: I would like to say, yeah, we
19 need an affordability index. The question is, why
20 are people not adopting broadband? There is a -- a
21 price index would be useful to understand consumer
22 choices. We are really at the early stages at
0068
1 creating benchmarks for that or indicators who can't
2 tell you this is how much you are paying to be on
3 line, because there is different factors that it
4 depends on, like the function, the capacity, the
5 application, the power. There is all sorts of types
6 of digital power that you have that dictate how much
7 you had to pay.
8 Thinking about how we can develop an
9 affordable index and price index would be helpful to
10 help us make those choices and really understand what
11 are the barriers, how does that function, and what
12 kind of mechanisms can we institute so it's
13 affordable for everyone. Thanks.
14 MR. WEITZNER: Anne, please.
15 MS. NEVILLE: Anne Neville from NTIA.
16 On the timeliness of data, just a
17 clarification. The grants that are going out now to
18 states, that is actually for two years of broadband
19 mapping data that will be updated every six months.
20 And NTIA has an ongoing responsibility to maintain
21 the National Broadband Map, So it won't be just a
22 one-time shot in terms of data collection.
0069
1 PROFESSOR FLAMM: You have a budget to
2 maintain this? I'm just curious. Sorry.
3 $100 million.
4 MR. WEITZNER: So the e-mail address for
5 those of you coming in through the Webcast is
6 bbdata@ntia.gov.
7 We have a question. I think we have a
8 question in the back, please. Could you identify
9 yourself?
10 MR. BEST: My name is Tom Best.
11 MR. WEITZNER: Why don't you come up to
12 the table.
13 MR. BEST: My name is Tom Best. I'm an
14 independent consultant.
15 Just a reaction to the suggestion of
16 deploying things that would measure traffic widely
17 across the Internet.
18 I think you would run into a great deal of
19 operational, as well as privacy difficulties, with
20 that, but I think there might be a simple way to do
21 something quite similar which would address both the
22 operational challenges as well as the privacy
0070
1 concerns, and that would be to develop something like
2 a setting at home; a simple application that people
3 could voluntarily download which would occasionally
4 subject them to passive monitoring; in other words,
5 checking what their actual upload and download speeds
6 would be like, checking their patent loss and
7 technical features, and also occasionally prompt them
8 to respond to some direct questions. Specifically,
9 also to ask them to identify what census area they
10 live within. So then you could then use that data to
11 associate with all the other metrics you would like
12 to use from other sources. Just a thought.
13 MR. WEITZNERr: Thanks. I think we are
14 going to move on to the third category of questions.
15 I have asked both Blanca Gordo and Nicol
16 Turner-Lee to give us information on economic and
17 policy issues that we ought to be thinking about and
18 ways to address those questions.
19 Blanca, go ahead.
20 MS. GORDO: It's better for me because I
21 can see everybody. It's a beautiful day to be asking
22 those questions to a researcher. What kind of data
0071
1 do you need? That's beautiful. I am happy to be
2 here and talk to you about my view on the
3 constitutional problems of broadband, and why this
4 research, as it relates to development in the poor,
5 is well designed for coordinated policy that gets at
6 new forms of inequality that are rife with the
7 technology in the production of society institutions
8 like you have heard.
9 We need data to inform the design of a long
10 lasting technology development plan. And by
11 development I mean education, jobs and wealth
12 generation. Scientific broadband research, I
13 believe, is key to examining contemporary development
14 of problems. To see the importance of why we need
15 to -- the importance of this type of research is, we
16 need to view, how to see the importance is viewing
17 technology as more than just a tool. It's a process
18 of production, consumption and exchange. We need to
19 pay attention to what these are to network into
20 society's form of production. That's what is new.
21 Also what is new is a new set of rules
22 that determine how we can get to determine and govern
0072
1 how society uses technology and broadband to
2 participate, compete and prosper.
3 This is important as you think about
4 conceptual data indicators for data collection,
5 Precisely because private and public institutions
6 have integrated this all-purpose technology into work
7 processes, into governing structures and service
8 delivery. And because they are replacing paper based
9 and face-to-face forms of service delivery, we are
10 replacing face-to-face operation with online forms of
11 transaction. Without broadband, many low-income
12 people and places will not be able to connect to or
13 reproduce their own means of production, consumption
14 and exchange online.
15 This can only result in what I call
16 digital destitution. This process of alienation and
17 deprivation from productive members of society will
18 cost institutions inefficiencies in productivity and
19 cost savings and reproduce atrocious results. It
20 comes down to recreating unprepared labor markets,
21 dependent and unproductive populations, and fueling
22 the types of informal and criminal economy. Though I
0073
1 do have some indicators for benefit, I can't tell you
2 how many minorities are actually generating a return
3 on their investment from broadband use with the same
4 level of confidence that I can tell you for how many
5 are actually excluded from this process.
6 Since Internet access is a necessary
7 component to participating in this technology
8 process, then, we know -- according to the most
9 reliable data set available, based on the census
10 abstract that Jim just spoke of, the community
11 population survey, we know that 60 percent of blacks
12 and 68 percent of Latinos are excluded from
13 institutional processes based on technology. Native
14 Americans are not accounted for. This compares to 68
15 percent of whites and Asians who use the network.
16 This is only complicated by historic
17 disadvantages, and we need to contextualize that
18 problem in that way. The mechanism of advantage and
19 exclusion evolves new sets of technology and human
20 skills, and we also need to account for that. The
21 changes in the rules of operation and even changing
22 institutional uses of this type of technology. Lack
0074
1 of institutional planning and preparation of
2 proposition to proliferate the productive form of
3 technology. The mechanism of advantage and exclusion
4 evolves under mechanisms of poverty that are
5 underway. But we know less, and Eszter pointed out,
6 how to start with this research with what are the
7 questions, we need to know to structure our data
8 collection that way.
9 One of my favorite questions that remains
10 to be answered and we would love data on is not
11 whether, but under what conditions through which
12 social processes and government structures can all
13 American society, including the poor and the
14 disabled, benefit from the use of digital network
15 technology.
16 And we do need, I believe, qualitative
17 broadband research and a comprehensive data set that
18 are needed for institutions to figure out with
19 certainty how to break technology market bottlenecks
20 for development. How places can obtain the
21 technology mechanism and how institutions can prepare
22 populations to develop the necessary skills for
0075
1 employment, for furthering their education and
2 ultimately generating wealth.
3 A statistically significant sample
4 representing the low income, the disabled, the
5 African Americans, Latinos, Native Americans in low
6 income places is lacking, making it difficult for
7 local planners to assess the problems and how they
8 can address them.
9 I think the NTIA would definitely serve
10 science and society by designing a culturally
11 sensitive survey instrument that gets at technology
12 integration. What are we measuring and when and how,
13 even to get at the issues that we were talking about
14 measuring the price index. The social institutional
15 processes and rules that gets to determine how you
16 use technology.
17 A case in point. In Los Angeles, I was
18 doing a study in southeast Los Angeles, every school
19 has a computer, but they are not connected to the
20 Internet. They have a computer lab in one of the
21 most over crowded high schools, and it's based on
22 quota in terms of time and accessibility. There are
0076
1 filters that prevent even very legitimate information
2 from being uploaded, and that's a decentralized form
3 of mechanism. If you bring your class for one hour
4 quota per week per class, in your one hour, like
5 those teachers said, I want to teach my students how
6 mythology and language is interconnected, and all
7 this information comes up, but I can't open it. By
8 the time you call the district, they are done with
9 the hour.
10 Rules are important. Also the types of
11 technology ownership and appropriation become key.
12 The utility and standard productive functions of
13 technology by price is important, so are cultural and
14 contextual differences with measurable numbers.
15 Those are needed for policy benchmarks. This type of
16 information would help identify solutions.
17 I want to leave you with the idea that the
18 stakes are high. Without public access to technology
19 and training, to manipulate the productive function
20 of technology, populations will definitely be
21 unfavorably positioned to participate in online
22 production consumption and exchange. They will be
0077
1 systemically limited from being in the public sphere,
2 and via these needs of telecommunications, services
3 economy, networked society, these things matter.
4 They will lack the materials that are extended online
5 by institutions, and they will be unable to
6 communicate and exchange necessary information with
7 institutions. The least bad outcome I think is that
8 generations of unprepared children, minority children
9 and their children will lose the possibility for
10 mobility.
11 And all of this -- I like for this -- I
12 want to leave you with this idea, that the lack of
13 broadband can lead to a three tier public service
14 system. That would be one -- the high quality and
15 efficient information exchange of necessary materials
16 that reduce production and transaction costs for
17 those with anytime, anywhere access, a limited low
18 rate public service, and basically no service at all
19 for some Americans. And this system could prove to
20 be unconstitutional for failing to meet requirements
21 for universal access to public service precisely
22 because institutions have integrated this into the
0078
1 formal aspects of our daily living.
2 We at UC Berkeley would be happy to create
3 a broadband technology research design and survey
4 instrument for development. At the Center for Latino
5 Policy Research we are working on measures for the
6 problem identifying potential policy solutions and
7 benchmarks for program evaluation. Thank you for
8 your attention. It's an honor.
9 MR. WEITZNER: Thank you, Blanca, very
10 much.
11 Nicol, would you like to add something?
12 Let's try to focus on data collection questions that
13 you want to throw out to us. I want to get through
14 the other, this topic and the other two as well.
15 Thanks.
16 MS. TURNER-LEE: I actually want to, before
17 I go into the data questions, as a sociologist, drill
18 down into why this matters, because I think it's very
19 significant, as we are framing the research, that we
20 are interested in, much like Blanca talked about, we
21 have to be clear about what the research question is.
22 The Commerce Department has provided
0079
1 incredible data sets, but part of what has maybe
2 driven all of us into this frenzy about understanding
3 the impact of broadband is the analyses behind that
4 will drive a certain answer to a particular question.
5 As I was just preparing a few remarks about
6 this, I thought about the out of government and
7 freestanding policy institutes as well as
8 universities that we started with a digital divide.
9 And we began to talk about digital inclusion and now
10 we were at a point where the FCC has a plan about
11 digital exclusion. What that means. That speaks for
12 the fact that we have to mitigate barriers and we
13 have to collect data that helps us get through the
14 barriers of relevance, costs and availability,
15 particularly for communities that are on the fringe
16 of participation access.
17 As a leader at the Joint Center for
18 Political and Economic Studies, the primary and
19 leading policy institute on issues related to African
20 Americans and other people of color, it's imperative
21 that we get all people engaged in the information
22 economy. This is really about -- if we are to look
0080
1 at the broad question that Susan put out on the table
2 earlier, this is about getting people included in the
3 information ecosystem and getting people involved in
4 the economic democracy that we are quickly tipping
5 towards, that with excluded resources you do not have
6 a voice, essentially, and it has impacts on your
7 ability to be prosperous, healthier, as well as more
8 efficient in how you do things on the energy level.
9 So when we talk about the type of data
10 that we need, I am going to be a little selfish,
11 because as a sociologist, there is this lack of
12 qualitative information.
13 I think it's clear that we can look at the
14 information that is generated today through the CPS
15 study and others, but it's that objective data that
16 tells us the relational and causal factors as to how
17 many computers you have, how many people in your
18 household, and it allows us to make assumptions about
19 broadband use.
20 If we start with a theoretical question or
21 a frame that broadband is a multiplier, then I think
22 we look at different ways to collect data and
0081
1 different ways to analyze that. And the multiplier
2 effect for now, something we could use immediately in
3 data collection is matching broadband mapping with
4 the state of broadband adoption activities and to see
5 if there are correlations between where broadband
6 service is and is not, and what activities and
7 programs are overlaying to move the meter on access.
8 So I think that is something I can actually do now as
9 we move through these investments through commerce.
10 I think another area that is part of this
11 discussion that helps inform policy and research is
12 we have also got to get at --
13 MR. WEITZNER: Could I ask you a question
14 about that before you get on to the next?
15 One of the questions that we have an
16 immediate concern about, as we go forward with this
17 first round of awards, is to really understand the
18 performance of these particular programs. We talked
19 a lot about national level statistics. What are we
20 able to learn, not just for minority populations, but
21 for all populations at a local level, at a micro
22 level, about the impact of the money that we put out.
0082
1 But it strikes me that it may get to some of the
2 qualitative questions that you are asking.
3 We will have a connection to, you know, a
4 very human scale activity with many of the programs.
5 What should we be asking in those programs, aside
6 from just numbers about job creation and all the
7 things we have talked about? Are there qualitative
8 questions that we could be asking that would help?
9 MS. TURNER-LEE: I was actually going to
10 put that as my point three, because as researchers,
11 we look at it as qualitative research and observable
12 analysis, but we have to gather information without
13 making judgment, because we don't know enough about
14 good or bad programs, but have a repository where we
15 collect information and allow the research community
16 to come up with factors that come up with specific
17 trends of success.
18 The prior work I did in the nonprofit
19 community, starting with CTCs with broadband adoption
20 efforts through my work at One Economy, there are
21 particular trends we see and what drives demand for
22 broadband. And I think if we could figure out -- and
0083
1 I don't know the answer, maybe my colleagues do,
2 because qualitative research is always somewhat
3 ambiguous and always subject to the interpretation of
4 the researcher -- but if there is a way to have maybe
5 separate, as to not compromise the scientific value
6 of the data that is being placed out there, to have a
7 repository for the collection and aggregation of
8 these stories. That may be a start to really list
9 what these programs look like.
10 Because right now we see on the Web site
11 the executive summary, for example. And many of us
12 have spent time in our prior lives analyzing
13 executive summaries just to get at the nuts and bolts
14 to what a program looks like and making comparisons.
15 But if there is a way, using the Internet, to create
16 this exchange of knowledge. I think many of the
17 nongovernmental actors will rise to the occasion to
18 help identify trends.
19 I was also going to say, in a similar note,
20 in the interest of time, there are two other factors
21 that I think we should pay attention to in research.
22 One are quality of life factors, what Richard talked
0084
1 about with workforce development. If this is going
2 to be right, it has to impact workforce development;
3 health, energy, entrepreneurship and small business
4 development, et cetera. So we also have to figure
5 out how to make those data sets available and
6 benchmarked, or at least somewhat aligned and timely
7 as the data that you are producing, or figure out
8 ways to incorporate those types of questions in the
9 work that you are doing. Because I think before we
10 knew little about, you know, certain factors.
11 If the focus is on digital inclusion --
12 exclusion, excuse me -- that we need to get out how
13 are people using broadband, and I think it's an
14 appropriate thing to do now to relook at those
15 questions, to find jobs. How are they using
16 broadband to conserve energy? How are they using
17 broadband to identify healthcare?
18 I think that was the beauty of Pew, because
19 it gave us insight into an area that we couldn't get,
20 into the living rooms of families. And although it
21 was a smaller sample, I think there's an opportunity
22 for entities to, government entities to actually
0085
1 start collecting that or to work across interagencies
2 to figure out what their data sets look like.
3 I think the research community, we need to
4 work multidisciplinary. On the government side, you
5 need to work interagency-wide if data is going to be
6 important to drive the ultimate effect. And on a
7 cultural aspect, the comment about the disabled
8 community, the comment about Latino Americans and
9 African Americans, it's because, Blanco said, 65
10 percent of the African Americans are excluded. But
11 what is even more surprising about African American
12 participation in broadband is, it has been
13 insignificant in growth for two years, and for
14 Latinos, and much we know about are for English
15 speaking Latinos. California may be different. If
16 there is some way to figure out how to get at more of
17 the demographics.
18 I was at the FCC earlier where we talked
19 about going deeper on subgroup data. There is an
20 opportunity to look at the data that we are
21 collecting and going deeper by not just if you are
22 African or Latino, I would push a conversation. A
0086
1 question that came to me today whether you are of
2 American descent, if you are a recent immigrant to
3 this country, what does broadband represent to them?
4 It's tough for me, in my sociological hat, to say you
5 all should do this, because that is on the burden of
6 research to push the analysis as we all continue to
7 care about this issue.
8 If we work together to find more flexible
9 questions, multiple methodologies, longitudinal work
10 in data sets that allow us to look out into the
11 future, I think we will be coming back into this room
12 with a different story that will look at whether it's
13 a pricing index and quality index or service index
14 where we could say we made progress with investments.
15 I put that out there for consideration, and I think
16 all of us are willing to continue that conversation,
17 but it's not an easy one to answer.
18 MR. WEITZNER: Thanks very much. I am
19 going to move us along. Any final quick comments?
20 Very quick.
21 MS. GORDO: I think you have to keep in
22 mind --
0087
1 MR. WEITZNER: Sorry.
2 MS. GORDO: I think it depends on how these
3 programs are structured. This is a time where there
4 is a pilot experimentation. What is important, when
5 you were evaluating this, is to take the preexisting
6 conditions into account. This kind of effort is
7 within a particular economy; political governance,
8 social dynamics, institutional operations and where
9 they are at. So I think it's important for us to
10 begin -- and I think there were missed opportunities
11 in the past. There is not a model, and there is
12 still no model. But what we can uncover are the
13 types of conditions and social processes that are
14 underway to isolate what are the benefits that this
15 technology is serving certain populations, and they
16 may be different. So we need to pay attention to the
17 inputs, the structure of the programs. And there is
18 some indicators that I do have and can share with
19 you.
20 MS. TURNER-LEE: One quick comment. We
21 need to go ahead cautiously so we keep the data also
22 open enough to come up with different types of
0088
1 discovery. So I put that out there, as we are
2 exploring this, so it allows for the discovery that
3 we are also looking for in the research world.
4 MR. WEITZNER: Okay. We are going to spend
5 the next seven and a half minutes on emerging trends
6 in research and ask Eszter Hargittai to kick us off
7 there. And we will go on to data format questions.
8 So please.
9 MS. HARGITTAI: So I am going to lead us
10 back to the discussion of the CPS, because, like
11 Shane and Ken, I think the CPS was an amazing
12 resource, and then it sort of lost, it didn't, over
13 the years, and there is hope that we can maneuver
14 back. So I have some concrete suggestions on how we
15 can do that. Not simply go back to the '01-'03
16 version, but actually move past that to make it
17 better.
18 Just a point. I thought Jim did a good job
19 of explaining why the CPS is so amazing. One more
20 point is just to address the comments by Richard,
21 actually is, because of these different supplements
22 that get administered at the same time, we are able
0089
1 to join the data. So, for example, I think it was
2 '01 or '03 when they had a separate disabilities
3 section, so I was able to write a paper specifically
4 looking at people with disabilities and their use of
5 ITTs, because you have no other data set where you
6 have thousands of such people. So that's another
7 reason why the potential of the CPS is so huge.
8 What I'm going to advocate here is in
9 terms of emergent trends. I think these have been
10 emerging for about the last decade, is that we really
11 need to move past focusing on simply access data and
12 simply infrastructural data.
13 So, for example, the -- what I got from
14 the slides we saw about BTOP, is that most of the
15 data there is infrastructural, and I think there is
16 serious limits to what we can do with that. And we
17 saw that both the '07 and '09 super tiny CPS
18 supplements have nothing but access data. So there
19 are serious limits there.
20 Where should we go? First we need to go
21 back to the '01-'03, because those are questions
22 about useful questions that Shane referred to. We
0090
1 definitely need to get questions back on there about
2 types of uses. However, that is not enough, either,
3 because actually we can sometimes do more harm if we
4 don't have all the relevant questions.
5 So, for example, one of the things that we
6 sometimes see in studies is that if we have questions
7 about types of uses, and then all we see is that
8 people are only using the Internet for certain types
9 of activities from which they might not benefit, then
10 some people walk away with the conclusion, well, why
11 should we support this, because it won't help our
12 workforce or economy or what not.
13 This assumes -- since we are lacking data,
14 it often assumes that it's actually people's
15 preferences that drive their uses, and we don't have
16 data to look at what else might be driving the
17 diversity of their uses.
18 One of the things I focused on in my work
19 over the past decade is skill and ability to use
20 technologies. That is one of the factors that we
21 also need to collect data on. And it's much related
22 to questions of inequality and inclusion because
0091
1 skill is predicted by various socioeconomic factors.
2 But we don't know this for a large national sample
3 because we simply don't have data on this for a
4 national sample.
5 In addition to skill, something else
6 specifically that we need to collect data on is
7 context of uses. So, for example, people's support
8 systems. I am not suggesting these are easy things
9 to measure, but I think to make the CPS all that it
10 can be, this is something that we should think about
11 including on there.
12 One of the ways to address skill is
13 precisely through education, training, what else it
14 supports, and it may be informal support through
15 people's personal networks is what matters.
16 By doing this, that would then -- okay, so
17 one of the reasons to focus on skill is it's one of
18 the things we can actually intervene on through
19 policy. We could help them move more diverse types
20 of uses out to a larger part of the population.
21 Just a couple of logistical notes. One of
22 the features of the CPS that I actually don't think
0092
1 works for the Internet use supplement is, in addition
2 to asking the respondent about his or hers' own
3 Internet uses, that person acts as a proxy to respond
4 about other household members' Internet uses.
5 Now, I think it's reasonable that maybe
6 your spouse knows your employment status. I don't
7 think it's reasonable that your spouse knows what you
8 do online. I think that's a methodological issue.
9 (Laughter.)
10 MR. WEITZNER: You hope, right?
11 MS. HARGITTAI: I think that's
12 definitely a methodological challenge with how the
13 CPS is set up, and something we need to think about
14 if we are to apply it to Internet measures.
15 Finally, if we are able to answer some of
16 the big and important questions that Susan asked, we
17 need longitudinal data. We have to stick with the
18 same people over the same time. We just can't get
19 snapshots.
20 MR. WEITZNER: Eszter, thanks.
21 Let me take a couple of additional points
22 on this topic, Richard, and we will go on to the last
0093
1 question.
2 MR. HOME: I am from the Department of
3 Labor and a defender of the disability population.
4 From my experience, we started to put
5 disability questions into the current population
6 survey back in 1999. Think about that. We
7 encountered all kinds of issues from the -- oh, we
8 can't define it -- the population with disabilities
9 to issues on severity to issues of self-report. And
10 it gets worse when it is reported on head of
11 household. Okay. All of that aside.
12 As the advocate for getting these questions
13 in there, because I believe the economic situation,
14 as described by the Department of Labor, should
15 include a description of people with disabilities in
16 this country. Go figure. We went through a lot of
17 issues around threats to response rates, over
18 sampling with people with disabilities, under
19 sampling certain finite groups, and it went on and on
20 and on. And in the end, we only got a limited number
21 of questions that we were able to ask, as is the case
22 with the CPS.
0094
1 But I just did want to mention that we are
2 designing a CPS disability supplement, based on these
3 new questions, that we hope to do in 2012, which is
4 the supplements that they filled out over the years.
5 So I would be very interested in your input into some
6 of these research areas that we might be able to
7 include on the disability supplement.
8 MS. HARGITTAI: I will say, having actually
9 looked at that data, it was limited data and it was
10 very hard, and we got to a certain point, but then we
11 were unable to continue. It was limited.
12 As a graduate student in the late '90s, I
13 tried to figure out how the Internet use supplement
14 happens, and I am satisfied -- with this transparency
15 workshop, I think I can figure out how this actually
16 happens. So we have been working on it for quite a
17 while.
18 MR. WEITZNER: We have been impacted in
19 time, but, George, we were interested in your
20 thoughts about the issue about data formats that will
21 work best for the research community, and
22 particularly how we take advantage of the fact that
0095
1 we have the web to get data out there more
2 transparently.
3 MR. FORD: I will be quick by necessity and
4 move on to that issue and move on for more general
5 comments on what has been said.
6 First, I think the data should be provided
7 in some kind of format that is easily imported into
8 packages or mapping packages or whatever you are
9 using. The Census Bureau is quite good at that. A
10 file or tire maps or whatever it is. That is not a
11 difficult issue. You can't run a regression on a
12 map. So there has to be some connection between the
13 data that allows demographic information to be linked
14 to whatever it is that we are collecting. Census
15 block, please not zip codes, because that doesn't
16 work with the census data.
17 Guidelines would be useful. I'm not going
18 to jump there. I want to move to general comments,
19 because I think the Census Bureau and you guys are
20 pretty good about making data available. That's not
21 a big problem.
22 I think that Shane's point about the RDC
0096
1 is excellent as well. People want to get to the
2 proprietary data. They can go in and access it in a
3 way we have been doing for years. That's a very good
4 idea and solves that problem in part.
5 I think more generally speaking, now that
6 the academics has spent your $7 billion on data
7 collection, what are we really, what do we really do?
8 I was discussing this with a professor
9 friend of mine the other day. Every outcome that we
10 observe today is optimal under the constraints that
11 we face. And the constraint is, is you got to beat
12 your head on the wall for ten years to get one
13 question added to the Census Bureau's information.
14 That point right down there. We have been trying to
15 get this done for 10, 20 years, whatever. It's just
16 to get a few questions asked.
17 We have to be realistic about what you
18 guys are going to be doing in the next year. You are
19 already on the hook for spending a lot of money in a
20 short amount of time. That will keep you pretty
21 busy.
22 So the question is, what is the highest
0097
1 priority questions that can efficiently be addressed
2 from the projects underway, considering the
3 constraints that we face? You are probably going to
4 get 10 to 20 questions out of all this discussion
5 that you can realistically expect in the next two to
6 five years. Okay. That's the way I see it.
7 But that's not really a problem, because
8 this is the federal government, and the federal
9 government is trying to make federal policy. And you
10 always deal in first order effects. Okay. We need
11 to ask: What's the first 50 percent of the R
12 squared, not the second 50 percent of the R squared.
13 What are the questions that are very, very
14 important? Given the fact that you got somebody on
15 the hook to give you data, what can you tell them to
16 give you? Not to ask a lot of questions.
17 Now, we have questions. We've got to ask
18 this race one group of questions and this race
19 another group of questions, and the people who live
20 over here one group of questions. That's not
21 practical under the constraints we have. We don't
22 have $7 billion to ask questions. We have $7 billion
0098
1 to deploy things and $100 million to ask questions.
2 Many of these issues will be dealt with at the state
3 levels.
4 Maybe the NTIA can give guidance to the
5 state organizations that are doing the mapping and
6 things like that, because they all do surveys in the
7 end. What are the kind of questions that you should
8 be asking so maybe we can get some comparability
9 across the states. Okay? What are those questions?
10 Well, a lot of the points here. What are
11 your incomes? Do you have it available? Check to
12 see if they do or don't, because many people who do
13 don't believe they do. How many people are in your
14 family? What are your incomes? All the questions we
15 looked at before. And to whatever extent we can ask
16 a few other things. One question that was
17 interesting that was done in Alabama, "Why don't you
18 have Internet in your home?" And one of the most
19 frequent responses was, "I don't want to let
20 pornography into my house." That is a very
21 interesting response.
22 Now, do we need to ask everybody that
0099
1 question? I don't know. Because now we know that
2 that's an important issue. We already discovered
3 that, not because of what you have done, but because
4 an organization that was founded by the fact that we
5 thought we were getting big money from the NTIA asked
6 that important question. If you ask all the
7 questions, you are not going to get enough energy in
8 the questions. We only need to ask one or two places
9 to ask the same question, and we get a nugget of
10 information we use. You get no competition in the
11 information that is provided.
12 So, I mean, it's a little general, but I'm
13 short on time here at the end. But I think the
14 question is: What is the priority and what can you
15 realistically do and what is the federal government's
16 role in all this? And I think it's very, very high,
17 a very high level. Maybe I will be surprised.
18 MR. STRICKLING: My remarks is that we are
19 going to have some number of grant recipients who we
20 can have supply information, and today I didn't hear
21 anything from any of you about what we might do with
22 those folks in terms of supplying information, and so
0100
1 I guess my question is: Did we not invite the right
2 set of presenters to answer that question, or, B, is
3 there a general sense that that is not a productive
4 way to go about collecting information on these
5 topics, or is it that you all didn't think about it
6 before today and take it as a homework assignment?
7 MS. HARGITTAI: I have a response. Are you
8 referring to the the BTOP program?
9 MR. STRICKLING: The BTOP program.
10 MS. HARGITTAI: I don't think, let me just
11 say why, because I think it's important to recognize
12 this, that these programs don't have -- I don't see
13 why they would have the staff to collect data that we
14 would need to then be able to understand what is
15 going on. So we have some of the very basic
16 infrastructural questions, maybe, but I'm not sure
17 what we do with the fact that there are five
18 computers in a computer center if we don't know
19 anything about what they are being used for.
20 On the other hand, there is not going to
21 be staff on these programs to do surveys of this
22 nature.
0101
1 PROFESSOR FLAMM: One quick comment. So
2 the mapping part of the --
3 MR. STRICKLING: I'm talking about the data
4 collection.
5 PROFESSOR FLAMM: The mapping part. The
6 BTOP grants, on the other hand, as an economist, you
7 are telling me you are providing me with estimates of
8 jobs created, for example? I got to go, "Come on."
9 The bottom line is, any estimate of job
10 creation always begs the question of what's the
11 baseline, what would have happened absent. That
12 requires some kind of scientific model of job
13 creation. Do you really expect these people who are
14 reporting this information to have sort of the same
15 common framework, scientific model they are going to
16 be running a model and producing an estimate of jobs?
17 What does that give you exactly? I don't know. So,
18 you know.
19 Now sort of basic questions. The people
20 who are receiving these grants have access to
21 information of things like what they are investing in
22 their networks. They might be able to tell you what
0102
1 they would invest absent the grant. Again, you ask
2 people these questions whether you give 100 percent
3 reliability.
4 MR. STRICKLING: Technically, we are not
5 supposed to give them the money if they didn't invest
6 in the infrastructure.
7 MR. FLAMM: Zero of this would have
8 happened absent the grant, or when? Because that's
9 another interesting question, whether it's next year
10 or five years hence. When you start asking the
11 question, there are subtleties that you have to think
12 about, and the idea that you are going to get a
13 diverse collection of 100 or 200 or 300 institutions
14 that think about this in a coherent way, as Eszter
15 was saying, it's a hard sell.
16 MR. STRICKLING: We have the ability to
17 provide templates, if we want to standardize that
18 sort of thing, but it comes down to: What is
19 reasonable to collect and what is useful to collect,
20 and then we deal with the resource question as part
21 of that. But at the end of the day, we are going to
22 be called upon to provide an evaluation of this
0103
1 program, and I guess we were hoping the research
2 community might assist us in that; otherwise, we will
3 figure out another way to do it.
4 PROFESSOR FLAMM: The bottom line, speaking
5 as an economist, the things that we use for the
6 things they measure, as part of our operations, that
7 they may be used if they are a business organization,
8 what they are investing, how many people they are
9 employing, what kinds of projects they are assessing,
10 announcing different kinds of projects. That may be
11 useful in analyzing what occurred. But what if
12 questions? Hypothetical. How much of what is going
13 on would have occurred absent the program? You are
14 not going to get a useful answer to that question.
15 The useful thing is to try to document, as
16 much as possible, the operation of the organization,
17 including, but not limited to, what is receiving the
18 money. That would be my response.
19 MR. FORD: I think you would ask those
20 types of organizational questions, but you probably
21 have to as a matter of procedure. I think that some
22 information about the market that is intended to be
0104
1 served. Okay. We might get some interesting
2 demographic information. We might say the BTOP went
3 to areas of poor or white or black, or, in the end,
4 we are assessing BTOP and not so much broadband.
5 That's what I view the purpose of this is,
6 and I did make notes for this -- because we got a
7 little sidetracked. How much money did you get?
8 What did you intend to serve? What did you serve?
9 And there we have the expected gain versus the
10 realized gain. What did it cost us to produce that
11 gain, and we should have some comparability across
12 places. It costs $4,000 for computer users in area A
13 versus $2,000 in area B, and we get information on
14 the efficiency of money spent in different ways, and
15 we say, "If we do this again, we may understand that
16 the middle mile didn't work so good," or whatever it
17 may be.
18 So, how do we better allocate money in the
19 future? Those are the types of information we need.
20 What did we get per dollar spent is probably the best
21 measure to guide you on that policy. So that -- and
22 I think we probably need, to some extent, classify
0105
1 the programs. And you guys know better the data you
2 work with and the variety. I looked at many of these
3 things, but I don't know them all. Is there some way
4 to classify this program, in a granular sense, so the
5 researcher can say, "This program was a local
6 community center program. This program was a middle
7 mile program. This program was a middle mile program
8 that also served a Wi-Fi network," or whatever it
9 would be, so we actually have some classification
10 that may probably be you guys driving that with your
11 intimate knowledge of what people are asking for.
12 But I think the primary purpose of
13 collecting data through BTOP is to say: What was
14 successful and what was not successful, so if we
15 spend another $7 billion, we know how to do it better
16 next time.
17 MR. WEITZNER: For those of you in the
18 back, if you can go to the mic that now works. We
19 will have Nicol and Shane.
20 MS. TURNER-LEE: Larry, I think you bring
21 up a good question, because one of the tragedies of
22 the first round of the Technology Opportunities
0106
1 program is, we didn't have a lot of data about
2 projects that were doing innovative things and what
3 worked and didn't work. I constantly tell people of
4 Commerce that I know of a handful of top programs
5 that are still surviving that no one would know was
6 there because they didn't have a chance to tell their
7 story.
8 If we were to look not only at the
9 infrastructure on the sustainable broadband adoption,
10 I think the data that went into the application
11 provides a good framework for preassessment data, and
12 the followup measurements we will see on the Web
13 site, as far as the indicators people have on the
14 update, are good.
15 What I would suggest from the conversation,
16 and the state of Illinois toyed around with this, is
17 to put in more qualitative questions around
18 challenges, opportunities and stories that you have
19 to tell. And the greatest value I think some of
20 these BTOP applicants can have with the technology
21 center movement was a good example.
22 There were cases where people had an
0107
1 opportunity to tell their story, and they had
2 processes where they had media releases, and they
3 shared about Ms. Johnson, who, because of that
4 investment, such and such happened in her life.
5 If we have that, then we can begin to
6 match some of the infrastructure investments what we
7 know about the broadband map, and having come from a
8 nonprofit background, the opportunity to talk about
9 what you have done, particularly in something of
10 strategic importance as this, is incredible for many
11 of the nonprofits. You saw the application number.
12 So I think there are ways to mix the types
13 of questions we are answering, because it's usually
14 hard to put in more than a sentence. But to find
15 ways where people can talk more about what they are
16 experiencing throughout the process. That would be
17 my suggestion.
18 MR. WEITZNER: Shane, and we will go to the
19 microphone.
20 PROFESSOR GREENSTEIN: Just to build on
21 Ken's spirited remarks is, you build off operational
22 data. Get revenue per user, because they will have a
0108
1 variety of different services and the through the
2 network. But revenue per user will be valuable, and
3 download speeds. You might even ask upload speed.
4 But the number one thing, download pamphlets, to be
5 precise, download bandwidth, and if you can get
6 actual, that's wonderful, though generally we don't
7 get that.
8 Finally, to Ken's remark, to do an impact
9 study, one would want to know before and after. So
10 we will get after from these operational data. So we
11 would want to know before. If you can ask for before
12 in a way that allows you to do a before and after,
13 then we can do an impact study, that goes directly --
14 that means we got to know what was available before,
15 what its speed was and price was, something to
16 compare, and then we will look at wages directly five
17 years out, two years out, one year out. We will look
18 at productivity, number of jobs, and we will go
19 straight to the county data, which we can do
20 immediately, and that comes out pretty fast.
21 We can do something pretty quick.
22 MR. WEITZNER: Could I ask you at the mic
0109
1 to identify yourself, and please go ahead with your
2 question.
3 MR. McCRANE: My name is Andrew McCrane
4 with Broadband Census. I have one question.
5 You had over 2200 applicants apply for
6 various broadband BTOP or BIP grants. Why not make
7 some of the data that you required them to submit to
8 you publicly available? If you don't do that in this
9 round, because of proprietary reasons, why not
10 include an ability to opt in for round two so we can
11 see a lot of the data, which I believe is
12 infrastructure. It's technology. Some of it is
13 speeds, and it is based on census place as
opposed to
14 census block. That would be interesting for us to
15 see. We are in the private sector. Map it. Play
16 with it, and just to see what it is. And that would
17 be a playground for a lot of other people in the room
18 here. So release the data you have now, or, for
19 round two, create an opt in so applicants can include
20 that.
21 MR. WEITZNER: Thanks very much. That is
22 helpful.
0110
1 MS. NORIEGA: Hi. I am Raquel Noriega. I
2 want to add a couple of thoughts, what George Ford
3 stated earlier.
4 One important stat we should ask these
5 grant recipients are: What are your expected
6 adoption rates and what are your actual adoption
7 rates, because that is perhaps the most important
8 thing we need to learn about this.
9 I think the ladies back there certainly
10 made the point, this is not about building pipes;
11 this is about building pipes that change people's
12 lives. If people aren't adopting it, then we are not
13 accomplishing our goal. So understanding what is
14 driving adoption and how adoption is growing is
15 something very important, particularly adding to the
16 questions that George Ford, with the demographics of
17 those particular groups.
18 MR. BERNSTEIN: Dave Berstein. I have been
19 writing about this stuff since '99. I don't have to
20 -- let's put some things on the table. These are all
21 questions that weren't answered here and should be
22 answered. We didn't hear any questions whatsoever
0111
1 about the cost of broadband. We heard a fair amount
2 about the price to customers, but in order to decide
3 whether the stimulus money is being wasted or not, we
4 want to know what is involved in the costs.
5 There are a number of things there, about
6 60 percent to 80 percent of the difference can be
7 explained by the number of miles of fiber or coax
8 that are required for a particular project. So that
9 number absolutely must be there.
10 You want to know on the mapping. I have
11 looked at it. Apparently, they are not telling us
12 anything about the unserved areas that we are trying
13 to reach. You want to know who are the providers,
14 where are the towers, is there fiber, because we are
15 throwing money at it. And the middle mile question,
16 most of the middle mile money is going to be wasted.
17 It's mostly going to over building places where
18 plenty of fiber is there already but the price is too
19 high. So when you over build, what you are looking
20 at is the difference between the price now and the
21 price after the over build. None of the indirect
22 benefits make any difference unless there is a big
0112
1 price delta there.
2 The second question on there. I have just
3 been doing some of the work at Columbia and looking
4 at the technology. It's pretty obvious that about 80
5 percent or 90 percent of the availability problem is
6 going to be solved. 90 percent of the stimulus does
7 as well. It's just there. So the question for
8 researchers is: Once availability is solved --
9 because it's cheaper to give everybody broadband and
10 to keep the PSTN running, so you put broadband on
11 every line unless policy fails -- how do you look at
12 the problems when adoption and availability don't
13 mean anything except for those who don't know, and
14 all the policy affects is speed?
15 And the last one I want to put on there is
16 -- really dramatic to see who came up here -- is:
17 How you prove a negative if in fact it's true that
18 nothing works significantly on broadband adoption
19 other than bringing down the price?
20 When you look at things like the first
21 work at Connected Nation -- and hopefully the stuff
22 they are doing now is going to be more effective --
0113
1 overwhelmingly, everything failed at sustainable
2 broadband adoption. There is no data on succeeding
3 so far. I wish it were otherwise.
4 But if in fact you could prove that people
5 know what the Internet is, so there is no way telling
6 them about it is going to make a big difference. If
7 that's true, that means we are wasting all the money
8 on sustainable adoption.
9 I wanted to throw the question out, how
10 could BTOP, or a researcher, prove, to the
11 satisfaction to people for a billion dollars, that
12 this is a waste of public money if in fact that's
13 true? I think it is.
14 MR. WEITZNER: Thanks very much. I think
15 those are helpful questions, Dave.
16 I am conscious of the time. We are 15
17 minutes past the point that we said we would hold you
18 here. So let me just ask who -- I know that Blanca
19 has a final comment, and if no one else does, I think
20 we will give you the last word, Blanca.
21 MS. GORDO: I think that the BTOP program
22 serves as a great opportunity for creation and
0114
1 evaluation of these types of policies, and as we
2 evaluate them, take into account the changes that are
3 underway and any set of challenges that nonprofits,
4 probably a lot of these providers serving the under
5 served are already faced with. I would ask them:
6 What are the challenges and how do you overcome them
7 to provide that.
8 I would be careful to evaluate them on the
9 basis of wages and all that. You can think that you
10 can design a perfect program that provides perfect
11 training, but the economy in that locality, there is
12 no economic demand, or you could potentially see that
13 you have a perfect economy that could capture people,
14 but maybe the type of training wasn't structured that
15 way.
16 So as we evaluate them, I think we need to
17 think about what are the inputs, what is the type of
18 service that is being provided, and how does that
19 coincide with the types of needs that the population
20 that they are serving are faced with. And what are
21 the current economic demands. What are the types of
22 networks. Are they structured to meet this need?
0115
1 Because they might not be. And to remember that, you
2 know, it's a learning curve. You know, because we
3 have adopted this type of technology, because we were
4 prepared and there was a plan and we adopted it, and
5 it worked. We are still wondering if we should
6 diffuse it or how we should diffuse it for areas that
7 -- while mainstream society has integrated this type
8 of technology, we were wondering should we take it
9 where. How.
10 So the operation of the organization is
11 important. We are charging organizations to provide
12 service or institutions to provide this type of
13 preparation, and yet themselves, you know, we still
14 haven't talked about how -- how does a community, an
15 economic development mechanism that was established
16 to address institutional shortfalls and
17 inefficiencies, how have we upgraded this mechanism
18 to be able to provide that type of service? I don't
19 think we know that just yet. So the structure of the
20 organization really does matter. So it would be
21 unfair to evaluate the kinds of contributions that
22 they are making.
0116
1 I think we are at questions of how would
2 we -- what's the learning curve? What are the inputs
3 that are necessary for those specific populations?
4 I think there was 84-year-old woman that I
5 met, and I said, "Do you like the technology?" And
6 she said, 84-year-old woman said, "Before I was
7 afraid of the computer. But today, the computer is
8 afraid of me, because I manage it." Is that a
9 contribution, in terms of how she makes that choice?
10 I met yet another Latino woman. "Is
11 technology important?" And she said, "Oh, yeah. The
12 other day, my daughter came home, and she was like,
13 Mama, I need the technology I need to do my
14 homework." "But we don't have technology." "Well,
15 then I'm going to get an F." She is running across,
16 "Can I use your technology? Can I use your
17 technology? Can I use your technology?"
18 These are the kinds of questions that
19 provide an experimental space that the valuation
20 where you select certain innovative programs and
21 control for the different conditions and the context
22 in the place and take that into account, and ask the
0117
1 people, what they are getting from this? We might be
2 able to isolate the benefit.
3 MR. WEITZNER: Thank you, Blanca. Thank
4 you, Anne.
5 Let me close now and just thank all of you.
6 We feel incredibly lucky that we have been able to
7 have the attention and thought of so many of you who
8 have done so much work in this area.
9 For us, we certainly hope this is the
10 first and beginning of a set of discussions as we
11 move forward with this.
12 I think we have heard a lot about ways to
13 look at the national data collection efforts.
14 Certainly we will take all this back and look at it
15 very carefully from the NTIA perspective.
16 We have been lucky to have a number of
17 people from the federal government, Commerce
18 Department, FCC, and from the different areas of
19 government. I think a lot of you have been helpful,
20 and you will probably have to hold all of our feet to
21 the fire on that and give us help, but we certainly
22 recognize the need to do that.
0118
1 And finally, I will say that some of the
2 questions that we posed are useful to you in
3 motivating the research that you choose to do.
4 We invited all of you here. We invited the
5 public here partly to let us -- to let you know what
6 kind of data we are going to have, but just as
7 importantly, to let you know what kinds of questions
8 we are going to have, whether it's the program
9 evaluation questions or the broader macro policy
10 questions. So we are counting on all of you who
11 provide us help and guidance, and look forward to
12 working with you all in the future.
13 Thanks very much and happy Halloween.
14 (The meeting was adjourned at 3:25 p.m.)
15
16
17
18
19
20
21
22
