personalize the information you get from us. just go to cspan.org/connect and sign up for the e-mail. the program guide is a daily e-mail with the most updateded prime time schedule and upcoming live coverage. word for word gives you the most interesting daily video highlight. with no commentary. the book tv newsletter sent weekly is an insider's look at upcoming authors and book festivals and the american history tv weekly newsletter gives you the upcoming programming exploring our nation's past. visit cspan.org/connect. and sign up today. john thompson join ed a discussion on the 2020 u.s. census. focused on the proposed question. the american institute host ed this discussion. it's an hour.

good afternoon. i'm carla, a senior fellow and i'd like to welcome you and our cspan audience on this session of making the 2020 census succeed. when mike asked if i was moderate, i responded enthusiastically. i know it is virtually impossible for many of my colleagues to do the important work they to without the data the census provide. the same is true for policymakers and for the business community across the country. any opportunity to talk about the value of our premier statistical agencies always worthwhile.

the dress reher stahl for the census is underway in rhode island. underfunding is a serious issue. there are other hearings that present challenges to this one. there are public concerns about information security and privacy as we learned last week. declining response rates and information explosion and rapidly changing technologies. different language requirements. complex living arrangements and finally, a mobile population. this is the first census that will have an online option. funlly, there's the inclusion of citizenship status and what that could do to response rates.

i hope our panelists can touch on some of these issues today. let me just say a word about each of them. they'll speak for about five to seven minutes then we'll have an open conversation. john thompson is executive director of the counsel of professional associations on federal statistics and more pornlt for this panel, the 24th director of the census where he served from august 2013 to june 2017. he introduced the director's blog, which i read when ever it appeared. he was president of norc at the university of chicago. diane is director of the institute for verge at northwestern and the margaret walker alexander professor in the school of education and social policy there. she is also a research associate and was director of the hamilton project at brookings. that's an initiative to promote policy to ens hans economic growth.

and manager of the heritage foundation's law initiative and a senior legal fellow at the me center for legal and judicial studies. he writes about civil rights, election integrity, immigration and government regulation. heed fseved for two years at th federal election commission b ab bd before that, worked at the justice department. my colleague here is director of economic policy studies here and the john g. serls scholar. he oversees a broad portfolio. all of the institutes in financial markets, tech policy, energy economics and health care policy. mike worked at the census bureau. we'll begin with john thompson. >> thank you very much. i'm delighted to be here. michael asked me to go back five minutes and sort of put the census in context. describe the readiness so

attempt to do that. so i think everyone knows that the census is constitutional mandated. article i, section ii. it has a number of important uses including reapportionment of the congress. drawing of districts, both congressional and local voting districts. it's used to allocate over $600 billion a year of federal fund annually. it is essential for state, local and tribal governments to make many decisions it supports business investments in local communities. finally,s it's used to ensure the accuracy of every demographic survey collected in t united states, which could include the current population survey, which produces month to month survey, and employment.

the national health survey and the american community survey. that's why there are so many concerns about the accuracy of the census because of all the important uses. it's important to understand that the census process has e violaed. in some way, and in others, it hasn't. so gabeginning in 1970, the cens bureau went to a mail back census. that process has been used until today. that is you prepare an address list. electronic caption the information off the paper questionnaires and then you go back and collect the information from those households that do not self-respond. until today, that operation has been paper and pencil. and so as our country has grown more diverse since 1970, the only they a paper and pencil

operation, you have to have more and r more people. and faster an population growth. and again, it's because of using a a paper and pen sip sill ration to do something that's very difficult. for the 2020 census, the idea was to modernize and use automation intelligently. that went from developing the address list so you don't have to walk every street in the country to use the internet at the primary response option and the census bure realizes there are households that don't have access to the internet and they provide those a paper questionnaire and they also allow people to call in for the first time to give their information over the phone. importantly, they are using

mobile technology to deploy the workforce to collect the information. they're going to be using iphones and ipads to do this. it introduces tremendous efficiencies into the operation. the census bureau estimates in 2024, if they received the resources they needed, they could have reduced the cost of receiving the 2020 census from about 17 billion to a little over 12 billion. that would have been about a $5 billion savings. however, if you look at the am of money the bureau asked for and the amount they receive, you'll see from 2012 through 2017, the bureau was underfunded by $200 million.

this forced them to make a number of decisions. they had to defer some operations. they had to reduce three test sites to one. which was a function of thchlt they had to defer and this was very important. we can talk about this later. they had to defer activities on their combined advertising and partnership program, which is how they achieve accuracy in the census. i think the good news is that the census bureau in their fy 18 budget did receive more than than they asked for. over a billion dollars more. this really indicates bipartisan support and it will help them tremendously to overcome difficulty, so it was a r very, very good move.

they still need a lot more money though. the census recently commerce secretary decided to add a citizenship question to the census. i and five other directors have written a letter, which is publicly available. expressing our concerns about th that, and i'll speak for myself, i don't believe there has been adequate testing of such a question in a contemporary census environment and there's a great risk to including it on the census. an undercount will be with them for ten years. tgs very important we have an accurate census. with that,ly conclude.

>> can you tell us about how the dress rehearsal is going? >> well, i was at the project management review and they mentioneded at the tiat the timr self-response rate was about 36%. they said that was on track with what they're looking for. the next big step in the census bureau said their automation was ready to go. we need to see it and how it works in the end to end test. make sure it does work. the gao will be looking at it so everyone will know how well it went. >> diane. >> thanks so much. it's a pleasure to be back here. michael and i wrote a paper along with some other co-authors about a year ago where we tried to lay out a framework for why it's so important to have well

functionings census bureau and other statistical agencies. but those are very important to the united states. michael and i will each talk a little bit more about that. but in terms of the census, we know a a successful and accurate census is crucial to the united states for a number of reasons. of course this determines the number of seats each state has in the house of representatives. but it afgts the distribution of billions of dollars in federal funding every year to local communities for infrastructure, vital services like hospitals, schools. we're facing increasing challenges to collecting the census. there's the increasing cultural and economic diversity of the united states just makeses it hared to go out and survey everyone. laid on that is increasing distrust of the government. in the past, the census has cost overruns and they came up with a

very innovative plan to keep the accuracy high, but also reduce costs this time. they were they are still planning on implementing a range of new collection methods. including this new internet and telephone response options as well as the traditional paper questionnaire. hopefully it will make a lot of things more streamlineded do doe road. so by making these smart investments in technology, in messaging and advertisement and so on, to accurately count a population that's increasingly hard to count, maybkes a lot of sense. we need to make those investments today, not two years from now. if we wait until it's too late, we'll pour more money on the problem to make sure we get up to the accuracy that we need. it's the first time really modernizing the senn is sus ace.

we've scaled back some of these tests. it's important that those are conducted with fidelity and that we can troubleshoot. we're not just talking about undercounting. it's not just immigrants. although to be sure and i predict add iing a citizenship question will increase costs of this because more people will not want to answer this, but we need to worry about undercounting or miscount og f the elderly. when we were moving now to something that involves the internet trks predictable who's going to be less likely to par tis pate in that. also, the statisticians tell us it's going to be difficult to count people who have multiple residences and if we don't do that, we'll end up not just undercounting, but miscounting people. that could have real consequences not only for the number of people in the house of representatives, a portionment,

but also on dollars spent. to conclude, we've heard ar the need for smart veinvestments today. t we're also concerned about leadership. the it's long benefitted from exceptional bipartisan highly qualified leadership and we need to make sure we fill this role is someone who's highly qualified and nonpartisan who will be respected on both sides a as the new director. thanks. >> diane, you mention ed the online or the internet option. do you see that as enhancing participation or are you worried that it. >> certainly in the long run u. it should. it is well time we modernize the census, but it's just going to require a lot of tesing and a lot of understanding who can we reach better with this and who do we have to do more following up with. >> and the privacy concerns?

>> i think we'll get into that for sure. i want to thank aei for inviting me to be here today. michael asked me to talk about probably what's raised the greatest controversy over the census, that is the announcement at the end of march to add a citizen question to the census. now i think a lot of the criticism against the administration for doing that is unfair and unparented. in fact, i think those are wrong and i want to give you a quote from the president. and i think the president is correct when he says this. the quote is for the purpose also of more exact ly distinguishing the increase of population by birth and immigrati immigration, we know know the numbers of citizens, of foreign birth and aliens. now you think i'm just quoting president trump. you are incorrect. i was just quoting president thomas jefferson who said that

in 1800 and in 1820, the census followed that and added a citizenship question to the s n census on quote, foreigners not naturalized. if you've listened to some of the hysterical plame claims by opponents or read the coverage in the paper, you would probably not realize that what the trump administration is not adding a citizenship question, but reb instating the question. it was asked on different censuses until 1950. when the u.s. bureau switched to sending out a short and long form. those of you who are older in the audience will remember one in six households got the long form. it had a citizenship question on it. as demonstrated by the long form from the 2000 census. not that long ago.

after the to 2000 census, it was discontinued and they switched to the american community survey. this is sent out continuously, but om goes to one of 38 households. and it only provides basically estimates of demographic characteristic, which makes its use nationwide problem, the acs, which is used today, has always had a citizenship question on it. if you want to download the form, you'll find the question and the commerce department said thai going the add the same question to back on to the census and keep this in mind. it doesn't ask you whether you're here legally or not. it asks if you're a citizen of

the united states. there's a law that bars the federal law from sharing the information with any other government agency. that's a criminal violation of the law if they do that. so the fear that somehow this information is going to get sent to dhs for example is just not true u. by the way, when i hear people talk about how this is going make the response rate go down b b, there's no evidence it's going to make the rate go down. no evidence it's caused the response rate to go down to the acs. in fact, the acs, if you look at it, it has a lot more intrusive questions on it. i could give you samples that would upset people a lot more than a citizenship question. by the way, if you look at the memo, that was put out by the commerce department, you'll see ta talked to a former deputy director at the census who told commerce that when they added a

question to their surveys, they've seen quote, no apreesh bable decrease in response rates. another quote from there that i think is correct is this. from kevin kelly of the commerce department. even if there's impact on responses, the value of more complete and accurate data derived from surveying the entire population outweighs such concerns. because remember, acs only goes to one out of 38 households. diay a ta is used for everythin from enforcing the voting rights where census data is available for federal fund uing. we've been in a debate over immigration policy now for ten years. does anyone doubt that we should have accurate data? on is sip senne ship op patience if we want to have an informed debate about i and come up with

a remedy? finally, what i know will surely make the opponents of the administration and adding this question, grind their teeth. is that the trump administration is actually following recommend da made by the united statesed nations f. you look at their recommendations for housing in 2017, you'll find they recommend a citizenship question on their census surveys and in countries ranging from australia to germany to indonesia all ask this question on their censuses. only in the u.s. is this considered at all controversial and it shouldn't be. thanks. >> thank you. i suspect you'll get some disagreement from our next panelist, michael. >> the typically b one of the problems speaking last is that there's not much left to say. but that's a nice segway to provide a contrasting view on the citizenship question. i am concerned about the impact

the question will have on response rates. it is of course the case that we currently haask about citizensh on the american community survey. it's also the case that the survey is a much less salient exercise in data collection than the census. the census bureau has found evidence that people in immigrant and minority communities today are concerned about the 2020 census. in a way they are not concerned about the american community survey. and you know, the reasons are easy enough u to figure out i think. fist, the descending census is a major undertaking. knock on every door. count every person. the government takes out a lot of money to advertise. the exercise in order to get people to comply with the survey. it's in the constitution. it's something that you know, people notice. people may much less attention

to the american community survey and they should pay less attention to the american community survey. so it's hard to kind of map what happens with the american community survey on to what happens with the census and to expect that people will respond in the same way. it's also true that in conduct ing tests for the 2020 census, the bureau has found that members of immigrant and minority communities are concerned about answering the census at all. this concern was registered before there was a decision made about the citizenship question. so even a part from whether r or not the 2020 -- has a citizenship question, the bureau has found members of immigrant and minority communities are concerned about answering the 20 census at all. and the reasons for that are not terribly hard to understand. there's been a lot of antiimmigrant rhetoric coming

from washington. there's been a lot of rhetoric coming from the president of the united states, both after he took office and when he was a presidential candidate. there's been a debate b about m imfor sure. the status of the dreamers and all sorts of things have kind of led to this moment. where people in some immigrant communities are concerned about answers questions in general and it's quite reason bable based on the evidence we have to conclude they will be even more reticent to answer a government form that asks about their citizenship status. why should we care about that? an important reason is that if people are less likely to answer the question, the cost of the census goes way up. the census is in the

constitution. the bureau does not have a choice about whether or not it atee attempts to enumerate the population. that means if you don't answer the form f you don't fill out your form, the accept sus bureau does not just does not just sa we tried our best, instead they send someone to your house, i'm here from the government and have a few questions for you. the bureau doesn't say we tried our best, they send somebody back. if you want to answer the questions they ask your neighbor to say hey, how many people live in this house. sending human beings costs money. if we want to keep the costs of the census under control which i think we all should, we should not be putting questions on the census that will make it more necessary to send more people to gather this information in person. it also affects the accuracy of the census, and that has ripple

effects for the next ten years, the 2020 census will be used to benchmark the current population survey, the american community survey, other demographic and economic surveys for years and years and years, and any small errors in that benchmarking will be with us for a decade, and those small errors compound over time. very important decisions are made using those data sources. decisions about interest rates, if the federal reserve looks at the current population survey and the current population survey mismeasures unemployment a little bit and the fed tightens a little too fast or the fed lets the economy run hot a little too long, billions and billions and billions of dollars are at stake. if social security, which adjusts its, which adjusts its payments based on inflation, if those inflation measures are off just a little bit, billions and

billions of dollars are at stake. private sector businesses rely on data that are benchmarked to decide where to open distribution centers to decide what items to put on store shelves, to make all sorts of critical decisions. if the data they're relying on are just a little bit inaccurate, then those costs compound over time. so small errors have big effects when you consider how important the census is and how long we have to live with it. again, another reason not to introduce inaccuracy by including a citizenship question. those are in costs of potentially including this question, what's the benefit? i don't see much of one, if at all. we have accurate, valid information on citizenship that comes from other less visible data sources. we have been able to rely on that for decades and decades, so

to me there's no clear reason to do this, and when you have a proposition where there's very little, if any, benefit and significant costs that says that we shouldn't do that particular thing. so my hope is that congress intervenes and keeps that question off the dicennial and i think it would be better for business and better for policymakers if ultimately better for taxpayers if that ends up being the case. >> thank you, mike and john and diana, dump in, first a question fuse for hans. how do you respond to mike's points about the salience of the ucs versus the cost, the accuracy and the benefit? >> the problem with the acs is that they only go 1 to 38

households so they extrapolate as to how to apply it to the rest of the country. that causes problems in enforcement of the voting rights act. i know that because i worked at the justice department enforcing the voting rights act and the reason that's important is this. if you are enforcing section two of the voting rights act, which prevents and prohibits racial discrimination in voting and you are, for example, filing a lawsuit against a state or a county or a city, claiming that they have deluded the vote, by the way they have packed particular districts, and you are suing on behalf of say hispanic voters, well, if you're coming up with a remedy, your remedy is to come up with a district in which minority voters can elect their candidate of choice. now the way you do that is you have to ensure, for example, if you're suing on behalf of hispanic voters that when you create a district that they have

at least 50% plus one of the population in that district, so they can elect their candidates of choice. well, if you just use population and you only have a citizenship rate of 60%, putting a 60% hess pan hispanic population does not give you majority hispanic voters. the acs, because it goes to such a limited number of households does not give you good enough data to do that. you can look at u.s. supreme court decision from a couple of years ago where this was issue and the supreme court talked about the lack of data in this area on this. census question will get us that for the whole country and i would repeat what the commerce department said which is there is no empirical evidence this is going to cause any significant downturn in the response rate

and if it did, we would know about it for the acs, which the census bureau produces all the time. >> diane, would you like to join this point of the conversation? >>. >> a couple points. one, the theory behind the aca and we realized it was replacing the long form because it was going to produce more accurate data than the long form. the acs has two products. one product that comes out every year and another product that comes out on an average over five years. think of it as a five-year average and that produces data for very small areas, census track and block groups that have the same level of accuracy as the census long form. the american community survey produces the data the same quality for citizenship that we have observed since 1950.

we've been able to do that. so i think i will stop there. >> diane? >> i think my primary concern is about the accuracy of the census overall and that it sounds like hans disagrees about whether this will drive up costs but i think there's significant risk. i'm per swayed by michael's arguments this will reduce the accuracy of the census or drive up the cos. the benefit what have we get over and above the american community survey is quite small. >> mike? >> well, i don't think i have much more to add. i do think it's, again, just to repeat myself i think it's very difficult to compare from, to compare what happens with the american community survey to the

dicennial census. everybody understands it is happening. the u.s. census bureau does not spend millions of dollars to help with compliance on the american community survey in the same way they do with the dicennial census, the bureau found local community leaders in immigrant and minority communities like church pastors or other similar community leaders in some cases are telling people in their communities not to answer the dicennial census because of some of the anti-immigrant rhetoric that's been coming from the administration and from the president and coming from washington generally, that does not happen with the american community survey. the american community survey gives us the information that we need to do things like enforce the voting rights act, the american community survey's question on citizenship is as accurate as the previous

question that was asked once every ten years on citizenship before the american community survey was adopted, so we're not losing anything by way of accuracy. the short, simple, you know, ten or so question dicennial census has not included a question on citizenship in decades and we've been fine enforcing the voting -- maybe not fine, but we have not needed to add a citizenship question to the short form in order to enforce the voting rights act over that time as well. >> well i think we have some pretty profound disagreement on that and i think we'll probably have some questions from you in the audience later. i'd like to turn to john to ask just about how some of the kinds of questions that the kren cuss has been asking over time have evolved in terms of the complexities we have in this society today, the multiple languages, the complex living arrangements, the mobile population, how are you coping? how was the census coping with

that when you were there and how are we coping today? >> so i think that you're hitting on something that every statistical agency and every company that produces data is facing, that is our population is becoming much more diverse, much more complex to measure, and so you have seen an emergence of a variety of techniques to reach out in new ways to individuals. you've also seen questions that have evolved more to measure the population, so for example, the census relationship question now asks questions on same-sex marriage or same-sex partners, where it didn't before, so you see sort of an evolution of the questions that we ask that parallels our society. >> diane, do you have some thoughts about this? >> mostly harping back on and that makes everything more

expensive, because we have to think about how to measure this, we have to test how to measure it, et cetera, et cetera, and in order to stay accurate we need to make these wise investments especially. >> john, can you tell us a little bit about how it actually happens? it seems to me it's mounting an army to do a u.s. census every ten years. how many people are involved? >> i think at peak right now the census bureau is looking at hiring somewhere between 300 and 500,000 temporary workers in 2020 to collect the information from those that don't self-respond and they'll have a set up of about 250 temporary offices to manage those people from, and they have to set up the information technology and the infrastructure to manage that kind of operation. >> how many languages? >> so i'm not, when i left they

were looking at having the questionnaire in five languages and they'll provide assistance and up to probably 100 languages that they need, so they're very attuned to the languages of our population. >> yes, that's impressive. we certainly heard a lot last week when mr. zuckerberg testified about information security. how concerned should americans be about the privacy and information security of the census? >> so -- >> you're the point man, john. >> the krcensus bureau, their information is top priority in terms of building systems and they work with expert in the field. they work with the department of homeland security. they work with the national institutes of science and technology, they work with private sector companies, and before any system is imflemted

it has to undergo some serious testing, including penetration testing by an outside person. that being said, in the world of cyber security, you can never stop being worried. you always have to be working very, very hard at that. so i felt comfortable that the census bureau was addressing the cyber challenges but it's something you can't stop doing. >> other comments on this point? >> the one thing i would point out is that throughout our history the census bureau has been very good and keeping the information confidential, with one very big exception, the very big exception that i'm sure people know about is the fact that franklin roosevelt in essence ignored the law and they used census data during world war ii to locate japanese-americans and go after them. it was a violation of law, should not have happened but it

did. that's the only real significant example of census breaking the federal rules about confidentiality of census data. >> in my world of survey research, we worry a lot about declining response rates, and the wonderful work done by the pew research center shows that, for example, about 20 years ago we had a 36% average response rate for public opinion surveys that's down to about 9%, it's held pretty steady over the last few years in pew's data overall. could you give us clues how the response rates have changed over time? john you may be the person we turn to first.

>> we'd love that in the survey world. >> but if you look at some surveys like the current population survey, which is one of the highest voluntary response rate surveys that the federal government conducts, you can see a steady dropoff in response since 2010, and it's been a real problem for the census bureau. they're working very hard to address it, but it's just getting harder and harder to convince people to respond >> have you taken surveys why people aren't responding? diane, you can weigh in here, too. ? >> i want to the make sure that everyone understands when we do public opinion surveys or the current population survey, et cetera, you have to have a benchmark to compare the response and non-response and those two. that benchmark is the united states census, which is all the more reason why it's just critically important that we get

an accurate census in the year 2020. so this measuring stick that all of these other surveys are measured against, and sort of adjusted for, when there's non-response, when there's sampling and non-sampling error, it has to be accurate. we want to make sure we get this conducted to the level we need to be. >> i might mention, this often surprises members of the public, surprised me when i took a look at it, but i often get calls from people who are upset about the american survey and not upset about the citizenship question. they're upset about the detailed income questions. most people don't like the idea of having to tell the government what their income is. what you all should know is there's a federal statute, makes it a crime for to you refuse to answer a census.

you can be fined up to $500, some people might think it's worth it to not answer that question and there's reported cases, most of them are very old, where people were prosecuted by the justice department for refusing to answer the census. it's always an option the government has. they don't exercise it very often. >> we have 15 minutes left today and i think it's time to turn to your questions and please identify yourself and ask a question. we could start over here. please wait for a microphone. >> andrew reimer, george washington university. the secretary's announcement of what's on the questionnaire is really his proposal, and that there's a process now of census submitting to the office of management budget, the actual census form in omb through the paperwork reduction act has to approve the 2020 census as it

approves every federal survey, and i think we should see that announcement in a couple weeks, and that the public has 60 days to tell census what it thinks of this design, and then census submits it to omb and the public gets another 30 days later in the year. if you could just describe this process to people, and omb's power to decide what goes on the survey form. for hans i think you said two things was incorrect. in 1940 law passed may have been morally wrong. the second thing is the fine for not answering the census while the organic census loss is 500 bucks for lying or 100 bucks not filling something out it's up to $5,000 because in the '80s

congress passed crime control legislation that overrode statutes that were specific to individual agencies so if you go to the census website it says can be fined up to $5,000. the third thing, sorry, i'll be done, there is broad evidence to its national advisory committee regarding recent as in the past year fears of respondents to answering the citizenship question. so there was a presentation with descriptions in great detail, this was raised to the level of alarm among census stuff that people were lying, people were walking out of the room, people were visibly sweating and this was not mentioned by the secretary. so the census internally has seen evidence compared to a few years ago of change in people's

attitudes regarding the citizenship question. could you respond to that? >> hans? >> andrew, you were just about right-on with the process. participator work reduction act requires information collection to be approved by the office of information regulatory affairs at the office of management and budget, and so the census bureau will put out the questionnaire for public comment, that's the question they'll use in 2020, and they'll have 60 days, then i believe sometime around september/october that whole package including the comments will go to the office of information and regulatory affairs, and they can approve or disprove any of the questions on there and have documents why the questions are important and the

research and that's the next step in the process. >> hans? >> on the process question, i will answer that by mentioning the fact that as i'm sure you know more than a dozen states have sued the government over the addition of the citizenship question. i think the suits will all fail and the reason is if you look at the constitution, the power to conduct the census is given to the federal government, not the state states and there's a specific statute that gives the commerce secretary to the power to "determine the inquiries" to the census form. the states don't have any claim to this, their claims are without merit. you're talking about anecdotal evidence. there's no impeerical data that shows a reduction in response rate. the nielsen firm said in their surveys they've seen no reduction in response rates and again i've said this more than once but i'll repeat it, the

census bureau sends out the acs all the time with a citizenship question on it and if it had any empirical evidence that the citizenship question reduced the response rate on the acs, they would have produced that and they didn't. >> question from this side of the room and right here? please wait for the mike, right here. thank you. >> thank you. my name is afshibat. i just wanted to ask a question, trying to find if the census, we' we've, knowing what's happened with the lecks, meddling and what happened on facebook, how do you go ahead and open a window that will try to prevent what could be happening in this census bureau? could it be some kind of

external meddling in the census, knowing what is happening now? >> john, is the census or diane, prepared for this possibility? >> so i participate part of the issue, any kind there's a breach of public confidence it falls on all of the different agencies collecting data. the census bureau has a harder job in communicating, which they will, that the data they have is confidential and they have processes in place to protect the integrity of the data, but every time something like that happens, it makes it harder and harder for them to get that message out. >> diane? >> i think similar answer, which is that we need to make these investments in technology to make the whole thing more

productive, but alongside that, we need to make sure that we are doing everything we can to secure the data especially anything that's not online. the census bureau has a terrific track record, historically in terms of this. they take confidentiality extremely seriously, they take data security extremely seriously so we'd need to get somebody who is an expert in those specific things to testify to that. i have utmost confidence in the bureau. >> a question over here, and then over here we'll go to that one in the corner first and then to you. >> thanks. peggy, i'm the congressional correspondent for the hispanic outlook. two questions. it seems lucy goosy stuff, the number of latino in the community, hispanics which of course are not a race recognized by the census bureau, so how do they come to this, i understand

last namsz. congressman tancredo gets stuff all the time, how do they determine the hispanic commute not less the latino vote and then along with that is for a long time they could not estimate the number of illegal immigrants in the country, suddenly we get this it seems like in concrete 11 million, that hasn't changed despite so-called record numbers of deportations so how do they arrive at that? >> john? >> i can address the hispanic question. right now the census asked two questions related to that. the first question asks about hispanic origin, and the second -- and it has different categories to check, and the second question asks about race.

part of the issue that the census bureau had been addressing over the decade was understanding some of the issues associated with an other race category and those who were hispanic it would sometimes mark that category. they had done some significant research on combining the two questions together, and they concluded that that was the optimum way to ask race and ethnicity. however, they have not been able to implement that, combined question, technology, because the office of management and budget hasn't changed the 1997 standards on race and ethnicity that requires two questions. it's self-reported. >> yes. >> nobody is looking at last names. they're asking people how do you identify. >> not just spanish, includes people from not spain. >> it's how anybody chooses to answer the question. >> it's self-response. >> it's self-responded, how

individuals choose to answer the question. >> number of immigrants or -- >> the census bureau when i was there wasn't estimating the illegal population so i'd have to defer to some of my colleagues and maybe at pew or something. >> that number comes from the department of homeland security and i'm not quite sure how they come up with the number. >> question here in the middle. >> karen gagliano. i think i heard you say there was going to be an electronic distribution of the census? is that correct? >> online. >> online. >> the census bureau will send out an invitation by mail to respond over the internet. they're not going to send emails

to people saying, respond. >> thank you for that clarification. what intrigues me is there's so much focus on the legal/illegal question versus the fact that not everyone is electronically literate, and that cuts across a lot of populations, aging, you know, poor, people who don't have access to computers, et cetera, and i would think that that, in this day and age, would have an even bigger impact on response rate and the skewing of responses. >> so let me just say a little something about that. the census bureau will mail out, they have two ways that they invite people to respond. one is called internet first,

where you get a letter or a postcard that says please respond by the internet. and at some point, in the process, if people don't respond, they will get a paper questionnaire but it's internet first and the last mailing would be a paper questionnaire. the second method is internet choice where they identify areas in the country using the american community survey and other data where they believe that moat people would rather respond by paper and for those areas they mail out a questionnaire. that will be about 20% of the households in the united states will get a questionnaire in the mail, and then they'll have the choice, they can either respond by the internet or fill out the paper questionnaire, and also, if they want to call in, they can call in and give their interview over the telephone. this is the first for the 2020 census. >> i get all that, but from a research standpoint, everyone in this room is probably electronically savvy, and i

would be concerned that the response rate among certain populations will be higher because we'll just go to our computers and respond, and other people -- i understand they have these other options, but this is quick. >> absolutely, so that's i think one reason why the end to end test is very important, so that we can understand before we field the full 2020 census, who is going to respond and not going to respond, and what cases do we need to send those 300,000 to 500,000 enumerateors to the hope to help people respond. in order to make this step forward which i think is needed for financial reasons, and to modernize the census, we need to test and understand the strengths and weaknesses of the new approach, that's why we need an end test but we've got got one where we should have had three. >> there's a question in the

back here, then up here. >> my question for hans. in your opinion what are the effects of having the citizenship question, what i'm thinking about is like california, they implement policies to attract population of illegal immigrants, versus alabama, let's say, when if you count all that illegal immigration, immigrant population in california that will give california more seats in congress, and more votes in the electoral college, what is your opinion about this situation, because our state will lose representation with higher citizen population. >> well that happens anyway. the portion is based on total population. it asks whether you're citizen, not whether you're here legally or illegally. because the census counts total

population, because a portion of it is done based on that total population, what it means is that states with large numbers of aliens who are here illegally actually do get more congressional seats. there's a number of studies done about this, one of the last ones that i saw had an estimate that california has between five and six extra congressional seats because of their large illegal alien population. texas is estimated to have two and there are some states that, if apportionment was based only on citizenship population they'd get additional seats which they are not getting. that's not going to change because you have a citizenship question on here. that would only change if the law and you might have to have a constitutional amendment governing apportionment was changed. >> time for two. >> one second, it's important to

be clear about that last point. the constitution is clear that congressional the apportionment of seats in the house is not determined by citizenship. it's determined by the number of people who were living there, and so including the citizenship question won't affect, well depending on how it affect accuracy and response rates et cetera, et cetera, but knowing with more precision the number of citizens in each state doesn't affect how many house seats each state gets which in my view is another reason not to include it. >> hi there. i struggle a little bit with a little bit of a chicken and egg problem when it comes to the citizenship question and responses. of course there say level of fear in the immigrant community

from the government. immigrants are one of the undercounted populations in every census and there's a census automatically looks at that, we have an undercount and we figure out how much it is every year. but it seems to me that the more we talk about the krn about this the more people might get scared so you have this sort of chicken and egg. where as everybody talks about how scared people are to respond, people would be more scared to respond. in past years, there have been efforts by immigrant advocacy organizations to encourage responses to the census, recognizing that the law says it cannot be used trying to advertise that, and so one of the questions i have for mr. thompson and others is how does this census or does this census work with other organizations nonprofits to try to organically support response to the census, try to alleviate some of the fears and how can we sort of balance, i mean if this is going to happen, whether it does or doesn't, if it's going to happen i think there's still valid

reason to have everybody counted for every reason that you have said, how can we encourage more responses in this age of fear? >> sure, since the 2000 census the census bureau started a program of paid advertising, combined with partnership, when they hire individuals to work in local communities with local governments, with local places, local leaders, and get local voices to talk about the importance of the census and the qualify depp shalt confidentiality of the census. the last census approximately 250,000 partners nationally to help them get the word out about the importance of the census. so they're going to do their best in 2020 to form as many partnerships as they can. the issue with the citizenship question is that it may make it

harder for them to get that word out about the confidentiality of the census >> one question in the back. you have the final question, sir. >> good afternoon my name is todd wiggins. i want ask if we could end with a happy note. i don't believe most of what is collected in the census is accurate because it's based on voluntary response and so you have to skew that. it's kind of like advertising. so where is the census most accurately and ideally collected? what country in the planet does the best job of getting accurate information and has the best incentives for people to answer accurately, and how does that work, and can we repeat that here. >> so -- i don't even know if

there is an answer. it may not be knowable. my first thought, after hearing your question, which is a good one, i think, is just to point out, how simple and sparse the dicennial census is. it's very basic questions, how many people live in the household, how old are they, what are their races and ethnicity, again, is an argument against putting controversial questions on there, and that is likely collected quite accurately, but some of the hard-to-reach groups who live in rural areas, members of immigrant and minority communities, the elderly, people who live in what census calls group quarters like college

dorms, that sort of stuff, there it starts to get difficult, and i think that difficulty highlights the need for the census bureau to be adequately funded in the run-up to the 2020 census, because that money needs to be spent today in order to make sure that, when it comes time to send out the questionnaires and when it comes time to send out the enumera enumerateors to knock on the doors that people know how to get to the hard-to-reach communities and people less likely to answer and get inaccurate response from them. >> we've covered an enormous amount of ground in a short period of time. i'm sure our panelists will stay to answer any questions you have. we want this civic project to be an enormous success and what all of our panel is are committed to. i'm remined of thomas jefferson and george washington thought

there was an undercount so to be conditioned these debates go on. thank you for coming and thank our panelists.

friday morning we're in salt lake city for the next stop on the see pan bus 50 capitals tour. gary herbert with be our guest at "washington journal" starting at 9:45 a.m. eastern. saturday our live coverage of the 16th annual annapolis book festival starts at 10:00 a.m. eastern and includes msnbc's chris matthews with "bobby kennedy: a raging spirit" april ryan with her book "the presidency in black and white: my up close view of four presidents and race in america" tech entrepreneur "the sentient machine" and co-founder amy sis kind "the list: trump's

first year" and "raven rock" the story of the u.s. government secret plan to save itself while the rest of us die. watch live coverage of the festival saturday at 10:00 a.m. eastern on c-span2's book tv. "new york times" versus united states the pentagon papers case. president nixon is using his presidential authority to prevent the documents related to the vietnam war. lower court judge stopped the presses for the first time in american history. the presses have been stopped someone fears the exposure might be dangerous to national security. another judge refuses to stop the presses and is proud of that. all up to the supreme court. >> the pentagon papers case, i haven't seen the movie but it's a great story but it only stands

for the proposition that the government can't stop the presses in advance, but the court actually acknowledges there's a possibility that once the "new york times" and "the washington post" published this, there could be prosecutions afterward. >> i think the gravitational force of the "new york times" case has created a political atmosphere where, within hugely broad bounds, we do not go after the press for publishing things, even where the statute seemed to say that we could. >> watch landmark cases, "new york times" v. united states with floyd abrams who represented the "new york times" in its case against the nixon administration and ted olson live monday night at 9:00 eastern on c-span. joshua geltzer, former counterterrorism senior director at the national security council under the obama