U.S. CENSUS BUREAU
WEBINAR
“PREPARATION FOR RELEASE OF 2005 ACS DEMOGRAPHIC AND SOCIAL DATA”
SPEAKERS:
STEPHEN BUCKNER,
PUBLIC INFORMATION OFFICE,
THE CENSUS BUREAU
DOUG HILLMER,
AMERICAN COMMUNITY SURVEY
J.R. WYCINSKY,
AMERICAN COMMUNITY SURVEY
AUGUST 2006
Transcript by:
Federal News Service
Washington, D.C.
STEPHEN BUCKNER: (In progress) – everyone. This is Stephen Buckner in the Public Information Office at the Census Bureau. I’m glad to see a lot of familiar names and voices on the line. For those of you who have not worked on the American Community Survey stuff or are recently getting thrown into the Census Bureau beat, we welcome you as well and look forward to working with you.
Today what we hope to accomplish is – through this webinar is to provide you with some updated materials and information about the upcoming release of the 2005 American Community Survey. We sent out a media advisory a little over a week ago to try to provide you with as much information as possible, and we’re going to keep on trying to give you as much information as possible on a flow basis.
One of the most recent updates is a 2005 ACS data user guide that is available on the Internet, which will be changing some time late this week, but right now it is on the ACS main page. It also includes all of the 2005 table shells that I know many of you have been interested in.
At the Census Bureau, we’re really excited about this release because it marks the first time that we have had American Community Survey data. That is the key socioeconomic data that we have historically only had one per decade. We are going to start – (audio break) – data release. So we are certainly excited about that. This type of information provides sort of the critical information that local decision-makers, businesses, and community leaders use in trying to plan what kind of social service programs to implement, and it’s also the type of information that is used by federal officials to distribute about $200 billion back to state and local governments every year.
The power of having data every year rather than data that could be up to 10 and 12 years old is just now starting to be seen, and we certainly look forward to working with you guys as we introduce you not just to the American Community Survey, but the way the Census Bureau is trying to put out more timely and relevant data about the population and economic.
With us today – we are going to have two presenters, the first of which is Doug Hillmer. He is on the American Community Survey office staff and does a lot of the technical work, including developing the tables and providing the data analysis review on it. The second presenter will go through the presentation of data on American FactFinder, of which many of you are familiar with. But certainly he is going to walk you through on how the ’05 data will show up on there, so a little tutorial for American FactFinder users.
Let me start quickly – before I hand it over to Doug to start the presentation on ACS, a couple of housekeeping things. As you’ll see from the media advisory that we did we put out, we have tried to provide you with as much lead time as possible to look at the data. There are nearly 7,000 different areas that we are going to be providing data for on – (audio break) – and we try to give you the maximum amount of time as possible during this embargo period to allow you to sip through it, answer questions to the Census Bureau staff, which we will have on standby to help you weed through a lot of this data, but also to allow you to develop your stories.
There is a lot of information in these two tables, and we know there is going to be a lot of questions. So I hope that the three working days, not including the weekend, but certainly we threw it in there for the weekend for those hardcore users as well. But if there are questions, please feel free to get in touch with us at the Public Information Office and we will try to help you the best as possible.
On the heels – (audio break) – starts on the 10th, which is a Thursday, for public release on Tuesday at 12:01. That means it’s able to be put into the papers on Tuesday. Following that release on the 29th of August, we will be releasing the economic characteristics as part of our annual income poverty and health insurance coverage news conference which we will be holding here at the Census Bureau once again. A media advisory is forthcoming this week on the notification of that and how it will be handled. We do plan on looking at trying to do a webcast again where you could also have a listen-in line to ask questions if you are unable to join us out here in lovely Suitland, and also to admire our new building that we are getting to move into.
On the heels of that, we will have some other characteristics we will be rolling out on a flow basis between October and November. So certainly there is a lot of information that is going to be coming out, and we’re looking forward to working with you to answer those questions on each one of the data releases.
So, barring any questions via e-mail, your phones have been muted just so that there wouldn’t be the environmental background noise that would distract other participants. We will open up those lines as soon as the presentations are done so that you can ask your questions orally if you would like, but in the meantime we do have people on the webinar system that are going to be trying to answer your questions as they are submitted. So if you have a question, you might want to take a look and see what types of questions are being asked. And certainly we know that there is going to be a lot of questions that you will want to just contact us individually.
Once again, we would love to have some feedback from you on how this webinar goes in terms of a briefing tool. We would like to maybe do these in the future, but we would love to have some feedback from you. So any of that you could provide would be greatly appreciated.
At this time I’m going to turn it over to Doug Hillmer with the American Community Survey Office, who is going to start through his presentation. Thank you.
DOUG HILLMER: Thanks, Steve. Good afternoon, everybody; at least it’s afternoon on the East Coast. As Steve said, we are going to do this presentation in two parts. I am going to give you an overview of the American Community Survey annual data products that are coming out. J.R. Wycinsky is going to step you through the details on these products on how to access them.
You should be looking at the slide right now that says “American Community Survey 2005 Data Release.” And the picture you’re seeing is actually a snapshot of our poster. And what it says underneath the American Community Survey is really one of the driving forces behind the creation of this survey. America is changing; its needs are expanding; and more current up-to-date information is needed than can be provided once every 10 years.
So what is the American Community Survey? The ACS annual data updates summations, giving a timely and accurate detailed look at how America is changing. This is information needed more often than once a decade. The ACS provides a new moving picture of change rather than a once-in-a-decade snapshot.
The ACS gives elected officials, businesses, and others best up-to-date information available to make better informed – (audio break). The ACS helps local officials know where they should plan new schools, senior centers. It also helps businesses pinpoint where to open a restaurant or a store. It helps transportation planners know where to add new bus routes or roads.
The survey will collect detailed information on the characteristics of population and housing on an ongoing basis instead of once every 10 years. And the Census Bureau will provide tabulations of these data on an annual basis rather than once each decade.
The other reason that the Census Bureau wanted to begin the American Community Survey is to allow the Census Bureau itself to restructure and greatly simplify the process of census taking itself. This will allow the 2010 Census to focus more on improving the accuracy and reducing the operational risk of the census taking in 2010.
The American Community Survey is a large, continuous survey, and that means that the data is being collected continuously in the field via interviews throughout the year every year. It is the largest survey ever undertaken by the Census Bureau, except for the Decennial Census sample data collections – (audio break) – single and multiyear estimates of characteristics in the population and housing of communities throughout the United States and Puerto Rico.
These characteristics include demographic, social, and economic information about individuals in households. They also include physical and financial characteristics of housing units where these households live.
It is important to emphasize that the ACS is not a population count. The Census Bureau, besides the Decennial Census every 10 years, has an intercensal estimates program, and that is the source for official counts from the Census Bureau. The ACS produces information for small areas and small population subgroups. The Census Bureau does keep all of its information about individual people and households strictly confidential. The Census Bureau only releases data for statistical aggregates, geographic areas, and population categories.
The ACS is part of the Census Bureau’s Decennial Census program. Part of that plan involves collecting detailed information on the characteristics of population housing on an ongoing basis rather than collecting these sample data once every 10 years.
So let’s take a look at how the ACS evolved. The American Community Survey actually began as a test in 1996, and it has been through a total of nine years of testing evaluation. Nineteen-ninety-six through 1999, selected counties were examined and the survey was conducted in those counties to determine how well it worked in counties with different kinds of characteristics.
Starting in 2000, the ACS expanded to a nationwide sample. Over 1200 counties were included in that sample. And from 2000 to 2004, the ACS produced estimates also at very high levels of geography such as the nations and the states. In 2005, full implementation of the American Community Survey began. Since the beginning of 2005, data has been collected every month in each county in the United States, nationwide, in every municipio in Puerto Rico as well.
The first data products from this fully implemented ACS will be available starting on August 15th. Every year thereafter, data products will be released from the ACS based on the interviews conducted through the previous – (audio break).
In 2005, the American Community Survey expanded to a monthly census size of about 250,000 addresses throughout the U.S. and Puerto Rico. That adds up to 3 million addresses in sample each year. At this current time, this sample covers the population living in housing units. The population living in what we call group quarters is not covered at this time. By the way, group quarters is our term to describe people who live – have other living arrangements other than housing units such as – (audio break) – prison, and so forth.
So as I said, the current ACS is limited to population in housing units, which is the overwhelming majority of the population. Single-year estimates will be available for communities of 65,000 or more beginning this summer in August. Data for communities of 20,000 or more will be available as three-year-period estimates starting in 2008. Data for all areas, including those smaller than 20,000 will be available as five-year period estimates starting in 2010.
In that same year, we will be releasing single-year estimates coming to 2009, three-year estimates coming to period 2007 through 2009, as well as the five-year estimates covering the period 2005 to 2009. And then, every year thereafter, we will be updating each of those separate estimates.
This chart gives you some idea of the kinds of geographic areas for which we will be publishing. If we look at that from the top in the left-hand column, top going down, of course we will be publishing our estimates for the nation, for the states, including the District of Columbia and Puerto Rico. The census regions are four of them and the nine census divisions. We will be publishing our estimates to all of the congressional districts. We will also publish our estimates for the rural and urban components of each state, except for the District of Columbia, which has no rural component.
And finally, something that may not be clear at first, we will be publishing our estimates for areas known as PUMS, or public use microdata sample areas. These are areas that the state defined in the year 2000 to the Census Bureau. For each state they defined these sub-state areas, which partition the state. Each of these areas is about 100,000 in population, and we will be publishing for each of them this year. There are over 2,000 of these areas, and there is also over 30 of these areas in – or maybe exactly 30 of these areas in Puerto Rico.
Another way that data users can arrive to data for where there – the minimum population threshold is not met is through the use of components at the state – (audio break) – for things such as inside or outside a metro area. That is people who live inside or outside of a metropolitan area. These components will also be available at the nation level as well as the regions and divisions.
Continuing on, we have a number of other kinds of geographies for which we will be publishing. And this chart shows geographic areas – (audio break) – all of that type of geography. For instance, there are 775 counties, which, using our 2005 population estimates, have a population of 65,000 or more. That is 775 out of 3,141 counties in the United States, or roughly 25 percent. And that is what is true for the other kinds of geographic areas here. Places – which is the Census Bureau term for cities and small towns – we will be publishing 492 places. That is a very tiny percent of the total number, but those are the places which have a population of 65,000 ore more, and the same is true for the other geographic areas you see here, including school districts.
Minor civil divisions – and if that term doesn’t mean much to you, think of it as townships in places like New England, sub-county areas which actually have functioning governments. All in all, there will been over 6,900 geographic areas for which we will publish estimates.
This chart here shows you the release schedule – (audio break). We will release estimates of demographic and social characteristics. Then on August 29th we will release estimates of economic characteristics. In October, we will be releasing estimates of physical and financial characteristics of housing. That includes things such as housing occupancy, tenure, rent, cost, and utilities. In November, we will be releasing select population profiles for a series of race, ethnicity, and ancestry groups. Examples of these include native Hawaiians and other Pacific Islander race groups, people of Mexican origin, and people of Italian ancestry. There are over 200 of these groups that we will publish.
Well, there are a long – I’m going to give you a very brief description of all of the different kinds of data products we have, as many of these J.R. we’ll be describing in more detail. I will also tell you which of these products will be available during the embargo period. Starting with the base table, this is our fundamental product. Over a thousand of these will be published across all four releases. The base tables of course that we will release on August 15th will cover demographic and social characteristics. (Audio break) – embargo site.
Next we had data profiles. These are more summarized products. They are derived from the base tables. The summarized key demographic social economic and housing characteristics. They will also be available on the embargo site. The narrative profiles are actually taken from the data profiles, and they provide a report that is in text form with embedded charts and graphs. The subject tables are somewhat like the data profiles, but they are – each of these addresses a specific topic. There are over 40 of these subject tables.
Selective populations profiles are what I had just described. These are reports of a variety of characteristics for specific population groups. The geographic ranking table – (audio break) – key indicators would be publishing these ranking tables at the state level, and for the District of Columbia. These ranking tables will also be on the embargo site.
We also will have – for those same indicators, we will have state-level phonetic maps, and we have something called geographic comparison tables which compliment the ranking table by showing those same indicators for other geographic areas.
We have some analytic reports as well, which are on the ACS website. And finally we have our public-use microdata sample. These are data files that contain records of a sub-sample of all of the housing units and the people in those housing units who responded to the survey. Those files can be downloaded from the ACS website, and those files will be available at the end of August.
There are some factors to consider in interpreting the data. The three main factors are interview and residence rules, time periods, and reference periods. I’m going to address them one at a time.
What are the interview and residence rules? The other way of saying it, what is the target population that we are trying to describe? Interview and residence rules define the universe of a target population for a survey. These rules therefore identify the units of the people eligible for including in the survey. The sampling frame reflects the choice of the universe, as do the instructions on the forms and in the procedures used by survey interviewers.
In 2004, the universe was the household population of the United States and Puerto Rico. Like the Decennial Census, the ACS interviews the population residing in the U.S. – (audio break) – legal status or citizenship. Starting in 2006, the ACS also included people residing in group quarters, or what we called the – making up the entire residence population, i.e., people living in either housing units or group quarters.
The ACS excludes people residing in housing units only if the ACS residence rules define their current residence to be elsewhere than the sample address. So what are these residence rules?
Basically, the residence rules are driven by the goal of – (audio break) – of all types of communities each year. Given this goal, deciding who should be treated by the survey as a resident of an area and who should not be treated as a resident is critical. We wanted to – we wanted to the ACS to reflect the population that uses community resources and should be considered by community planners. For example, in designing facilities for seniors or assessing the extent of the disabled population or the language needs in an area, we wanted to be sure to represent the population. We may not live there most of the time but stay there longer than a brief vacation.
So here they are in brief. The residence of – actually, it says community here, but the residence rules are really based on the sample address that we interview. So a resident on an address – a person who is considered a resident of an address – if that person lives there year round, or if that person lives there more than two months but not year round, or if that person has, you know, another place to live. If that person is away, but is going to be away a total of less than two months, that person is still considered a resident of that sample address. A person is not a resident of the address if the person lived there two months or less and has another residence, or if the person is away for more than two months.
Finally, the ACS produces what we call period estimates. What do we mean by that? Period estimates describe a period of time. Data are collected over that period of time. In order to estimate the characteristics that are to represent 2005, for example, the ACS uses annual or yearly average estimates rather than choosing one particular point in time during the year. The fact – (audio break) – the population housing units as of a certain point in time, April 1st of the year of the census.
In particular, the ACS produces period estimates to describe people and their characteristics as well as the characteristics of housing during the calendar year 2005. most 2005 ACS estimates can be thought of as representing the average characteristics of an area over that calendar year. Because the interviews take place throughout the year, the population represented by the survey changes from month to month. Therefore the yearly estimates for an area can be seen as an average of 12 monthly populations. This is important if the population of an area changes substantially during the year. The 2005 ACS population is the average population of an area in that calendar year.
Finally, we come to reference periods – to continue reference periods, excuse me. As Jeff explained, the data collection timeframes for the census and ACS are very different. And most of the questions are asking about things as they are right now on the day of the interview or enumeration. Both the ACS and census questionings also tend to use that date as the end point of any reference period. Only income, migration, and agricultural sales do not. In Census 2000, income referred to the 1999 calendar year. Migration referred to five years ago based on April 1st, and agricultural sales referred to in 1999.
Let me illustrate this point concerning characteristics that are not time dependent with one of the questions, the question we have about tenure, which has to do with owning or renting. In 2004, during the demonstration phase, the ACS estimates for renter-occupied units in Alabama was about 492,000. They way that should be interpreted is that that estimate means there are an average of 492,000 renter-occupied units across all 12 months in the year 2004 of ACS interviews. That does not mean – (audio break) – the entire year; it simply means that when we average the responses throughout the year, we estimate that in 2004, the number of renter-occupied units was 492,000.
And finally we come to a question which I imagine is a question in the minds of many of you: What comparisons can be made between these estimates and other estimates? First of all, when we look at previous years of the American community survey, can comparisons be made? Our answer is generally yes, but there are two cautions. Number one, the sample sizes of the survey, the American Community Survey between the years 2000, 2004, were much smaller than the sample size in 2005.
We went from 800,000 samples selected in each of those years – the 2000-to-2004 period – to 3 million selected in 2005. Therefore the sampling variability of the estimates in the 2000-through-2004 period is much greater than the sampling variability of the 2005 – (audio break) – trends from the past may be due to random noise and not real change.
Secondly, there are more geographic areas that will have 2005 ACS data than was possible from the 2004 ACS. In 2005, data were collected in all of the 3,141 U.S. counties, American India, Alaska native tribal areas, as well as Hawaiian homelands. Between 2000 and 2004, data were collected in only 1239 U.S. counties. All of the – and during that period, those counties were really the counties with a population of 250,000 or more. So those are two cautions to bear in mind when thinking about comparing to past years with the American Community Survey.
And then finally, what about comparing back to Census 2000? The primary limitation comparing a 2005 ACS with Census 2000 is that target populations that I mentioned earlier. The Census 2000 data includes a population living in both housing units and group quarter. The 2005 ACS only includes the housing unit population.
In areas where you feel that the contribution from group quarters is limited, it is reasonable to make comparisons to Census 2000. The ACS website includes a link to a set of subject definitions. You can find advice there about making comparisons back to Census 2000 for each topic.
That concludes my part of this talk, and I am going to hand it over to J.R. Wycinsky now, who will – (audio break).
J.R. WYCINSKY: Okay, thanks, Doug. Again, my name is J.R. Wycinsky. I work with the ACS communications area. And I want to run you through the portion of this presentation to actually access the data, and the slides and the screens you’re going to see to actually get into the data when it’s released.
Just a couple of notes before we get into it. There are already some slides that have – (audio break) – right off of our production site. So we have given you the shell, and then we are going to kind of – you’ll see a blurred out data area, so don’t be alarmed by that. Also, just to – it may be obvious, but for some of you who might not be aware, obviously some of these screens are cut off at the bottom. So for many of these tables, there is actually a lot more data below, so just be aware that it’s not just what we show on these little screen shots.
And finally, also, just to begin, you’ll note at the top there, the factfinder.census.gov. That is actually a direct link to get right to this American FactFinder site. And of course you can get to AFF off of the census.gov website. There is a link on your left-hand side when you enter through that portal. But for now, if you go straight into American FactFinder, you can follow down the left-hand, and you’ll see data sets, and then to the – (audio break) – and that is the area you would clink to actually get into the data sets page, which we see here.
And, again, I’m going to run through the various different data products that Doug touched on. And you’ll see right now the default pulls up. This kind of gray middle box in the middle portion of the slide there is the 2005 American Community Survey area with a little blurb that lights up and tells you a little bit about it. If you go on AFF right now, the default would be the 2004 area, but you would get a similar set of products on the right-hand side. So for the purposes of this presentation, you’re going to see what you’ll actually get on August 15th.
So we again we begin here. There is the American Community Survey listed under data sets. That is where we are at. We are going to down one by one through these different products on the right-hand side, again, in this gray box here. We’re going to begin with data profiles, and, again, those summarized key demographics, social, economic, and housing characteristics. Excuse me. Below that, you have got selected population profiles, which are tabular profiles for about 200 different race, Hispanic origin, and ancestry groups.
Below that, you have got ranking tables. Those compare 90 key indicators for the U.S., all states, and D.C. Below that, you have got subject tables. These present over 40 summarized – (audio break) – 2000 quick tables. Below that, we have got detailed tables, which we also refer to as base tables in the ACS area. These are the foundation on which other ACS data projects are built. There are over one thousand of these tables. And again, on AFF, they are called detail tables, but we also refer to them as base tables.
Below that, you’ll see geographic comparison tables, or GCTs for short. These compliment the ranking tables by showing key indicators for other geographic areas. And as an example of a search you can do with this, you can actually pull up all of the counties in the state, or all of the states in the nation. You get it all in one nice neat table. We are going to go through that in a little bit. Below that – well, actually not below that, but as part of the data profiles, when we get into them, you won’t actually see them, but they will – (audio break) – files which are essentially almost like a newspaper article, which have kind of facts dropped into text, a little write-up with charts and drafts and what not. So those will also be available through the data profiles.
And then ultimately we have got thematic maps. Those present the information in the geographic ranking tables on the U.S. – on a U.S. map to share the relationships for these indicators. Now we are going to go through the ones that will be available with the August 15 release. Just to point in specific, the selected population profiles – that is the second link on your right-hand side – those will not be available until October of 2006.
Also, just to note at the top right-hand corner of the page, you’ll see this other resources box. That is a – below that you’ll see a link to the American community main page, and that is a page that has all kinds of information. The data user guide will be linked up there – (audio break) – things to get through the data. We actually have links to some of our prior presentations. There is a lot of good information there. We also have quality measures, and then below that, public use for microdata samples, or PUMS, that is actually how you actually link into our PUMS site to get the extracts of the ACS data.
And now we are going to move on to what we would do to get a data profile. For those of you that are familiar with AFF, this will not be that much different; it will just be new data. Across the top of the page there, you are going to see what we call – or what are called breadcrumb links, which tell you where you’re going, step-by-step approach. You’ll see we started out from the main AFF page; we lengthened the data sets. Now, we are going to pick our geography, and the next will be results.
(Audio break) – so this is the first, being a list method, which is highlighted on this slide. You could also do a name search. You could actually type in an address search using the third tab, or you could pick it out from a map, a reference map. Some of the available geographies in this pull-down menu, you see we have got counties selected, but data for all geographies of 65,000 or greater are available. Some of those types include the nation, region, state, congressional district, school district, public use microdata area, the metropolitan statistical area, or MSA.
So for now we have pulled down the county. We selected the state of Maryland, and we selected Prince George’s County. At this point, if you want to get an idea of what that county looks like, you could actually click on the map-it button, and that will pop up a separate window that will give you a little – kind of an overview of the boundaries of that county or that geography in general.
But for now we are going to stay with PG County and hit “show result.” And again, if you kind of follow on the top, you’ll see the steps we got – we used to get here. So here is the data profile for Prince George’s County, Maryland. Again, these display key estimates of demographics, social, economic, and housing topics. For the August 15th release, only the social and demographic characteristics will be available. Actually, if you wanted to get an idea of what the other two looked like, you could go back to the ’04 data because it will follow a very similar pattern.
This particular profile will display the numerical estimates along with the margin of error for that estimate. You can derive confidence bounds by adding the margin of error to the estimate for an upper bound, and subtracting the margin of error from the estimates for a lower bound. And all of these – (audio break) – confidence level. Some other options you have on this page, as you go across the top right-hand corner, you’ll see you can print or download the table. You can download it into an excel spreadsheet, a text file. There is about three or four other different methods of downloading this table.
But for now we will go to an example of a ranking table. And, again, to actually get here, we skipped over the steps just for this presentation, but to get here, you go back to the data sets page, select the ’05 ACS, and you click on ranking tables. And here are these ranking tables, and there is going to be 47 ranking tables on social and demographic characteristics released for the August 15th date, and here is an example of some age and sex, age, and children, and again, it keeps on going down the page.
Here is an example of a ranking table by state for a median age. And here we have got, again, the percentage estimate and the margin of error. And, again, it’s ranked for a single characteristic, in this case being media age. And it starts out ranking highest to lowest, but you could also manipulate that. You’ll see kind of the middle of the page, it says rank, and a little arrow to the right of that. You can click on that to make it ascending or – (audio break). If the geography is available, high enough, you can go back to the ’04, ’03. You could see it with statistical significance, and you could also view it as a chart, which gives you a better kind of graphical representation of the margins of error, the confidence levels. For the ’05 data, the ranking table will be available at the state level.
Now we’re going to move on to an example of our subject tables. For access to the subject tables you need to specify a geography of interests, just like you did for the data profiles, and these will be released for all available geographies over 65,000 or more in population. These are the before specific topic – (audio break) – and you’ll here Section H, school enrollment, movers between regions, the language spoken at home. And from here we’re going to go to an example of these school enrollments, so it will be just like if you clicked on school enrollment in the middle of the page. Here you’ve got the actual subject table for U.S. school enrollment. And again, you’ve got the data blurred out, but you’ll see you’ve got some estimates, the margin of error. You also have distributions with some of the key universes, and these other – some of the tables display measures such as medians, aggregates where appropriate, and also include tables of education rates for the relevant measures.
From here we’re going to go to an example of our base or detail tables. Again, to get to this step right now you’d go back to datasets, you’d select the ’05 ACS, you pick your geography, and then now you’re getting to the point where you’re picking your table. You can do that by various methods: by subject, by keyword, or just do a show-all table search and then just scroll through and pick out the table you want. In this case we’ve selected six-by-H (ph), the very first table. We’ve clicked on it and then hit the add button to put it into the lower box down below. You can actually make more than one selection in this area, and then if you wanted to edit that, you could remove some of those by clicking on it in this lower box, using the remove button to get that out of there.
So for now we’ll stick with six-by-H. We’ll go ahead and click on “show results.” And just a little note; there’s about – there’s more than 200 of these base tables – (audio break). To give you an example of the base table of six-by-H; we’ll go ahead and click on show result. And just a little note; there’s about – there’s more than 200 of these base tables – (audio break). To give you an example of the base table of six-by-H for the U.S. – and again, we don’t stop at 62 to 64 at the bottom there; it keeps on going down the page – as you’ll see, across the top there you’ve got the title of the table, you’ve got the dataset from where it comes, in this case the ’05 ACS, and again, you have the estimate and the margin of error.
Okay, now we’re going to take a look at some of the available geographies for the geographic comparison tables, or the GCTs. These show key indicators for geographies that are organized alphabetically instead of being ranked. And they serve as kind of a compliment to the ranking tables. And, again, right now you could do – from this example on this particular slide you’ve got nation selected, so you’ve got U.S. by county, so theoretically it would list all the counties in the U.S., but you could also do it by state, and that will be what our example is.
(Audio break) – table we want to see in this geographic comparison table, so we will actually go to this slide here, and you’ll see again – just follow along step by step. We picked our geography and now we’re picking our table. There are 47 geographic comparison tables, which will be released on August 15th, and they cover the same topics as the 47 ranking tables. So again, we’ve selected median age. Go ahead and click on show result, and that will take us to the results, which is median age for counties by state. In this case we have all of the counties in Alabama. So you’ve got the median age and then you’ve got the – so you’ve got the median age estimate and then you have the margin of error as well. Again, from this page you could click on view as a schematic map on the left hand – (audio break) – in an overall map of the state of Alabama. That’s another option to help you get a more graphic sense.
So that kind of wraps up the actual how to get into the data. This is just an example of our newsroom page right off our Census.gov site. You’ll see in the upper right hand corner of this newsroom page the embargoed releases. This is where you all would go in to get access to that data ahead of time. You will need a user name and password, which you can get through our Public Information Office, and there is more information on those contacts from the very last slide, which we’re getting to. Once again, with that embargoed release you will be able to get the base tables, the data profiles and the ranking tables.
And that will wrap up the presentation. There is the contact information. And now I guess we will open it up to questions.
MR. BUCKNER: If you can bear with us, we’re going to have the operator look at the mute off. You can continue submitting questions. Just one moment.
OPERATOR: Please have your pass code and conference leader’s name available. A coordinator will assist you momentarily.
(Pause.)
MR. BUCKNER: Okay, I think the mute has been lifted. Are there some questions for either PIO or the APS presenters about the presentation?
Q: When will the embargoed material be available on the 10th?
MR. BUCKNER: We’re going to shoot for the morning of the 10th, probably around nineish or so. We shoot for nine. Sometimes it happens a little bit afterwards but we’re definitely going to get it up in the morning period.
Q: And if we’re looking at something like New York State or a particular state and we added up all the group quarters of people in colleges, mental institutions, prisons, and then added that on to the ACS count, would it be then fair to combine the 2005 ACS with the 2000 census in terms of total population?
MR. HILLMER: This is Doug Hillmer. Actually we missed part of your question, but I think – because it just timed out a little bit here – I think you’re talking about population count.
Q: Right.
MR. HILLMER: One of the things that I mentioned – maybe I didn’t dwell on it enough – is that if you’re looking for population counts, it really – we suggest you go to our intercensal estimates program and you – right off of our census main page there is a link to the population estimates program.
Q: But there will be a count in the ACS, right?
MR. HILLMER: There will be counts in the ACS –
Q: So should we – I mean, how should we treat those in terms of saying, well, this is the count of the ACS 2005 and it looks like it went up or down from 2000, but that doesn’t matter, or does matter or –
MR. HILLMER: Well, I’ll let Lisa Blumerman –
LISA BLUMERMAN: Sure, hi. This is Lisa Blumerman. I think it depends on the question you’re trying to answer. I think when you’re looking at data you need to look at – you need to use data with caution, and if you’re trying to compare the 2005 ACS with Census 2000, you can certainly do that but you need to be aware of the limitations. And as Doug was saying, one of the big limitations between making comparisons from 2005 ACS to Census 2000 is the fact that the ACS data do not contain reported population.
Q: Right,
MS. BLUMERMAN: Now, for some universes, that won’t have much of an effect, but for some characteristics, that may. So you need to look at that and analyze that and ensure that it is a comparable statistic that you’re looking at.
Q: Great. I don’t want to get too specific, obviously, but my question is obviously we’ve had a lot of questions about population counts in New York City and they’ve been adjusted occasionally. If I took the 2005 ACS and added on all the people I know are in group quarters, and then found a big difference between that and the 2000 census, how significant would that be?
MS. BLUMERMAN: I don’t think that we would recommend that you combine the two data products together. I think that if you’re – again, if you’re trying to look at counts and if you’re interested in the total population for parts of New York City, for example, the population estimates program does release those. The estimates that are coming out with the ACS are sample estimates; they’re not population estimates, so they’re subject to all of the sample characteristics and all of the variants that a sample in a survey would have. So they’re representing a different universe.
Q: Okay, and who is this speaking? I'm sorry.
MS. BLUMERMAN: This is Lisa Blumerman.
Q: Great. Thanks.
Q: I’d like to piggyback on that question.
MR. BUCKNER: Can I have your name and –
Q: It’s Lori Weisberg, San Diego Union Tribune.
MR. BUCKNER: Hey, Lori.
Q: I don’t mean to belabor this issue of comparing to the 2000 census, but for me the census is all about tracking change in our country, and in an area like San Diego County where you do have a large military population, I’m starting to wonder whether it’s worthwhile to compare to the 2000 census if you’ve excluded group quarters. I’m really struggling with what to do about this.
MS. BLUMERMAN: Well, the American Community Survey, starting in 2006, did start again collecting group-quarters information. So the data that we’re releasing this August, while it won’t have it, we hope that future data will.
Now, in terms of making comparisons with 2000, I think there is a lot of utility in using the ACS data in order to measure change. I think you do need to use caution and you need to be aware of the characteristics. However, if we didn’t have the ACS, you wouldn’t have any of these detailed information until 2012, so this gives you a picture of what things are like now that you wouldn’t have had otherwise.
Q: Hi, this is John Cammon (ph) of the Arizona Republic in Phoenix. I’m wondering, in your survey of 250,000 households – (inaudible) – what is the response rate, and do you continue sampling until you get 250,000?
MS. BLUMERMAN: Sure. Let me talk with you a little bit about the operations for data collection for the American Community Survey. As we told you, we do have a sample of about 250,000 addresses per month. We actually collect data from three different modes. We collect – our first attempt with those addresses is by mail, so we mail out about 250,000 addresses, and that’s about a one-month window. So every sample, every group of 250,000 addresses that actually are eligible – (audio break) – over three months.
If at the end of that month we haven’t received a completed response by mail, we’ll then move on to a telephone operation where we’ll send about four a week for addresses that we’ve been able to identify, phone numbers, trying to get completed questionnaires from them over the telephone.
At the end of that four-week period – (audio break) – have not received completed – (audio break) – we will do – (audio break).
MR. : I’m sorry, this is all broken up. Someone is talking.
MR. BUCKNER: If everybody could just mute their line during the question if you’re having any conversations, that would be helpful. (Cross talk.) Okay, here we go. I’ll let you get back to the explanation – (cross talk).
MR. : Yeah, this is not working very well.
MR. BUCKNER: Okay, let me jump – this is Stephen Buckner at PIO. Any questions that we’re not able to answer because of some of the audio issues that have been going on – and we are trying to work those out – any issues or questions that you have, please contact the Public Information Office – (cross talk).
MS. BLUMERMAN: Who is this guy?
MR. BUCKNER: Hello? Hello? Sidebar conversations we need to hold – (cross talk).
MR. : There is a lesson here.
MR. BUCKNER: Yes, I think there is. Anyway, like I was saying, if there are any questions that you have following the – (cross talk).
MR. : I think he’s with the Census Bureau, isn’t he?
MS. BLUMERMAN: No. (Cross talk.) Hey, San Diego, hush.
MS. WEISBERG: It’s not San Diego.
(Cross talk.)
MR. BUCKNER: Hey, is that – Arizona or New York or whoever that was – (cross talk).
Q: No, no, no, it’s not from Arizona. I’m just wondering if we can to the chase and say how many of the 250,000 do you respond to? I mean –
MR. BUCKNER: We get about a 97 – we get about a 97 percent weighted response rate. (Cross talk.) There we go. All right. We get about a 97 percent weighted response weight.
Q: What is weighted response?
MR. BUCKNER: Very simply, it’s the total response rate once we get all the collections by mail. Those that we get by telephone operation, if we still haven’t reached the respondent through the mail or the telephone, we will do a subsample, a one-in-three follow up visit in person. We don’t go to every single household like we do in the census.
Q: Okay.
Q: And how much is that rate again? I’m sorry.
MR. BUCKNER: Ninety-seven percent response rate.
Q: Thank you.
Q: So, for instance, will we be able to tell how many people responded from a particular community, and what the number is that you used to make your estimates?
MR. BUCKNER: We do have some numbers on the ACS website that does at least tell the number of addresses that were in sample per county.
Q: But not per community? I mean, how do we know that you had enough responses from a community of 65,000 to be accurate? I mean, wouldn’t the numbers be scattered throughout the state?
MS. BLUMERMAN: We’re collecting data from every county nationwide, and it’s designed to allow us to produce reliable estimates for single-year estimates of populations larger than 65,000.
Q: And so the margin of error is printed per each one of those –
MS. BLUMERMAN: That’s right; the margin of error is available for every estimate that we publish.
Q: Okay, great.
Q: Is that at a 95-percent confidence level?
MS. BLUMERMAN: It’s a 90 percent.
Q: Ninety percent.
MR. : Ninety percent, yes.
Q: What was the response rate from the 2000 long form?
MR. BUCKNER: The overall response rate for the census overall was 67 percent. We halted a three-decade decline and the response rate was 65 percent in 1990; in 2000 it was 67 percent. The long form response rate I think is on our website – I’ll have to double check – but because it is the census we follow up with every household and make every single attempt we can to reach that individual respondent.
Q: What’s a comparable number – like in our previous stories of the 2000 census we would say like, you know, one in 250 households responded or were interviewed.
MS. BLUMERMAN: One in 480 households are eligible to receive the ACS sample – the ACS questionnaire –
Q: Okay – on a yearly basis.
MS. BLUMERMAN: An important point to remember with that is the ACS is designed to be accumulated over time. So what we’ll do is we’ll accumulate three years to give you data for areas larger than 20,000 and five years for all areas, so that one in 480 is really for those annual single-year estimates.
Q: And so writing about this, the base period is just – these are the characteristics for 2005, all of 2005?
MS. BLUMERMAN: That’s correct.
MR. BUCKNER: That’s right.
Q: What’s the Census Bureau’s policy on discussing this information with other people? We’re trying to do interviews.
MR. BUCKNER: Hi, this is Stephen with PIO again. We, as part of the embargo policy, and what you sign up for is that you’re going to maintain the integrity of the embargo, which means you’re not going to run it in any form of our outlets via the Web, the Internet, Web pages that you have, the newspapers in print, on radio or on television in any form or fashion. If you need to work with local experts that know data or know particular trends, we certainly allow that as long as they also maintain the integrity of the embargo as well.
So as you follow up with your local data experts, which we certainly urge you to do, we do have a data dissemination network which are partners that we have that are very familiar with this data at both the state and local levels that you can certainly contact, or your own sources that you have, but they need to understand that any of the data that you share with them needs to be held in confidence until the public release. They can’t do anything with it.
Q: Okay, thank you.
Q: Will you guys know before August 10th exactly which tables in the SS3 (ph) – I guess they’re history tables – that we can compare this stuff to?
MR. HILLMER: Well, I don’t think we’ll be able to provide you an exact list, but one thing you can use as a guide is – if you go to ACS-3 you will see many tables that deal with households and household populations. Some of our tables are dealing with the structure of the household or family. There are tables on language spoken, and tables covering – that are in a household.
So those are the kinds of things where you will be – those tables themselves are restricted to the housing unit population and – (audio break) – covering demographic and social characteristics. So that includes some of the things that are covered in those tables.
We will attempt to provide more information on that in terms of the exact ACS tables, but I can’t promise before the embargo that we will have that.
(Cross talk.)
Q: It would just be nice to be able to download the comparison tables ahead of time.
Q: Can you clarify one thing please? There are four sets of information that are going to be available in August. That is age and sex, school enrollment, movement between regions, and language spoken at the home. Those are all four things?
MR. HILLMER: Yeah, those four subject tables that J.R. showed –
Q: Right.
MR. HILLMER: – they were also available on August 15th.
Q: Okay, good. Thanks.
(Cross talk.)
MR. BUCKNER: There’s many more, but those four will be available on August 15th.
Q: Yeah, and the other tables, say the housing data tables you were just talking about, those will be available or will not be available on August 15th?
MR. HILLMER: We will have – just to review – in fact, maybe J.R. can bring up the thing that covers – the page that covers all of our products. For each of our separate products, for instance space tables, we will have space tables that are covering the demographic – (audio break) – characteristics. We will have the first two data profile tables – the first one is on demographic characteristics, the second on social characteristics. We will have those four subject tables, as we’ve just talked about. We will have ranking tables covering, I don’t know, 40-some topics. Those are indicators that are in the area of demographic and social characteristics. And by the way, you will have that information available to you very soon. This is something that we should have mentioned and haven’t mentioned yet.
Off of the ACS website, which you can get to easily by going to Census main – the Census main page and then clicking on the American Community Survey. Later this week we should have an updated version of our guide to our data products. I’ll just tell you it’s there right now. It’s in the upper right hand corner of the ACS website. However, it’s not up to date yet. When data is made available, by the end of this week, you will have, for each of those kinds of products – base tables, profiles, subject tables and so forth – you will have the detailed list of exactly which topics are covered in the August 15th release, and then the same will hold true for the other releases as well.
MR. WICINSKY: Let me clarify one thing. The demographic characteristics covered in this release can be found on the media advisory that we sent out about a week ago, but they cover age, sex, race, Hispanic or Latino origins, relationship and households by type, the social characteristics. They cover school enrollment, educational attainment, marital status, fertility, grandparents as caregivers, veteran status, disability status, residence one year ago, place of birth, U.S. citizenship status, year of entry, world’s region of birth, language spoken at home, and ancestry.
Q: Back to the methodology of the survey, are the 250,000 questionnaires sent out proportional to what you believe the population is in each one of the counties of the country? I mean, how do we know that they don’t all go to Alabama once – (chuckles)? I mean, most all of them go to one area and then one question – I mean, one questionnaire apiece to all of the other counties.
MR. HILLMER: Basically, the way the Census Bureau conducts the survey is we start with a master list of addresses. Those addresses – we have mailing addresses, we know where they are in the country, and we are shooting for a roughly 2 ½ percent sample rate, and so a systematic sampling procedure, a form of a probability sample, is applied, and that guarantees that there will be a good proportional spread in the sample. So in a very short fashion, that’s the way we make sure that they don’t all end up just in one part of the country.
Q: (Off mike.)
MR. WYCINSKY: The question that was just asked about whether or not those subjects are still coming on August 10th, absolutely. If you go to the American Fact Center right now and look at the 2004 data and look at the demographic characteristics or social characteristics from your area that you’re located in, those are the same characteristics that are coming out on embargo beginning on the 10th.
Q: Going back to the comparison of the 2005 – (inaudible) – estimates, I know that when the 2000 census came out and we were comparing it back to the 1990 estimates we had to do calculations for the margins of error. Are we going to have to do it for these too? Because there was a discussion on Census L (ph) where they said that the census was going to do that for us.
MR. HILLMER: I’m not aware of that.
MR. WICINSKY: Yeah, no, the census has not agreed to do any kind of calculations of margin of errors between 2000 and 2005.
Q: Okay, so we’ll have to do it ourselves?
MR. HILLMER: Well, you have the ACS margins of error –
Q: Right, but if we want to calculate the margin of error for the difference between the two, we’re going to have to do that ourselves, right?
MR. HILLMER: You would have to calculate the standard error or the margin or error on the Census 2000 side, that’s correct.
MR. BUCKNER: Okay, if there aren’t any more questions I’ll try to summarize just quickly. Again, the data will be coming out of embargo on the 10th of August. It will be for all demographic and social characteristics from the American Community Survey. The economic characteristics, which include income and poverty but also things like – (inaudible) – to work will be coming out on the 29th of August. Those will not be on an embargo; those will be at a news conference here at the Census Bureau on the 29th of August.
We have not made up a determination on the housing characteristics and the other selected population profiles, whether or not we’ll embargo those, but we certainly will look at that as we get a little closer to it.
Q: Stephen, I have a really stupid question.
MR. BUCKNER: Okay.
Q: There was something earlier – I guess that somebody had posted a question about a table shell. What exactly is that?
MR. HILLMER: A table shell is basically the table without any data in it. It shows you the layout of the table – title, descriptions for each line, and so forth, and right now on that same website I was talking about called the Guide to the Data Products, we do have a file – it’s a fairly big file, but you have a file that contains all of the table shells for the 2005 ACS base tables.
So that’s one place to start. I will caution you; that file is not limited to the ones released on August 15th.
Q: So that will be all four releases?
MR. HILLMER: That would cover all four releases, correct. Later this week something nicer will be out, much more, I think user friendly. But the table shell is simply – it’s from the spreadsheet and it simply shows you the way the table will appear.
Q: Incidentally, if it means anything, I would urge you to release the later releases on embargo. It really helps.
MR. BUCKNER: Yeah, we will certainly take a look at that for the housing characteristics. Our limitations with the economic characteristics are that the income and poverty figures in the ACS are contained within those tables, and we do those for immediate release because of the sensitive nature of those characteristics, as combined with the current population survey’s annual estimates of poverty for the nation.
So we’re a little bit limited and hand-tied in terms of what we can do for the economic – (audio break) – characteristics. More than likely we’ll be doing that as well.
Q: I think we can all concur that we support the idea of an embargo for those other two releases coming up too: housing and selected populations.
MR. BUCKNER: Okay, thank you.
Q: Thanks.
MR. BUCKNER: There was a question on the Web that indicated where a questionnaire was. Let me also point out – I think I failed to mention – there is an online media toolkit that is also on the American Community Survey page that lifts out a lot of the facts and overviews of the survey. It talks about the rates – the one in 480 that Lisa had mentioned and other –
(Audio break.)
Q: – annual estimate, or is that going to be a three-year estimate?
MR. HILLMER: That will be a three-year estimate. So for instance, in 2008, the estimate for the – well, for every area of 20,000 population or greater will be based on samples collected during 2005 to 2007. And the same goes for the five-year estimates. The five-year estimate, the first one coming out in 2010, that will be based on samples collected during the period 2005 through 2009.
MR. WICINSKY: And just sort of in lay terms, the way I help try to explain the whole three-year and five-year averages, the lower levels of geography that you go to try to produce data, you need more responses to have better estimates, and that’s why you need those three- and five-year averages because you have more responses from that geography.
So that’s just sort of in lay terms. You need those responses to really get a much better estimate. Without it, the variability would be huge.
Q: But it does not pinpoint a time – you know, a single time. It’s a little mushy.
MR. WICINSKY: It’s a moving average. It is a moving average, yes.
Q: Is there a master list of all the places of 65,000 or more? Several of my cities are wafflers. I mean, they’re real close to 65 but I don’t know if you have them at 65,000.
MR. HILLMER: Oh that same website coming up at the end of this week you’ll have a master list of all the geographic areas we’re publishing, which areas we have at having a population of 65,000 or greater. For the cities and the counties, we are using the numbers that that population estimates program provides, but that list will be outside the end of this week.
MR. WICINSKY: And just to look at one of the questions that was just posed on the reference point for the estimates, it is the 2005 population estimates that are being used, but the 65,000 or more population threshold that we use to publish the data, we use that for the – that’s the total population. So there may be a few geographies that the household population is slightly below 65,000 or more. So if you start to say, wait a minute; the household population is slightly below 65,000, that’s the influence of group quarters and not the total population, which is the cutoff (?).
Q: Can you let us know what format the data will be in the embargo off the FTP site? Is it in Excel?
MR. HILLMER: Yes, it will be; it will be an Excel spreadsheet, and we will be providing information as soon as we can about the actual organization of the data at the FTP site.
Q: Could you help those of us who got to you just with a phone number and didn’t see anything you posted; have no idea how to email you, have no phone number to call you back on and do not have the media advisory that you keep referring to?
MR. BUCKNER: Okay, the media advisory is on the Census Bureau homepage under – (audio break) – and it talks about the ACS data being released in August. You can contact the Public Information Office. We will be happy to walk you through everything.
Q: What’s your phone number?
MR. BUCKNER: 301-763-3691.
Q: Thank you.
MR. BUCKNER: And if anybody is on the line that is not on our email listserv and/or embargo access, you can also fill out something online to submit to get access to the embargo site. So certainly take a look and see whether or not you have your embargo access available.
Q: I have a quick question. There is something on the Web here where they answered a question, saying that the ACS is a one-in-40 sample? So is it not a one-in-480 household sample?
MR. BUCKNER: One-in-40 is over the course of the year whereas one-in-480 is over the course of a month.
Q: Oh.
Q: So when we’re looking at the 2005, we should say one-in-40, right?
MR. BUCKNER: That’s correct.
MR. HILLMER: Yes.
Q: Okay.
MR. HILLMER: Lisa was talking about the monthly sample.
Q: Okay. So one-in-40 households get the ACS a year.
MR. WICINSKY: That is correct.
Q: Okay.
MR. WICINSKY: But just coincidentally, when we get questions about what is the American Community Survey, from respondents or from other – (audio break) – is getting it. Well, they’re not. It’s not like the census where everybody gets a questionnaire and we’re trying to reach everybody. So often they’ll ask their neighbor, hey, did you get this? They’ll say no, and then they start asking the questions. But despite that, we have a very good response rate because we have really good field staff that works on our ongoing surveys that’s able to talk to them and walk them through it.
MR. BUCKNER: Okay, if there are no other questions, we’re going to end it. Certainly please feel free to give us a call at our Public Information Office and we will put you in touch with the right people in terms of getting some answers to questions that were either not being able to be heard because of the audio, and we do apologize to that, but, again, we’re trying to work through the process here of maybe doing these webinars on a more frequent basis for more of our releases in terms of briefings before the data coming out, so we certainly would like to have some feedback, and you can send that to our email address, either mine directly if you have it – it should be on the Web – or you can send it to our general PIO mailbox, which is also on the Web on most of our news releases.
Q: Great. Stephen and everyone, thank you very much.
Q: Thank you.
MR. BUCKNER: Thank you, everyone. Have a wonderful day.
(END)
Source: U.S. Census Bureau
Public Information Office
(301) 763-3030
Last Revised:
August 14, 2006 at 02:30:25 PM
