Press Briefing -- April 25, 2000
|MR. STEVE JOST: This is another one of our
regular census operational press briefings where we're providing you and
the public with a status of all census operations. Today we're going to
focus on the largest of all those operations, known as Non-Response
Follow-Up, and try and provide you in as much detail, as time will permit,
how those operations function. I just want to remind you if you have not
been part of our operational press briefings before, we have reporters on
the phone from around the country. Once the Director finishes with some
opening remarks we'll turn the questions over to you and we'll alternate
from the room to the phones. All we ask is that you identify yourself and
your affiliation before your question. We have staff with microphones so
that we can pick you up on audio, as well as the director.
With that, I give you Dr. Kenneth Prewitt.
DR. KENNETH PREWITT: Thank you, Steve.
It is, of course, a cliche, but nevertheless true, that the census is the
nation's largest mobilization in peacetime. As you know, we concluded the
first key operation, the mail-back phase, successfully, I might say, but
the response rate was, as we announced earlier, better than we had
anticipated. Indeed, I can report to you today that we will post tonight
at 6 o'clock a final tally for both the nation and all 39,000
jurisdictions. That tally will show 65 percent, which I indicated last
week we were certain that we would reach. It does show some movement from
last week. That is, now 17 percent of our communities around the country,
or more than 6,600, have bettered their 1990 performance by 5 percent or
more. This is really a dramatic improvement for those communities.
However, as we have feared and announced last week, the response rates for the long and the short forms continue to show a much greater differential than in 1990. In 1990, as you recall, it was 6 percent, and today it is twice that rate, or 12 percent. I won't dwell on the particulars of the response rate phenomena. I can answer your questions, if you wish, during Q&A, and, of course, direct you to our Web site, where there's an icon on the Census 2000 initial response rate.
But now, of course, we enter our most demanding phase. Starting
tomorrow we need to visit every household from which no census form was
returned or cleared through our data capture centers. That's
approximately 42 million addresses. We need to identify them, we need to
get information on each of them, and do that in 10 weeks.
Who is going to do this job? A few words
about our employment pool. Remembering that as recently as six months ago
there were many skeptics who said that the Census Bureau would never get
enough people to do the task of Non-Response Follow-Up. We recruited
aggressively. More than 5,000 print ads, over 160,000 radio spots, 25,000
bus posters, over 150 million flyers and brochures. The recruitment
strategy worked. We generated 6 million calls to our toll free jobs line
and over 10 million hits on our Internet job site. The bottom line? We
have 3.4 million applicants, which is well over our goal, 113 percent of
our goal, and from that applicant pool we identified 2.59 million
qualified applications, again well over our target, 108 percent of our
goal. Or, put differently, this is approximately a 5 to 1 ratio for
reaching our actual hiring goal.
As of today, we have 492,245 people on the
payroll. They are being trained at approximately 40,000 sites around the
country, and this, of course, represents the most qualified selection from
the larger pool of 2.6 million.
Now a word or two about this workforce.
It is what we call front-loaded. That is, it is approximately twice the
minimum needed to do the job. This is our way of protecting against
expected attrition. Indeed, we want to conclude our Non-Response
Follow-Up phase with about a quarter million people on the job. So, we
have front-loaded it roughly 2 to 1 to make certain that as we have
encountered the expected attrition, we nevertheless will have an adequate
number to complete the task. Of course, if we do not have attrition, it
simply means that we get the job finished sooner.
As a further redundancy, we have already
identified 53,000 for replacement training. That is, even if we have as
many as 10 percent of our additional hires who do not complete training --
and of course that happens in an operation like this -- we already have
replacements in place.
Now across this, as far as I'm concerned, quite spectacular recruitment success at the national level, there are, of course, a few spots where there are weak areas. In three of our 520 local census offices -- that is, about a half a percent, there are specific training sites where fewer than twice the minimum did arrive for training. For example, let us say we have a training site where we needed a minimum of 15 people but we wanted 30. That is, the front-loaded number was twice the minimum. And we only got, say, 23. This is a hypothetical case. We're working on those cases, but I say these are specific sites in fewer than -- well, there are only three LCO's out of 520.
But the important fact is, every one of our training sites has been
started, or will be started if it's on Thursday as against Monday -- we
started in two phases -- every one of the training sites is starting with
an excess above the minimum required. That is, none of our workload is
not going to go forward with the requisite number of census-takers.
Obviously, the basic task of these
census-takers is to convince someone in nonresponding households, those
who could not be bothered to mail a form back, to now cooperate. And this
really is the fundamental message to the American people: if, as of today,
you've not mailed in your form, we ask you then to please cooperate with
the census-taker. These are not professional survey-takers. They're
certainly not sales or marketing people. They are part-time workers who
accepted the challenge to do this very difficult job for America. They
are committed to doing it well and to doing it right.
I'm going to return in a moment to the
Non-Response Follow-Up operation and how it works, but first let me give
you a very quick update on other field operations. We have completed
operations with our questionnaire assistance service in six languages, and
our thousands of local questionnaire assistance centers and other
locations where "Be Counted" forms were available. We made available
about 16 million "Be Counted" forms, and approximately 650,000 have now
Also, as you know, we set up a process by
which language forms could be requested and about 2.2 million language
forms were requested, and of those more than 80 percent have now been
returned. It's a response rate well above the national average of 65
percent, which is to suggest, of course, that our language program did
work, as we had hoped that it would.
At the end of March we completed our
special place enumeration that was 8,600 emergency shelters, hotels,
motels used as shelters, 3,500 soup kitchens, 300 mobile food vans, 5,800
targeted shelter outdoor locations, transient night operations, 13,000
locations, racetracks, carnivals, marinas and so forth. All of those
operations have now been completed on schedule, and we've concluded our
enumeration of 240 villages in remote Alaska. We've completed military
enumeration of uniformed personnel in 500 installations overseas, and we
still have underway some other special operations like List Enumerate. I
won't go into that in detail, but that's about 375,000 housing units, and
also a process we call Update/Enumerate -- American Indian reservation
areas, resort areas with high rates of seasonal housing. And that's about
930,000 housing units. And group quarters, which we start soon, 11,000
enumerators will go to about 200,000 college dorms, hospitals, nursing
homes and so forth.
All of those operations -- and there are,
of course, a number of them that I've just listed -- all of those
operations are either completed, if they were scheduled to be completed by
now, or they're under way and, as of now, operating on schedule without
any major difficulties.
Now let me return quickly to Non-Response Follow-Up. Just to let you know how it works operationally, because obviously over the next 10 weeks it will be the story that many of you will want to be following. Bear in mind that the census universe that we start the census with is 120 million addresses in the country. We now have shifted to the second universe, which is the universe of nonrespondents, as I say, about 42 million of those addresses. So that's where our focus is now. Whereas a month ago, of course, it was on getting mail back from 120 (million) addresses, we now have about 65 percent of those that have come in, and so we now talk about a new universe that is the universe of nonrespondent households.
For this universe we must generate address files and geographic files
for each one of our local offices, and load these files onto our
databases. From these then the local offices print file assignment
directories, address listings and labels for questionnaires. In each
local office we calculate the assignment based upon the nonresponse rate,
and also specific characteristics of the blocks, including block sizes and
shapes. Larger distances take more travel time. Therefore, you do your
We then carry that workload down to the
next level of geography, first, to the level of what we call field
operation supervisor districts, and then from that, of course, down to our
crew leader districts, which is the smallest unit that we operate with.
That's the 40,000, where the actual door-to-door operations take place.
For each of these operations we produce census field maps with boundaries
drawn for field operations to crew leader levels and, of course, we do
different maps differently depending on whether it's city-style or rural
Here now is an issue that I want to explain
as clearly as I can. We actually have to do that. That is, we have to
create a non-response file or universe in order to make all of these
assignments, to get the maps, to get the kits in all of the right places.
We started that work on about April 18th. Of course, since we started
training yesterday on the 25th, we had to have all of that work finished
by then. Which means forms are still coming in, even after we have cut
for our nonresponse universe.
We do the best we can at striking late
forms that come in from that nonresponse universe, but you cannot find all
of them. You can find all of them but they're still coming in. Which is
to say, we're still getting forms coming in today, and we're also getting
what we call non-ID forms, which are "Be Counted" forms, and it's
impossible instantaneously to make certain we know what household that "Be
Counted" form came back in. So there's necessarily some overlap between
the nonresponse universe and the responding households.
That universe is hard for us to calculate,
but out of the 42 million, that could easily be a million or 1.5 million
addresses, where we will actually get a form but we don't know we've got a
form when we have to do this work, which means we nevertheless have to
send somebody out to that address, knock on the door, and ask for a form.
That, of course, will irritate some members of the American public,
understandably. They'll say, "I mailed it in and why in the world, if you
know that, why are you bothering me again?" and we'll try to explain as
best we can.
Some of you will have already seen because
you visited a training site, we are trying to get roughly 500,000 people
to understand these kinds of subtleties in just a few days of training,
but nevertheless, that will be one of the things that we encounter. Many
of you will pick up that story and will understandably be saying, well,
why in the world couldn't they get it right? I'm just trying to explain
it's not that we could not get it right, it's just that at a certain point
you have to stop and get this other operation on track.
Obviously, address lists must be customized
to fit target areas in which special enumerations are required, such as
"blitz enumeration" or "paired enumeration." Blitz enumeration is when we
actually send a crew of enumerators into an area, let us say a large
apartment building in a hard-to-enumerate area, because we simply think
that's the most effective way to do it, send a lot of enumerators in on
one day and try to blitz the entire complex. Paired enumeration is a bit
different, as we send out enumerators, obviously, in pairs, and that's for
safety reasons where we're sending them in to high crime areas.
All these addresses are loaded onto our
automated operation control system, and this system tracks every census
household ID in the Non-Response Follow-Up universe. All enumerator
assignments and all cases selected for the quality control re-interview
program, we obviously have a quality control process. We're selecting a
number of cases where we go back for every enumerator, for all 500,000, we
go back out and check some of their work to make certain that we are
dealing with honest employees who are doing the job that they are saying
they are doing. We only know that by going out and doing a quality check.
We then produce address registers for each
enumerator assignment area, and identify whether it's a short form or a
long form. And each register contains about 40 cases. That's each
enumerator initially is sent out in the field on about 40 cases. They, of
course, carry both the forms, but also the labels for each address. That
is, our Non-Response Follow-Up universe now has to have an address label
for each of those addresses. The right enumerator has to have the right
label forms so that when they actually get it and stand on the doorway,
they put it onto the forms to be mailed back in -- or not mailed back in
-- taken to the LCO and properly handled.
The labeled questionnaires are then bundled
for each enumerator assignment area. They are handed over to crew
leaders. The crew leader checks them, makes sure that it does meet all of
our quality standards, and then they come back into our local census
offices. Obviously the enumerators now are what the whole enterprise
rests upon. The quality of the work they do, their flexibility, their
ingenuity, their determination. We are very pleased with the quality of
the workforce that we have for this task. Nevertheless, we know it's
going to be a difficult, challenging, and even at times a stressful task.
One of the things that we've concentrated
on, I just mention it quickly, and this is the possibility that the census
enumerators, that there will be phony census enumerators who use the
census as an opportunity to get into homes inappropriately, illegally.
You should know that an individual is phony if they don't carry a census
ID card printed with their name, they don't carry this tote bag. This is
the census enumerator bag. If they don't have this on them then you
simply say, this is not a census-taker, if they don't have the tote bag.
If they also don't have the actual document itself with questionnaires.
And they should be providing -- to everyone when they knock on the door a
form that shows the phone number of the local census office, and any
resident who is concerned will then go back into their home, call that
number, and make certain that this is a census-taker.
So, we think that this is a number of
different ways in which to try to reassure the American public that they
will have a way to determine whether someone is a census-taker. And most
importantly, under no circumstances is a census-taker to ask to come into
the home. Obviously, if someone wants to invite them into their home,
fine, but if a census-taker asks to come in, that is a signal, an alert to
the resident that this could well be not a census-taker but someone trying
to scam the census.
So, we think we have five or six different
ways. You would have to fabricate the material, you would have to get a
fake badge, you would have to get a fake tote bag, you would have to get
fake papers, you would have to carry a phone number and make sure that
somebody back there would answer that phone and say yes, you are. And
then you would have to make the extra mistake of waiting and hoping they
invite you in. That's the only way that someone who's not a census-taker
could actually use the census operation to get in. Those, we believe, are
enough layers of protection that we hope we've reduced to the absolute
minimum any misuse of the census experience by members -- by criminals who
simply want to use this as a way to get into homes.
If no one is at home, then the census-taker
on the first visit, the census-taker then leaves a notice of visit form,
notes the time of day when he or she would like to make a follow-up visit.
If it's obviously occupied housing, then they should come back at least
three times and try to make phone calls at least three times, for a
maximum of six attempts. At that time the census-taker can interview a
neighbor, a building manager or some other knowledgeable person and get
the best possible information they can.
If an enumerator is unable to complete an
interview because of a refusal, the reasons given are documented on a
form, they're returned to the crew leader, who in most instances will
assign the case to an even more experienced or expert enumerator for yet
another attempt. That is, we make an attempt, of course, to convert
refusals into cooperation. If all else fails then we do try to get at
least basic information from a neighbor or building manager.
If we can determine that the house is
vacant, then it will be deleted from our nonresponse universe, if it's
been demolished or burnt out or obviously for sale or vacant.
Enumerators are not actively looking for missing housing units, though if
they discover one they will record the address, information, and if
possible complete a blank enumerator questionnaire that is unlabeled, or
non-ID enumerator questionnaire.
Every day census-takers return their
completed work in individual transmittal envelopes to their crew leader,
and every day their crew leader delivers this work to the local census
office, and there the questionnaires are reviewed in our assignment
control operation to make sure the quality standards are met. If they
fail that test then they are sent back to the enumerator for resolution.
Starting May 1st, questionnaires which pass the quality review are routed
for automated check-in, where they are properly ID-ed in their local
census office. They are then labeled with the address bar with the
enumerator's payroll materials and so forth. Status reports reaching the
LCO and each field operation crew leader are prepared daily again. And
also daily, each LCO will box all questionnaires checked out by operations
control, give a bar-code shipping label, and ship it Fed Ex back to our
central processing offices, and we'll start processing these
questionnaires as early as May 4th.
Throughout this Non-Response Follow-Up period, we will be juggling a variety of difficult kinds of cases. Census Day households with the usual home elsewhere, an apartment or trailer park address mix-ups, partial interview data, non-English-speaking respondent households, those that we selected randomly for quality control, there's a large number of units, will get special treatment because that's what has to happen in a census.
Finally, of course every census success is measured by what the people
do. That is, do the households cooperate? It's very hard for us sitting
here today to tell you how many households simply will not cooperate. We
are about to learn that, and will be reporting that back as we have the
Finally, I'd like to say that I know for
many of you good news is not easily written up as news, and I want to
apologize that I can only give you good news today. Just the facts.
Just the facts. We are on schedule. We have hired the people we need to
hire. We are doing Non-Response Follow-Up, as we have said that we will
do it. That doesn't mean we won't have some difficulty tomorrow morning
that we can then reconvene you and let you know about, but as of today the
census is where we would hope it would be. We are, of course, extremely
pleased. We start on this Non-Response Follow-Up period from the platform
of 65 percent response rate rather than 61 or a lower response rate.
The only bad news I could come up with to
read to you today comes from an issue of Business Week magazine, and I'd
like to read you a few quotes from that Business Week magazine. "This
week, in advance of the scheduled starting date, the census was bubbling
hot in the news. The census director had his hands full oiling the
machinery for the start-up of the big project and at the same time trying
to quiet the storm of protests stirred up by some of the census questions.
Though the Census Bureau explained that all information will be
confidential, the objections are piling up. The census has become a
political football. Legislatures are declaiming against violation of the
sanctity of the home by the Census Bureau snoopers." That was in Business
Week, March 9th, 1940. So here we are.
With that, I'm happy to take your questions. Steve will organize it.
DR. PREWITT: Well, the basic number in 1990, the completed number that we find reported was 98.6 percent of the population was counted. However, that's a net number. In 1990 about 8 million people were not counted and about 4 million people were double counted, so the net was about a little over 4 or 4.5 million. So, we have been saying all along with respect to this census that we are making every effort to count everyone in the United States. However, we do not believe that we can completely do that. You've been writing the stories. You've been quoting people back to us saying, I've thrown it away, I refuse to cooperate, I'm not going to answer the door, they're not going to get this information from me. There are X number of people in this country who either won't be found, or if found, won't cooperate. The magnitude of that we will not know until we have finished this basic census operation and then done our Accuracy and Coverage Evaluation, our own big quality survey. We will field that starting in August, and it won't be until next -- middle of winter or early spring that we'll report that number.
We will not know, of course -- by definition you won't know how many
people you did not count until you've done that piece of quality work.
There's no other way to know it. What is the true number? A census is an
approximation of the true number. That's what it is. We will have an
estimate based upon this enumeration, but we are not promising to actually
count a higher percentage of the American population from the basic
enumeration that we did in 1990. The conditions that make that difficult
have not lessened. High mobility, irregular housing, high percentages of
immigrant populations, linguistically isolated households and now, of
course, coupled with that has been the stirring of anti-census,
anti-government arguments, which simply will mean some people are going to
slam the door and say, "I won't cooperate."
D'Vera COHEN (Washington Post): You have
had some experience, haven't you, though, with people going door to door
already, the enumerators in the remote areas. What experience have they
reported back on, for example, refusal to answer questions on the long
form or slamming doors in faces? Does it seem to you that that has gone
up, and do you have any numbers to show how many folks?
DR. PREWITT: I don't have anything
systematic on that, Dee. We have not encountered any unusual problems.
Now these are very special populations, as you know. When you're doing
remote Alaska, it's a different kind of exercise. You're doing it with
the authorities, with the cooperation of the village chiefs. That's the
same thing in Indians lands, and so forth. So it's a slightly different
kind of knocking on the door operation than it is when we're now going
back into the rest of the country.
But, no, the good news is that in none of
those operations have we encountered an exceptionally high or unexpected
level of refusals. In fact, we think that we did a very good job in
remote Alaska, which is the one big operation that is complete. A very
complete job there.
MR. JOST: We'll go to our first call on the
telephone. It's Laurie Krise of the Arlington, Texas Morning News.
Before we take your question, Laurie, I can report to you that Arlington,
Texas, is one of 800 communities that reached their 90 + 5 goal over the
weekend, at 66 percent.
DR. PREWITT: Congratulations.
Q: Thank you very much. My question today
is, how helpful is the door-to-door phase? In other words, for example,
in 1990 how much did it help push up the numbers and get the response rate
DR. PREWITT: Well, there is no census
without Non-Response Follow-Up because the base therefore, would be
approximately 65 percent of the population, which would, by any count, be
a disaster if you could only count 65 percent of the population. So we
always expect a heavy Non-Response Follow-Up workload. That's been true
since we went to the mail-out census, which really starts partially in
1960, and then in 1970. So, we've really only had three censuses, which
are like the one we're conducting in 2000, where we first do the mail-out
and then we do the door-to-door follow-up.
But certainly we've always expected to have
a heavy Non-Response Follow-Up workload. That's what we'll have in 2000.
And we expect to get close to 100 percent, but we would be very surprised
if we actually got to 100 percent. As I say, we won't know that until
we've done our Accuracy and Coverage Evaluation.
Q: Congressman Coburn from Oklahoma, one of
the anti-census critics, he's using his two weeks away from Washington to
hold a series of town meetings and he's going to be talking about the
census. One of the things he's going to be doing is again urging his
constituents to politely refuse to give any more information to the
census-takers other than the basic questions. Would you respond directly
to that, and can your agency -- or will your agency enter that part of
Oklahoma with a different agenda or go to those households differently?
DR. PREWITT: No. We can't fine-tune the
census operation at this stage to try to respond to local campaigns that
have as their purpose creating less than complete census responses. I can
simply say what I've said before, that if there is serious degradation of
the quality of the census information that the country will suffer for 10
years. That is, the array of programs that depend upon high quality
census data range, as you know, across education, health, transportation,
housing and job training. They also, of course, do connect back to how we
calculate the Consumer Price Index, to which the Social Security is of
course indexed, as are all private pension plans. They connect back to
our unemployment numbers, which, of course, are what are used to make
judgments about the state of the economy.
So we put at risk, if we seriously degrade
the quality of long-form information, we put at risk a very substantial
number of government obligations. There's simply no doubt about that.
We are hoping that when we actually look at the data that only a
scattering of the American people will cooperate in these efforts to
degrade the data. But all I can simply say as the director of the Census
Bureau, that if the data are degraded, we have put at risk programs which,
after all, are federally mandated programs. These are not Census Bureau
programs. These are government programs.
I know there has been a lot of concern
about the race question, but I can only remind us all that the U.S.
Congress in 1964 did pass the Civil Rights Act, in 1965 the Voting Rights
Act, and short of repealing those acts, or deciding to administer them
differently, we need high-quality race information about this society.
On the other hand, all we can do is train these enumerators to do the best
job that they can. In this book, which is the training material, we
obviously have all kinds of special sections, what are the laws, how is
this question used, what happens when you get refusals, how do you try to
convince them. What happens if they will give you some information but
not all the information.
Just keep in mind how big this operation
is. We have 500,000 people today, 495,200, whatever the number was.
Close to 500,000. There will be more tomorrow, going through this kind of
material. These are not professional survey takers. Those of you who
have gone to training sessions will recognize that they are people, as we
have all along promised, hired from the local communities to try to count
those communities. So we are putting those people in a very difficult
position. We're saying it is the law that they are supposed to get this
information, and it's not a law the Census Bureau has passed, it's a law
that the United States Congress has passed.
They're going to knock on the door and someone's going to say, "But my congressman says I don't have to." All we can do at that moment is say, "Do the best job you can."
There's not much else that we can do. It is too bad that at this key
moment in the census process that there are efforts to convince the
American people not to cooperate with one of our most fundamental civic
MR JOST: We'll go to the phones again.
Mark Skertic of the Chicago Sun Times.
Q: Good afternoon, Dr. Prewitt. What kind
of training or what kind of precautions have been put in place for
enumerators who have to go into high crime areas, such as a housing
project or a neighborhood where crime is a major problem?
DR. PREWITT: I very quickly covered that,
but let me just repeat it. We have two processes. One we call blitz
enumeration. That's a particular place, like, say, the Robert Taylor
homes in Chicago. We will send a whole team in one day and try to simply
cover that entire complex over a period of days, in that case, because
that's a number of buildings. In other areas where simply it is a high
crime area, we send them in pairs. But beyond that we do have special
instructions to all of our enumerators, wherever they are in the country,
to say, here's how to psych out, if you will, or anticipate a troubling
situation, a possible crime situation, a place where you could be put at
risk. Here's what you do. Don't go back. Wait until you call your crew
leader, get help. We're very, very focused on making certain that we
don't put any of our enumerators in a vulnerable position.
On the other hand, we do have the
responsibility to the American people to count everyone, and so we're
constantly balancing off the two things. How do we make absolutely
certain that we don't put an enumerator at risk, and on the other hand,
that we don't create a situation in which whole areas don't get counted.
Because then they are suffering because they don't get the kinds of
benefits or programs or attention that they merit.
So we can't put law-abiding citizens at
risk by not counting them simply because they happen to live in a
neighborhood where there are crimes. So we are balancing that off as best
we can and trying to give our enumerators all kinds of assets, clues,
instructions, materials and so forth to make certain that nothing does
happen to them, and we will pay very close attention to that.
PAM FESSLER: Of the 42 million households
that the enumerators have to go to, how many of those people have to fill
out the long form rather than the short form? And how will that
complicate the follow-up job?
DR. PREWITT: Pam, I'm going to have to do some very quick arithmetic. It should have been 1 out of 6, but since the differential and the return rate is, as you know, higher -- that is, fewer of the long forms came in -- I'm worried that I'll get my arithmetic wrong. At any rate, it will be close to 20 percent of those households will have to get a long-form questionnaire, which is higher, as I say, than we had anticipated, since we anticipated 60 or 70 percent. Actually we anticipate a little higher because we always have the short forms. I'll have to do my arithmetic and give you the right number.
But certainly it puts a bigger strain on the enumerator task. The two things. One, it's more difficult to administer, and also you get higher levels of resistance. I want to say that we got resistance in 1990, as we got resistance in 1940, as I just read. There's always been resistance. It's always been harder to get the long form data than the short form because it takes a bit longer, 45 minutes. Obviously you would say, correctly, that 45 minutes doesn't seem like an awfully large investment to make every 10 years. It's not like we're asking the American population to stop everything and work hard at this for months and months and months. We are asking for 38, 45 minutes. Nevertheless, there will be people who say, I'm too busy, I don't want to take the time, I don't know the answer to this, and so forth. And that will put a strain on the enumerators.
They are being trained today to go ahead and be as persistent as they
are capable of being, consistent with being friendly and cooperative and
neighborly about this task. It's very hard for us today, Pam, to say
exactly the magnitude of additional burden that will occur. We simply are
going to work on that now, but what we worked on last week is getting
these 500,000 people at the right places with the right materials, and
MR. JOST: On the phone is Melissa Campanelli with Direct Marketing News.
Q: I'm just wondering about more on the
long-form. Are you sending out enumerators to households that have only
entered some of the long-form questions? Are you asking for the
households to fill out the rest of the questions?
DR. PREWITT: No, we do not -- in this
operation that we're starting tomorrow -- we are only going to households
from which we did not get any kind of form back. These are nonresponding
households. If a household sent back a long form but did not completely
fill it out, they are not part of our nonresponding household universe.
They become part of a different universe, and that doesn't happen for
about another two months. That is the universe of people who -- well,
actually some of that work starts a little sooner. Some of that work
starts in about two or three weeks.
If we get a long form in that says there is
so many people living here but there is not enough information on the long
form to determine with some level of certainty, then they will get a
follow-up phone call, or, if necessary, a field visit. We cannot put a
household into the census file unless we are certain that there are people
living there. As I've said in testimony and have said to the press
before, if someone sends a form in and says there are 23 of us living here
and no information about them, we simply can't put 23 into the
apportionment count or districting count, for obvious reasons. We have to
make certain that someone is there.
What we cannot do in a census environment,
in a census year, is when we get long forms that have the basic -- the
minimum amount of information to be certain the people are there but have
not all of the other information that you need for these other purposes --
we do not have the capacity, the budget or the operational tools in place
to go out and get that information. The country will simply have to get
along without that basic information.
Q: Thank you.
MR. JOST: That's you. Go.
Q: Hi, Simon Kennedy, Bridge News.
You touched on the role the census played
on the economy, in the jobs report and the March data, or the March
employment report, [which] was the last, kind of showed a combined through
this year of about 117,000 people. And I was wondering how many people
have been hired since then for your -- for the census?
DR. PREWITT: I think if I understood you
correctly, we had hired 117,000 earlier for other operations that we
talked about before --
Q: Through -- purely from a 2000
perspective. I think you employed 39,000 in January and February --
DR. PREWITT: That's right.
DR. PREWITT: Those people were hired for a
certain task. They then went off the payroll. Now, I don't know what
percentage of them are now part of our 492,000 but a high percentage of
them, anybody who did an earlier census job for us and did it well and
wanted to remain a census employee, we would then put them back on the
payroll. My guess is a very high percentage of the initial people who did
our early work for us have now come back in on the payroll, but I can't
give you the exact percentage.
Q: Is the 492,000 -- I don't want to kind
of compare apples and oranges -- is that people who've been employed
purely for carrying out the 2000 census or is that the total payroll of
the U.S. Census Bureau?
DR. PREWITT: Oh, no. We're just talking
about decennial census payroll now in all of these numbers. In addition,
the Census Bureau has others doing other kinds of surveys, other kinds of
work, economic statistics and so forth, including field work. But all of
these numbers are just decennial census numbers.
Q: And when did you start employing people for --
DR. PREWITT: What?
Q: When did you start employing people for the decennial -- ?
DR. PREWITT: For the decennial census?
Well, our first big field operation was where we did the address work and
that was almost two years ago.
Q: Okay. Thank you very much.
MR. JOST: Okay, on the telephone from the
Philly Daily News is Myung (Oak) Kim (ph).
Q: Dr. Prewitt, I have two quick
questions. First of all, of the total, the almost 500,000 workers, what
percentage are enumerators?
MR. JOST: Nearly all.
DR. PREWITT: Trying to do the arithmetic
in my head. It's kind of eight to one ratio. They are 40,000. Of those,
close to 440,000 are enumerators.
Q: Oh, okay.
DR. PREWITT: There are 40,000 crew
leaders, and there's another layer, about 5,000, what we call district
level supervisors and another level in the local offices. But I would say
approximately 435,000 to 440,000 of those are enumerators.
Q: Okay. And then going back to the
safety issue, are you aware of any enumerators or other census workers who
have been threatened or hurt during non-response follow-up since this
DR. PREWITT: Not yet in the 2000
experience. In previous censuses we certainly have had occurrences.
Someone was shot at in 1990, I think in the Bronx. We've had dog attacks.
We've had plates thrown at enumerators; I mean, angry kinds of outbreaks.
But, fortunately, we have not suffered any serious injury or fatality in
that kind of incident.
We have already suffered -- in this census
we've had, I believe, the current number is four census deaths; that is,
deaths of people working on census material. These have been traffic
fatalities. After all, there's a lot of people covering a lot of miles,
especially in our rural areas, mountain states, and so forth. And when
you have that many people driving that many miles, then you, of course,
are going to suffer certain kinds of accidents. And I believe the number
today is there've been four deaths of census workers since the decennial
census started. The number that I had at one time; this is now dated;
this was about two weeks ago. At that time we'd driven a total of 62
million miles getting the census materials out and, of course, that
number's been expanded since then somewhat.
Q: Okay. So as far as non-accident
related situations, you haven't had any so far then?
DR. PREWITT: No. And on the other hand,
the operations have not required us to do what we're now about to do,
which is to start knocking on lots of doors. That is, for that purpose,
for that part of the operation, we're still in training.
Q: Okay. And then the three local census
offices that you mentioned earlier, that was less than half a percent,
that had shortcomings of enumerator recruits --
DR. PREWITT: Yes.
Q: Are any of them in this -- in my area?
DR. PREWITT: Remind me where you are?
Q: In the Philadelphia area.
DR. PREWITT: No.
Q: Okay. Are they very far away? Can you say where they are?
DR. PREWITT: Oh, no, I'm not trying to be --
MR. JOST: You're at the five follow-up limit.
DR. PREWITT: There's one in Chicago, one in
Wisconsin and one in Florida.
Q: Thank you very much.
DR. PREWITT: So you can decide whether that's far away or not.
Q: Thank you very much.
Q: In terms of non-response follow-up,
would you say this year you have the more difficult task in rural areas or
in big cities, urban areas?
DR. PREWITT: Well, that's a tough
question. I think historically we've had more difficulty in urban areas.
One, people aren't at home and you have a higher concentration of
population groups which are resistant. Young, unemployed males, for
example, are very hard to count. You have a higher density of rental
units. It's harder to count for all kinds of reasons. We never get as
full a count of people who rent as we do of those who own. They feel less
settled, less connected to the community.
What I cannot tell is the long-form effect.
The long form, not to forget, there are -- a higher percentage of the
people in the rural areas get the long form because you have a higher
sampling ratio. You use the long-form data to make statistical estimates
about population characteristics as --
[MOMENTARY BREAK IN RECORDING.]
DR. PREWITT: -- New York City, you don't
need 1 out of 6; you need 1 out of 10. So if we have more long-form
resistance than we had in previous censuses, then the pattern of
resistance that we've encountered in the urban areas of today will shift
in part to the rural areas. But it's too early -- again, too early to
MR. JOST: We'll go back to Pennsylvania.
The Pittsburgh Post-Gazette. Gary, are you with us?
Q: Yes. Dr. Prewitt, two basic questions,
please. One, any demographic breakdown you can give of your enumerator
workforce by gender, age, race, et cetera? And second, for the 42 million
who have not returned the forms, based on past experience, any estimate as
to how many of those are people who actually received forms and
deliberately did not return them as opposed to the number who probably
just never got a form in the first place?
DR. PREWITT: Well, let me take the second
half. It's, again, very difficult to estimate that. We certainly know
that some parts of those 42 [million] households did not get forms. For
example, there is, in that household list, what we call new construction.
These are houses that simply weren't even built when we started the
mail-out process, and weren't lived in. And we've added those records to
our Master Address File, but we added them right up until April 1st. We
obviously didn't mail to them. And now we'll be knocking on the doors.
There are other areas where there were some
mail delivery problems, having to do with the fact that we don't mail to
post office boxes. But some people get their mail at post office boxes
but they weren't delivered to their homes because the post office treated
them as city-style addresses. I won't go into the details of how that
happened. But, yes, there are perhaps as many as a million households who
will not have gotten a form for one reason or another. So, when we go
knock on those doors we hope these are people saying, "Hi, we've been
waiting for you; glad you finally found me, and I want to answer my
questionnaire." It's not a large number and that we know because we did
have very good luck with our Mail Address File. We know the number of
people who did tell us they got the advance letter, the questionnaire and
We actually got a lot of calls and e-mails
back from people who said, "I got the advance letter and I got the
follow-up card, but I did not get the form. What happened?" Well, that's a
mail delivery problem; that wasn't an address file problem. If they got
one piece of mail from us that means they were on our address file and the
address file worked, but for some reason they did not get the form. You
know, bags and forms get lost, things happen. But we do not think that's a
huge number, and most of the people who did not mail the form in either
lost it, couldn't be bothered, put it aside, whatever happened, well, we
won't know until we knock on the door and start trying to get them.
The demographic breakdown, I'm sorry, I
don't have that in front of me. It's available, in effect. I do know
that we have recruited a -- well, I can only give you some anecdotal
evidence. We know that we've hired a large number of people who are
bilingual. We were yesterday in a training session here in the New York
area -- or in the Washington area. And I went to that training session.
And I would say at least a third of the people, 40 percent of the people
were bilingual, and I would say a quarter of them were not born in the
United States, and they were from all over, because this is a high
immigrant area. There was someone there from Nicaragua, from Bolivia,
from Russia, from --
MR. JOST: Ghana, Vietnam.
DR. PREWITT: -- Ghana, Vietnam, a quite
remarkable collection of kind of new immigrants, out-and-out trying to
count their population groups. I can also tell you, but this is not very
systematic. This is not the hires; this was the recruitment pool. About
60 percent were female and 40 percent were male, and about the same
number, 60 percent, were over 45, which means that we got a good supply of
probably women who either were part time in the workforce or had not been
in the workforce and wanted to do this work. We're very pleased with
those kinds of numbers. That means we have a more mature workforce than
we might have expected. Indeed, the session I went to yesterday, just
interacting with about 35 trainees, I would say there were two or three
senior college students, that is, maybe graduate students, that age grade.
But the preponderance were certainly mature workers and many of them
retired workers. Of course, that's in the Washington area with a lot of
federal retirees, and they've come back to do this job.
MR. JOST: We have time for two more. Bob?
Q: Yes. Just eyeballing the list, there
seemed to be a number of states that had significant drops from 1990.
And I wonder if there's anything you had heard or anything they have in
common. Pennsylvania, Ohio, Kansas, Michigan, Indiana, all were down at
least three or four points in their initial response rate.
DR. PREWITT: Yeah.
Q: I don't know if there's anything that they have in common from --
DR. PREWITT: No. This question, of
course, came up the last time we went over this in more detail. And we
simply, as I said then and have to repeat today, have not been able to do
any serious work on the response rate.
I would say it's very important when you
look at these response rates that any state which is down 4 percent or
less actually met what we thought it would get. Don't forget, the 1990
response rate was 65 percent. Our predicted national response rate was 61
percent, a 4 percentage point drop between '90 and 2000. So any state
which is 5 percentage points or lower actually did not perform as well as
we thought it would.
Meeting the 1990 target was no modest
accomplishment. And the fact that the country did it, as I've said
before, I consider to be a very serious achievement by the country and by
the Census Bureau. And so I'm hesitant. There're certainly a few states
that are below that 4 percentage points. And as a state, Delaware was one
I see; Pennsylvania, barely. The state or the area which looks the worst
on this piece of paper is Puerto Rico. And I do want to defend Puerto
Rico again, because I think we did not treat them as we should have in
this process. Just to repeat it. Puerto Rico did not have a mailout/
mailback census in 1990. Therefore, it had no base against which to
measure its performance. And we asked the governor of Puerto Rico, "What
would you like to set for yourself as the target?" He said "I will beat
the national target." That was a very, very high mark. So that 15
percent differential is really a little unfair to Puerto Rico. We had no
way ourselves. In fact, our best guess is that Puerto Rico would come in
at 50 percent, and that's exactly where it came in.
We budgeted in Puerto Rico for a mail-back
response rate of 50 percent. And, lo and behold, they got 50 percent,
which means the people down there who worked with the Census Bureau had a
really very subtle and complicated understanding of how Puerto Rico would
respond to its first mail-back census.
But, setting Puerto Rico aside, actually
very, very few states failed to meet the target that we had set, which, as
I say, 4 percent -- not target, but our predicted level -- 4 percentage
points below the 1990 number.
MR. JOST: And just to clarify, in your handouts, those asterisks indicate states whose rates went up since our last report to you last week.
With that, we'll go to our last question for Paul Johnson at the
Bergen County, N.J., Record.
Q: Yeah. My question was basically what
is your average pay scale for the enumerators? And I missed the
beginning, so I didn't know how many people you had hired as enumerators.
DR. PREWITT: Right. As I said when we
started, we currently have on the payroll 492,245 people, which is well,
well above the minimum we need to start this task. The wage rates vary a
lot, depending upon the community. They're set, depending upon the local
pay rate, from roughly $8.75 to $17.50, I believe, for enumerators.
Q: You wouldn't happen to know what the rate was in New Jersey.
DR. PREWITT: And the rates are slightly
higher, of course, for crew leaders and other supervisory work.
Am I right on that number? I think it's 17.50 is the top rate.
MR. JOST: Eighteen.
DR. PREWITT: It's actually 18 now.
MR. JOST: Eighteen-fifty.
DR. PREWITT: I should say, just to back up
on this, because some of you expressed interest in this earlier, we set
wage rates. We then went into our recruitment effort. And in those areas
where we felt like we were not getting a deep enough recruitment pool, we
changed wage rates and we changed wage rates in 10 percent of the local
offices. So that was about the number. So some of those earlier numbers
you saw, we set a wage rate. That's why it's probably 18 in some of those
places where we increased the wage rate to make sure we had a deep enough
Q: All right. Thanks a lot.
DR. PREWITT: Sure.
MR. JOST: Okay. With that, the one last
reminder. The response rates, the final posting of the response rates,
will be public tonight at 6:00 on our Web site, www.census. gov, for all
38,000-plus jurisdictions in the country. And you may see at smaller
levels of geography bigger changes than in the national rate if a region
had a large amount of mail that came in in the last six or seven days.
Thank you very much. We'll see you again in a few weeks.
[END OF PRESS CONFERENCE.]