Increasing U.S. Census Participation
One of the biggest issues with the U.S. Census and probably the one that wastes the most money is trying to count those who are hard to count. My personal fix would be to use sampling to solve the problem, but for now the task of the Census takers is to try and count everyone. My attention was brought to a project called “Census Hard to Count 2010” which maps the “hard to count” population nationwide (based on the Census Bureau’s analysis) to help local and national organizations target their outreach efforts for the 2010 Census and customize messages to communities at risk of being undercounted.
It features interactive maps at the state, metro, county, and tract level, along with detailed statistics for each area. You can search in various ways, and also add overlays showing Congressional districts, ZIP Codes, tract-level maps of 2000 Census mail return rates, and recent foreclosure risk. There’s a FAQ that goes into details about the data and their methodology.
Clearly larger states have a bigger problem with hard to count populations but Alaska, Hawaii and New Mexico probably point out that there are socioeconomic factors as well. Using the demographic layers available in the web app shows that this problem is very difficult to pinpoint and my hat is off to those trying to crack it.
The UI from the Census Hard to Count 2010 Application
Look at that… a flash based map site.
Is it a waste of money in your opinion because the effort of counting hard to find individuals is poorly managed or because the census should not try to count these hard to find individuals? I hope it is not the latter.
Well neither to be honest. I think sampling gives a better indication of population for the price. More value to my tax dollar.
Now don’t for a minute assume that I don’t think these hard to count folks should be ignored. On the contrary, I think sampling gives them a better voice than only counting those who fill out a form or can be counted in person.
While sampling is definitely a good approach, I think the 100% goal may still be achieved: with mini-robots that barge into your home like in Minority Report. I know what you’re all thinking – “the technology isn’t there”… and yes, you are correct (maybe). But do we not have the ability to reprogram Roombas? Free Roombas for every household in each HTC tract.
While cleaning,it also conducts its survey and once complete, it self destructs.
The problem with sampling in the census is…legally it shouldn’t be done. It’s suppose to be enumerated. Until the law is changed, the slow, hard head counting should be done.
My really problem with the census is all the other questions they ask.
edit…I see after posting, summitup stated the same thing.
Have to remember that, Constitutionally, the Census is an enumeration, not a count. That technically rules out sampling. We may never see the list of names, but the Census’ Constitutional purpose is to make a list of everyone in the country, not count everyone in the country.
(Edit: Interesting how we all jumped to the same thing without reading each other’s comments
CENSUS is defined as an enumeration. Since sampling, modeling, and estimating is not a census .. you can not substitute sampling for a census.
Agreed, you can’t do it today. Congress should change it because there is no point to counting everyone when you can’t count everyone.
When has this not been the case? When has it really been the case that the census could “count everyone”?
I doubt it, but I’m saying trying to do so is like trying to catch a unicorn. It might be possible, but I wouldn’t bet on it.
Better to not waste the time and come up with better solutions.
The point is that the official “number of people in the state of Nebraska” is the enumerated count, period. Even though an estimate based on a sampling methodology would be closer to the “truth”. The Census can generate those estimates, but they won’t be the official “population” used for trivial things like allocating seats in the House of Representatives, etc.
Sooo…how does the Census estimate the number of people they miss?
LOL, they sample!
Summitup is right: You can’t create realistic estimates and samples without having some “real” data on which to base them. I’m working on an IT project related to the UK 2011 census, and I know the census managers are very concerned to try and make their counts as reliable as possible, not only by reaching people and counting them, but also by comparing their counts against other known statistics. Of course, this may involve comparing census counts with samples e.g. from separate healthcare or education statistics, but the aim is to try and have a reliable baseline every 10 years (in the UK) on which to base subsequent policy. And for that you need real data, otherwise you’ll just get ever greater variation between what you think is a reliable sampling model and reality (especially given the political and financial issues relating to census counts and the corresponding pressure to falsify/distort statistics). Incidentally, I like the HTC website, but I didn’t see “students” in the “hard to count” categories, yet they can really distort the statistics for a district, especially if census day coincides with the hangover period after a big campus party…
This is slightly off topic, but it will want to make you laugh or cry. Our county received 10-20 paper maps and a box of red pencils from the census bureau earlier this week to update various boundaries.
We are not going to do it that way.
Eh that’s the way they’ve been doing it for a while, lowest common denominator I guess. As the 1 person GIS department in New Mexico a few years back we had to do the same thing and it worked out fine in the end.
Before you change the way they want it done check with the Census Bureau first because they not have the means in the local or regional offices to deal with some type of super whizbang GIS data.
I’m curious how the map boundaries actually look after the big CENSUS/Harris MTAIP project to clean up the census maps around the country? I have a bit of effort invested there so I’m curious how it looks in the end…
[...] of some) is a cake walk compared to some of the more infamous censuses in history.  James Fee points out an online mapping effort relating to the 2010 Census called “Census Hard to Count [...]
Glad you folks are at least talking about the Census. I didn’t want to make a million maps for our people trying to assess the Hard-To-Count – so we made a viewer too – Flex based. Here’s the link: http://gis2.lacounty.gov/drp/census/
It’s simple – just for Los Angeles County…but you get the idea.
Despite your insanity using the flex viewer (I kid, I kid), thanks for the link.
Very interesting stuff!
Just an FYI that today we added two new features to the http://www.CensusHardToCountMaps.org site:
I think the Twitter feature is especially neat, but will also be actually helpful to local groups. Be a good test of the value of the Twitter geo api.
[...] of some) is a cake walk compared to some of the more infamous censuses in history. James Fee points out an online mapping effort relating to the 2010 Census called Census Hard to Count [...]
What is wrong with our government when they spend huge amounts of tax dollars sending out a letter to millions of Americans telling us to expect another letter in a week?
See the letter, if you didn’t get one yet, and read about it at http://www.realvaluesite.com/?p=229.
What can we Americans do about stopping this horrific waste of our tax dollars?
Let’s see if it helps get those response rates up. Of course it MIGHT be a waste of our tax dollars – but this is our Constitution and they are required to do this complete count (I’m a fan of sampling myself). Since we’re not done yet, let’s not rush to judgment; we don’t know it’s a waste just yet!
[...] Increasing U.S. Census Participation (by James Fee) – Census is coming up and the challenge will be to count the ‘hard to count’ population. [...]