Today marks the first-ever release of 5-year American Community Survey (ACS) estimates, an agency goal nearly 20 years in the making with strong support and funding from Congress. These can be used by everyone to examine their neighborhood characteristics during the 2005-2009 period and compare them to those of adjacent areas.
The estimates will tell us key attributes of our communities – the prevalence of moving into the area; the mix of rental and owned homes; the average number of vehicles owned by households; how people in the area travel to work; how long a trip to work they have each day; whether there are multi-generation families in the area; what languages are spoken in the area; variation in income in the area; how many hours per week workers put in; how high are the housing expenses in the area; the prevalence of military veterans in the area; and citizenship status of persons in the area.
These will be released on over 670,000 different geographical units, as small as a group of census blocks, but also for school districts, congressional districts, state legislative districts, counties, and other places.
This is the first time such a release has occurred, and it is probably the largest release of statistical estimates on one day in our history. (Each estimate also is accompanied by a sampling margin of error, so there are really 22,000,000,000 estimates.)
Releasing such vast amounts of statistical information, we’ve re-doubled our efforts to review all the estimates. But as statisticians we know that if the expected 10% of the margins of error are randomly not covering the actual value because of sampling variability, this amounts to 1,100,000,000 estimates. We’re trying to get the word out that cumulating over the smallest areas to form larger areas will also reduce the sampling variability. Since this is the first time the country will be given these small area estimates, we and our thousands of users across the US need to learn about how best to use them. Paying attention to margins of error in these estimates is crucial to wise use.
Under the design of the American Community Survey, this is a beginning of annual releases. Next year we’ll all see estimates for 2006-2010; the year after, 2007-2011; and so on. So each year we’ll get an update that describes the last five years of our communities. As the years pass, the trends of these estimates will tell the story of change in our communities. With this, citizens will have documentation of how things are going in the community; civic officials will know what areas exhibit what attributes.
Before the American Community Survey, the Census Bureau released these types of estimates for small areas only once every 10 years. Within a few years they became out-of-date and useless for planning and community action. The American Community Survey provides more timely information and thereby makes it more useful to all of us.
You can take a look at the estimates right now.
ANYONE WHO THINKS THAT IT IS A GOOD THING FOR THE FEDERAL GOVERNMENT TO HAVE THIS MUCH PERSONEL INFORMATION OF THE PEOPLE WHO LIVE HERE SHOULD RETHINK THAT VIEW.
I think your discussion of the margin of error around the estimates is very confusing. How did you arrive at the figure of 22,000,000,000 estimates? Is this counting every point in between the margins of error as a separate estimtate? I think the following paragraph is equally, if not more, confusing. If I am having a hard time understanding this (I work at Census), the average reader must be scratching their heads over this explanation.
This is great news…have they mapped out the top 10 cities?
I’m confused. The three municipalities I looked at from the survey have a combined population of over 21,500 people, but this survey shows their total at just over 11,000. That’s some kind of margin for error!!
I find it interesting how the federal government isn’t for “racial profiling” yet it’s obvious they are when doign the census. Hypocrites.