Archive for pdACS2013

ACS vs. Summary File 3

The American Community Survey (ACS), which replaces the Census long form Summary File 3 (SF3), is a primarily mail-based household survey conducted by the U.S. Census Bureau with an annual sample size of about 3.5 million addresses and a response rate said to exceed 97 percent. Like SF3, it produces estimates for numerous social, economic, and housing characteristics. These estimates are summarized for geographic areas ranging from neighborhoods to Congressional districts to states to the entire nation. The smallest geographic entity presented is at the Census Block Group level.

The ACS shares many similarities with SF3. However, there are many differences. The chief advantage of ACS data is its far more frequent release. It collects responses continuously instead of every ten years. This gives planners at all levels of government, business, and the general public far more current data than the decennial long form, and provides for the first time information about temporary populations, such as beach and ski communities.

But this advantage is also a disadvantage. While the ACS is timelier, information is also smoothed (flattened) out and has a lower accuracy rate because it is conducted over years of time instead of at a single point in time. This is particularly prevalent for small geographic areas which must pool three or five years of data to accumulate a large enough sample for reliable estimates.

There are also differences in residence rules, boundaries and definitions of geographic areas, how and which questions are asked, and survey methodology.

Our pdACS2013 package is available for those wanting to try out the new ACS data.

What are jam values?

In the world of statistical databases, jam values are common. They are hard-coded information displayed instead of derived measures. They are used to represent unique situations where either the information to be conveyed is an explanation for the absence of data, represented by a symbol in the data display, such as a dot “.”, or the information to be conveyed is an open-ended distribution, such as 1,000,000 or greater, represented by 1000001. Even an empty value or a zero (“0”) will often have a special meaning.

Jam values can also be used to explain why information cannot be disclosed, such as for privacy reasons or because the data does not meet certain filtering criteria.

Data definitions are provided with jam values and normally they are not difficult to utilize. Depending on the parameters of the project, some may even choose to ignore them altogether. However, it is important to understand that they exist, and are not treated like numbers in tabulations and analysis.

Special consideration must be taken when importing data with jam values because they are often in alpha form. Non-numeric values will be converted to zeros (0) if appended into fields that accept only numeric information.

Some users convert jam values to special numbers during the import process so numeric fields can be used. Numeric fields are easier to work with because they do not have to be converted when counted or used in calculations.

You will find jam values frequently employed in American Community Survey (ACS) estimates and margins of error as well as in our own ACS demographics product, pdACS2013.