First, users should note that zip codes are determined by the U.S. Postal Service for the purpose of delivering mail, and are not technically geographic boundaries the way we typically see them on a map. Zip codes are used to pinpoint the mail delivery location, typically at an address or a Post Office. Alternately, the U.S. Census Bureau uses nested building blocks (such as census block groups and tracts) to form geographies, but these do not reliably line up with zip codes, which often cross county and other boundaries. An additional challenge is that while census blocks use streets as edge boundaries, postal delivery routes generally service both sides of a single street.
However, because data users often desire data by postal zip codes, some economic data are available by zip codes in the Census Bureau's County Business Patterns and Economic Census. However, data from the 2000 Census, 2010 Census, and the American Community Survey (beginning with the 2007-2011 five-year estimates) are only available in a special Census geography called Zip Code Tabulation Areas (ZCTAs, or zic-tuhs ). Census ZCTAs assign to each census block the most frequently occurring ZIP code for the addresses within the block, and in this way, approximate the true postal zip code.
How closely aligned are the ZCTA boundaries to Zip Codes?
They can differ a little or a lot. A ZCTA in a densely populated area is likely to more accurately represent the area covered by a zip code than ZCTA containing a smaller population. The best way to confirm the extent to which a ZCTA represents a zip code is to overlay zip codes and ZCTAs using mapping software.
Important notes for working with ZCTAs:
There are basically three ways to examine Somali and other African populations with Census data--by examining answers to questions in the U.S. Census Bureau's American Community Survey regarding country of birth, ancestry/ethnicity, or language spoken in the home.
Ancestry was first measured in the 1980 decennial Census, but Somali (typically appearing as Somalian in the table) ancestry first appears in the 2005 American Community Survey data tables for Minnesota. Examining ancestry data may include people born in Africa, and people born in Minnesota (or anywhere else). This means ancestry may be best for including children in Minnesota's African communities who were born in the U.S. However, because ancestry data result from an open-ended question, one challenge with ancestry data is that some people answer with a broader ancestry group. It's likely that some people of Somali or Liberian ancestry answer simply with "African" which would exclude them from ACS estimates of Somali or Liberian ancestry. Data are available by the first ancestry they selected, or total ancestry, where individuals are represented in all ancestry groups they have selected. However, users should note that in total ancestry tables, the sum of all the ancestry groups will far exceed the total population because individuals are represented multiple times.
These African ancestries are available in the tables from the 1990 and 2000 decennial Census long-form (SF3) data:
Ancestry does not appear in the 2010 decennial Census data.
These African ancestries are available in the ongoing American Community Survey data:
Find data about African groups on our Data By Topic: Immigration & Language page.
See data about languages spoken in MN on the U.S. Census Bureau's American FactFinder website.
See data about how many Minnesotans claimed Somali or other ancestries on the U.S. Census Bureau's American FactFinder website.
When conducting surveys like the American Community Survey, the U.S. Census Bureau samples a smaller number of people that they use to represent the total population, and produces estimates from those responses. With every survey estimate, there is a corresponding margin of error that attempts to measure everything that isn't captured by only asking questions of a portion of the population.
Adding and subtracting a margin of error from its estimate gives you a range within which the actual number is likely to fall. Census Bureau data is published at the 90% confidence level, meaning that nine times out of ten, the true value will fall somewhere within the range you've created around the estimate (by adding and subtracting the margin of error).