You are not alone.  Prior to the 2010 Census all of the socioeconomic characteristics such as income, poverty, educational attainment, housing value and how long it took commuters to drive to work were collected during the decennial census through a “long form” questionnaire.

Long form surveys, which were first collected during the 1940 Census and following the Great Depression, went out to approximately 1 in 6 households.  Every household and every person living in a group quarters environment, such as a dormitory, nursing home or correctional institution, were counted for reapportionment and redistricting purposes, but the houses that received the multi-page survey asking additional detail had the task of answering more detailed questions.

With the increasing need to have more timely information than once every 10 years for planning and programmatic purposes, Census Bureau’s statisticians developed a plan to collect the information on a continual basis.  This ongoing survey, the American Community Survey or the ACS, is sent to approximately 250,000 households per month.

This 250,000 monthly sample is approximately equal to 1 in 40 houses – far fewer than the 1 in 6 houses from the former decennial long form survey.  Because of the much smaller sample size, the Census Bureau must collect all of the surveys for one year before it can release data for even the largest areas – those with more than 65,000 persons.   From now on, each year all areas with 65,000 or more persons will get updated information.  For smaller areas with 20,000 or more persons the data will be collected over a three year period before it is released, and these areas will receive “refreshed” three-year data every year. The smallest geographies such as census tracts or block groups received their first data in December, 2010.  Those characteristics were collected over a five year period, from 2005-2009.

Each of these datasets is considered a period estimate because they are collected over a period of time as opposed to a single point in time like the decennial census.  The annual data are collected over 12 months.  The three year data are collected over 36 months and the five year data over 60 months.

Often when someone refers to a particular ACS multi-year file they erroneously attribute the data to the last year in the period.  For example, the 2005-2009 ACS should not be referred to as 2009 ACS data.   Instead, it represents the characteristics of persons in a particular area for the period between 2005-2009 and they should be referred to as such, for example, the median household income for census tract 7027 for the period 2005-2009 is $90,050.

As of December, 2010 all geographic areas are now covered by the ACS.   The latest ACS releases: 2009 one year estimates released in September, 2010; 2007-2009 three-year estimates released in January, 2011and 2005-2009 five-year estimates released in December, 2010.  NONE OF THESE ACS ESTIMATES ARE FROM THE 2010 CENSUS.

The first 2010 decennial Census data for Maryland was released on February 9, 2011.  It is considered a snapshot in time of the State on April 1, 2010.  The 2010 Census data currently available only provides data on total population, detailed race categories, Hispanic origin, persons 18 years and over by race and total, occupied and vacant housing units.  Additional age, gender, household relationship and group quarters information will be released later this spring and summer.

But from now on, all socioeconomic data like educational attainment, income, poverty, disability, foreign born, travel to work, etc., only come from the ACS.

Mark Goldstein was assisted in this article by Jane Traynham, Manager of Research & the Maryland State Data Center.