Calculating American Community Survey (ACS) Reliability

This is a quick and basic primer on how to properly leverage American Community Survey (ACS) data with the Online API.

 

 

The Simple Four-step Workflow

 

Step 1: Gather/Select the ACS estimate and corresponding Margin of Error (MOE) values that you will be working with.

Using the Get Summarizations service, you can view the list of analysis/summarization variables currently accessible with your Online API subscription. The list of variables can include the ACS variables which can be identified by their prefixes 'ACSXXXXX' and 'MOEXXXXX' where the XXXXX portion is part of the shared name of the variables (e.g. ACSSOMEHS and MOESOMEHS, respectively in the example below).

 

Step 2: Using the ACS Estimate and associated MOE value, calculate the Coefficient of Variation (CV).

[ACS] = ACS Estimate Value
[MOE] = ACS Estimate Margin of Error Value
CV = Coefficient of Variation

 

Step 3: Using the CV, determine if the associated ACS estimate is statistically reliable.

 

Step 4: Decide how you will handle the ACS variable based on the reliability.

In most cases, you may want to add value to your application or analysis workflow by reporting the ACS reliability metric. Depending on your use case, you may want to reject the variable if it has an unacceptably low reliability.

 

Examples Using Analysis/Summarization Variables from the Online API


Example 1: [Basic Example] Determine the reliability of the ACS estimate for the number of adults 25 and older who completed some high school but did not graduate for Hale County, Alabama

Step 1
AREA NAME: "Hale County"
AREA ID: "01065”
ACTIVE DATASET ID: "USA"
ONLINE API DATA LAYER ID: "US.Counties"

ACS VARIABLE ID: "ACSSOMEHS"
ACS VARIABLE VALUE: 2557
ACS VARIABLE DESCRIPTION: "Count of 2005-2009 ACS Population Age 25 Years or Older by Educational Attainment: 9th-12th (No Diploma)"

ACS MOE VARIABLE ID: "MOESOMEHS"
ACS MOE VARIABLE VALUE: 350
ACS MOE VARIABLE DESCRIPTION: "Margin of Error for 2005-2009 ACS Population Age 25 Years or Older by Educational Attainment: 9th-12th (No Diploma)"

 

Step 2


Step 3
Since 8.321 is less than or equal to 12, the ACS Estimate for MOESOMEHS (the number of adults 25 and older who completed some high school but did not graduate) for this area (Hale County) is 'highly reliable'.

 

 

Example 2: [Applied Example Using the NEW Smart Map Facts Data & Geometry Query Service] Query every county in the 48 contiguous states where 20% or more of the adult population (25 and older) attended high school but did not graduate (and only return counties where ACS values for 'ACSSOMEHS' and 'ACSEDUCBAS' are considered reliable estimates).

See the REST syntax with embedded ACS calculation.

See the results with the counties which satisfied the query criteria.

 

Further Reading