Skip to Main Content

Census Workshop

A research guide designed to be paired with the UVA Library's Research Data Services workshop on the Census. Learn the basics of Census data, and how and why you want to use it.

Data Librarian

Profile Photo
Jennifer Huck
Brown Library 152
Subjects: Data Science

Practice Exercise

  1. Select Tables on the left. 
  2. Select American Community Surveys (1-Year Estimates).
  3. Select American Community Survey (ACS) 2017 - Begin Report
  4. Under List Geographies, Select a Geographic Type, select 050 ... County
  5. Under Select a State, select Virginia
  6. Under Select one or more geographic areas, select Albemarle County, Virginia, and Loudoun County, Virginia, and click Add and Proceed to Tables.
    1. Note that independent cities are listed below counties in Virginia.
    2. Note that Charlottesville is not listed is too small to be in the 1-year estimates!
  7. On the next page, change the tab at the top to Search by Keyword.  Search for "educated," then search for "educational attainment."
    1. Note that "educated" pulled up zero results.  Social Explorer can be particular about the keyword search.  
    2. Note that there are several tables with an id that begins with "A" at the top, followed by tables that begin with either "B" or "C."  "A" indicates that it is a custom Social Explorer table, while "B" or "C" indicates it is a regular Census table.  (B = Base, C = Collapsed.)  Custom Social Explorer "A" tables do not come with Standard Error reported, and are often even more collapsed than the "C" tables.  Census "B" and "C" tables do report Standard Error. 
  8. Scroll down to B15003: Educational Attainment for the Population 25 Years and Over, then select Add, and Show Results. 
  9. Download to Excel

Standard Error

You are probably familiar with polling results showing "+/- 3 percentage points."  The error gives you an idea about how confident you can be that your estimates are correct.  Much like those ubiquitous polls, the ACS survey comes with error, too.  Because the sample size is so much smaller than the regular Census, the margins of error are much greater.  This workshop is too short to go into the finer details about standard error, but you should have some basic awareness about it. 

What is standard error?  The Census says "The standard error is a measure of the deviation of a sample estimate from the average of all possible samples."  

Note: Social Explorer reports standard error, while American FactFinder reports Margin of Error.  Margin of Error is a simple transformation of standard error.  The Margin of Error reported in American FactFinder reflects a 90 percent margin of error.  You can use the reported standard error in Social Explorer to create a Margin of Error. 

Why would you care?  If you are comparing estimates, for example comparing two counties within the same survey, or the same county across different survey years, you should conduct statistical testing to determine if the differences are statistically significant or not. 

You can learn more about simple statistical tests for the ACS here:

Practice Exercise - Standard Error

Using the "American Community Survey: Accuracy" link (above), use the formulas located in the "Instructions for Applying Statistical Testing" to determine if the different categories in the "Educational Attainment for the Population 25 Years and Over" table for Albemarle and Loudon Counties are statistically different at the 90% level.  


More Practice

Using Social Explorer:

  1. Find the Census Tract GEOID (FIPS code) for your home address using the Census Geocoder.  Use the 11 digit number to pull age statistics. 
  2. Find an interesting table for Charlottesville in the ACS Supplemental Estimates.  Find a comparable table in the ACS 5-year and note any differences. 
  3. Pull "sex by age" statistics for Crozet, Virginia.