Coefficient of Variation are super high in the bicycle commute data, ACS_14_5YR_S0801 , almost all higher than 30, poor reliability?

Curious how to interpret this data set? other than with extreme caution? Used the below method to calculate the CVS and almost all of them are greater than 30.

  • Is this data too unreliable to use? 

 

Calculating the Coefficient of Variation (CV)

Since ACS estimates are based on a sample, data are published with margins of error (MOEs) for every estimate. These MOEs are based on a 90‐percent confidence level and give users an idea of how reliable or precise estimates actually are. As the MOE gets larger, relative to the size of an estimate (the smaller the sample, the larger the MOE), the estimate becomes less reliable. A measure called the “coefficient of variation” (CV) is used to discern the level of reliability of an estimate. To calculate the CV, users first calculate a standard error from the MOE, and then divide that standard error by the estimate. The exact steps for the ACS MOE are below.

a. Calculate the standard error: SE = MOE/1.645
b. Calculate the Coefficient of Variation: CV = SE/Estimate * 100

CVs are a standardized indicator of the reliability of an estimate and can help users to quickly gauge the usability of that estimate. The lower the CV, the more reliable the data. There are no hard‐and‐fast rules for setting acceptable thresholds of reliability. The amount of acceptable error in an estimate depends on the analysis or situation at hand.

What is a good CV?

U.S. Census case studies:

  •   High reliability: CVs less than 15%

  •   Medium Reliability: CVs between 15‐30% ‐ be careful

  •   Low Reliability: CVs over 30% ‐ use with extreme caution

Parents Reply Children
No Data