​​​Rodriguez Data Solutions  LLC

a Small/Minority Business Enterprise

24%

Always Include the Margin of Error


As seen in Chart A, there is a large gap in the poverty rate between Hispanics and non-Hispanic Whites.  In 2010, the poverty rate for Hispanics was 24 percent but only 6 percent for non-Hispanic Whites.  We want to know how the gap has changing over time.


Chart B includes the margin of error associated with the same data.  It shows that the poverty rate for Hispanics increased from 24 percent to 26 percent from 2010 to 2014.  However, the increase was within the margin of error of 2 percent, which means that there was no change.  You might then infer that the poverty rate of Hispanics remained the same.  This would be correct.


Data for ZIP Code delivery areas are often used in socioeconomic analysis; however, the boundaries of ZIP Code delivery areas change.  Furthermore, a single ZIP Code delivery area can include parts of several municipalities.   


The United States Postal System does not provide maps of ZIP Code boundaries and the U.S. Census Bureau warns data users that its maps of ZIP Code delivery areas are inaccurate. 


A review of more than 600 ZIP Code delivery areas in Connecticut found that numerous ZIP Code delivery areas cover more than one municipality.  For example, the delivery area for ZIP Code 06754 in northwestern Connecticut includes parts of the municipalities of Sharon, Cornwall, Kent, and Warren.  In situations such as these, it would be misleading to assign data from  ZIP Code 06754 to any of the municipalities in the area.

​ZIP Code Delivery Areas in More Than One Municipality

Accurate Data

​Poverty Rate

Chart A: Not Showing the Margin of Error

22%

Data Source: ACS 1-yr, Confidence Interval = 90%

"Fast is fine, but accuracy is everything." 

- Wyatt Earp, a famous and very accurate gunslinger

Information from Chart A (without the margin  of error) could lead to a follow-up analysis to find out what caused the poverty rate to change among Hispanics.  This would be unnecessary because there was in-fact no measurable change in the poverty rate of Hispanics. 

Poverty Rate

Chart B: Showing the Margin of Error

Use ZIP Code Data with Caution

28%

26%