ACS Data Issues

1
Statistical Significance Testing
Why do it? When to do it? How to it?
2
Testing is Important
Red shading is not statistically significant
3
Statements you might want to make

• Estimate X is bigger than Y
• Estimate X this year is larger than X last
  year
• Estimate X is smaller than Census 2000 value
• State A has the highest Value

4
Comparing SF3 or CTPP 2000 to current ACS
Lets do a test
5
How do I do a significance test?
6

• Obtaining Standard Error from MOE is Key
• Sum or Difference of Estimates
• Proportions and Percents
• Means and Other Ratios

There are different formulas
Consult Appendix 3 in the Compass for
Understanding and Using ACS Data What General
Data Users Need to Know Report Series. JUST
CLICK HERE
Where.
http//www.census.gov/acs/www/Downloads/ACSGeneral
Handbook.pdf
7
However HELP is Here
PUSH BUTTON SIGNIFICANCE TESTER
trbcensus.com/notes/StatisticalCalculationsMenu.xl
s
8
But what if I am using 2000 non-ACS Data?
Click Graphic and review Chapter 8
where do I get my SEs?
9
You will need 2 things the Standard Error and
Survey Design Factor
Survey Design Factor
X
N Size of publication area (population) Y
Estimate of characteristic
10
Survey Design Factor
Get Survey Design Factors from the Census 2000
SF3 Technical Documentation, Chapter 8 Table C.
Mode to Work 1.4
1.2 0.9 0.7
http//www.census.gov/prod/cen2000/doc/tablec-fl.p
df
Insert state abbreviation for specific state
11
Solve for the unadjusted Standard Error (SE)
12
Adjust the SE by the Survey Design Factor
13
Populate the adjusted SEs in your worksheet for
2000
14
Calculate the ACS SE from the MOE
Remember at 90 Significance Standard Error
equals the MOE / 1.645
15
Calculate Z and test
is Z lt -1.645 or Z gt 1.645?
16
is Z lt -1.645 or Z gt 1.645?
Green is Significant This is NOT
17
Try to do an exercise on your own?