Tag: Census Data

A Science for Everyone, Community, Demography

Our Communities by ACS Numbers

michaelmccurryphd

I listened to a comment about the median household income in Trego – and defaulted to my professional statement before retirement – “That’s American Community Survey data, and it’s not very good for small communities.”  When I checked it, the $36,458 median household income for Trego translates as “somewhere between $27,478 and $45,438.  ACS data has its uses, but it has to be used with a lot of caution.

So here’s a little ACS data on our communities – you can check for margin of error (MOE) here.   I wouldn’t recommend using any of the numbers without reviewing MOE – but just sharing the data shows the variance.  It’s safe to admit that my household was one selected for the ACS. With two retirees at home, I didn’t hurt Trego’s school enrollment rate, I raised the percentage of bachelors degree or above, kept the employment rate down, and raised the median age.

Trego CDPFortine CDPEureka CCDRexford Town
Population5153176,47078
Median Age60.527.950.153.3
Median Household Income$36,458$68,036$40,827$30,481
Bachelor’s Degree or more26.10%19.20%22.40%0.00%
Veterans6.80%16.20%12.90%16.80%
Poverty9.50%5.20%20.40%23.60%
School Enrollment97.80%72.30%81.90%100%
Employment Rate40.20%59.50%38.30%20.60%
Housing Units2831773,71673
Occupied Housing Units2371442,79646
Disabilities31.10%18.80%26.70%65.90%
Children under 189.30%32.50%22.10%13%

It looks like the Fortine sample drew some younger respondents.  Eureka CCD with a larger population and larger sample is probably closer to correct, and the town of Rexford data is probably close to useless because the small sample size almost guarantees sampling bias

Community, Demography

Trego and the American Community Survey

michaelmccurryphd

Montana’s American Community Survey is composed from the final interviews conducted with 10,138 households in the state.  Since Montana has 519,935 households, the chance of any household being in the final interview is 10,138 out of 519,935 = 0.0195, right around 2% of the population is included in the survey.  Since Tregp shows 295 households, we can guess that our community data is assembled from somewhere around 6 completed interviews.

This table about sample sizes is from kenpro.org

As you will note, a population of 250 calls for 152 samples, and that the break is in the lower right corner – while a sample of 382 covers a population of 75,000, 384 is good for a million.  I’m not going into details about sampling – this is a blog, not a stats class. If enough people ask for the stats instruction, I’ll do another article.  Suffice to say, we can expect the numbers on the ACS to be pretty vague in small communities. So let’s look at the ACS data: data.census.gov

The top of the page shows:

Total Population                                           515

Median Household Income                          $36,458

Bachelor’s Degree or Higher                       26.1%

Employment Rate                                        40.2%

Total Housing Units                                      283

Without Health Care Coverage                   9.0%

Total Households                                         249

As we go further down the page, we start to encounter the variance – the range the number represents.  That 515 population is taken from the decennial census not the ACS. 

Median Age:    60.5 +/- 3.2 years 

16.8% of Trego folks speak a language other than English at home – plus or minus 15.4% so that’s somewhere between 6 and 166.  Probably not a particularly useful piece of information.

That Median Household Income turns out to be plus or minus $8,980: the number can be as much as 24.6% off either way.  It can be as low as $27,478 or as high as $45,438.  As we move into the full chart on that number, we see that the number of households lists the margin of error as plus or minus 67.  Could be as low as 182 or as high as 316. 

That 26.1% of Trego residents with a bachelor’s degree or more has a 12% margin of error – it could be as low as 14.1% or as high as 38.1%,  It shows 8.1% of our residents holding graduate or professional degrees, but doesn’t give a margin of error there.

The school board will be pleased to know that 97.8% (plus or minus 6.6%) of our kids are enrolled in Kindergarten to 12th grade.  Might even surprise the County Superintendent.  Pretty sure some kids out there are home-schooled.

That 40.2% employment rate (+/- 12.9%) looks low – but I guess it fits right in with a median age over 60 and 31.1% (+/-9.4%) disability. 

And finally, there are 48 women 15 to 50 years old – but the margin of error is 38, so it translates to somewhere between 10 and 86.

The ACS data is good – but the sample for Trego was small, and not checking the limitations lets us make blunders.