2010 census vs. 2013 data discrepancy - ideas?

[Coragus]

Hey all,

We have a local fire chief that is asking our agency's GIS to research an up-to-date number of housing units for his city. One of our GIS guys did so and came up with about 200 fewer housing units than the 2010 Census reports, even though 2013 population estimates suggest an overall population increase. There have been no de-annexations nor have there been 200 demolition permits issued in the last three years.

My hypothesis is that the Census Bureau counts the living quarters of the long term transient population, such as business people in extended stay hotels or residents of halfway houses, in their count of housing units. This population would be impossible to catch in a GIS analysis because such a structure would come up as "X-brand Motel".

It is known that said city does have at least one hotel that is used by advocacy agencies in the area for homeless shelter and post-rehab clients.

Does my hypothesis make sense? We owe an answer back to a pretty antsy fire chief. If not, what do you think is going on?

Thanks,

Coragus

[Wannaplan?]

Population and housing estimates available from governmental agencies and for-profit companies are rife with faulty numbers. I can't give you any advice on your hypothesis, other than it makes sense to me, and if you had to choose a number from the 2012 estimates or the numbers your GIS team gives you, I would without a doubt, go with your GIS team.

[terraplnr]

That seems like a good guess, based on the 2010 Census definition of a housing unit:

"A housing unit is a house, an apartment, a mobile home, a group of rooms, or a single room that is occupied (or if vacant, is intended for occupancy) as separate living quarters. Separate living quarters are those in which the occupants live and eat separately from any other persons in the building and which have direct access from the outside of the building or through a common hall."

I'm wondering if illegal units could be part of the problem, too.

[Cardinal]

I'm not sure I would agree. What may be more likely is a discrepency due to the estimation algorithms. Remember that they do not have someone sitting down to study your community individually and decide if a particular number is consistent with what they might see on the ground. Instead, there is a very small sampling of the population (that might not even include any participants from your community), from which they draw data to plug into the formulas. I often run into similar issues in small areas, which is why I often end up doing my own projections.

[Coragus]

I'm not sure I would agree. What may be more likely is a discrepency due to the estimation algorithms.

So, basically, you're criticizing the Census Bureau and the 2010 total? The GIS count isn't an acutal estimate, it's an actual count based on our most recent data that's coming up 200 units under the 2010 Census number.

Not harshing, just clarifying! Thanks.

[WSU MUP Student]

I would echo what Cardinal says. Additionally, if I recall correctly, I believe that places like halfway houses are counted as institutional and not as individual housing units.

What is the population of your community? Have you considered using a few years worth of American Community Survey data to see how the years compare to each other or to the decennial census numbers? I know that the ACS data has a much higher MOE than the decennial census data, but it is nice that you are able to compare multiple, consecutive years.

How large is your community (geographically)? Maybe contracting with a local university or community development agency to undertake a windshield/walking survey to physically count all of the residencies within the community would be worthwhile.

In the end, remember that the economic downturn has caused average household sizes to increase while the number of households has decreased as adult-children are not establishing their own households as early as they used to and are more likely to move back in with their parents over the past 5 years or so. Have their been a lot of demolitions while new construction has been slow to replace the old units? This could partially explain the population increase without a corresponding increase in housing

[terraplnr]

My hypothesis is that the Census Bureau counts the living quarters of the long term transient population, such as business people in extended stay hotels or residents of halfway houses, in their count of housing units. This population would be impossible to catch in a GIS analysis because such a structure would come up as "X-brand Motel".

It is known that said city does have at least one hotel that is used by advocacy agencies in the area for homeless shelter and post-rehab clients.

The 2010 Census said that “emergency and transitional shelters (with sleeping facilities) for people experiencing homelessness” are classified as “group quarters” instead of housing units.

http://www.census.gov/acs/www/Downlo...efinitions.pdf

But if the hotel you are talking about is more like a SRO hotel/building I think it would be classified as individual housing units, since they are effectively small apartments (quick Google search didn't give me the answer).

Don't know if this helps still, sorry. But it seems like for fire department purposes, the GIS total would be the one to work with.

[Cardinal]

The 2010 numbers will be the most accurate. Everything else is an estimate based on limited surveying and generic formulas. As WSU said, the margin of error is pretty large.

[mike gurnee]

The 2010 numbers will be the most accurate. Everything else is an estimate based on limited surveying and generic formulas. As WSU said, the margin of error is pretty large.

I agree. The most accurate estimate today would be the 2010 census, plus/minus building and demo permits. Factor in at least a 6 month time lag between the permit and occupancy.

[Cismontane]

yep.. use real census decennial data over small-sample ACS updates whenever possible.

[ColoGI]

I'm not sure I would agree. What may be more likely is a discrepency due to the estimation algorithms. Remember that they do not have someone sitting down to study your community individually and decide if a particular number is consistent with what they might see on the ground. Instead, there is a very small sampling of the population (that might not even include any participants from your community), from which they draw data to plug into the formulas. I often run into similar issues in small areas, which is why I often end up doing my own projections.

This.