Question on Data Projections

12-03-2014 04:55 AM
New Contributor



I have a quick question.  How far out can you project out to and do you have any concerns on the accuracy of said projections the farther we get out from the official 2010 Census numbers?  This would be mostly concerning HH/Population counts, income projections, demographic projections.  For a hypothetical let’s assume I needed projections out to 2025, is that something doable and moreso what level of confidence would you have in projecting that far out?



Tags (3)
0 Kudos
3 Replies
Frequent Contributor

I'm not sure if this is helpful, but if you have any need for employment projections, the BLS has some datasets for this, at the national level:

Employment Projections Home Page

There's also population projections from the Census:

Population Projections - Main - People and Households - U.S. Census Bureau

0 Kudos
MVP Emeritus

That is really a question for the more informed over at CrossValidated ... the statistics discussion forum.  Anything else you get would be opinion and I wouldn't base any decisions on those unless you have sound evidence.  So for my opinion...if the trend were replicable with little deviation...I might project beyond tomorrow...but this never worked in the stock market for me so who am I to say

And look how often our governments are right...

0 Kudos
MVP Alum

Dan has a good point - with that much time, there could be considerable variability, calling into question the predictions.  A real-life example, year 2000 Census data projected to 2010 for New Orleans.  Then along comes Hurricane Katrina, which results in a huge demographic diaspora....

I don't envy the people who have to quantify that variability... (i.e. that is not an easy job).

Chris Donohue, GISP

0 Kudos