Hacker News Reader

Comment by hnburnsy | Hacker News Reader

Maybe this isnt fair but we had demographers hired by our school district who could not reliably predict student population 2 years out. I watched them fail for 15 years.

View on HN

In many fields it's easier to predict the aggregate of the whole system than the quantity of any specific component. Weather, thermal behavior, and also population.

Some things are also just not predictable, like human behavior

If you go a level up from demography to biology, you'll see a very clear message that population might plateau, but then it will start going up again.

Just saying something and hoping it’s true doesn’t make it true

Sounds suspiciously like climate change deniers who think it's clever to point out we can't exactly predict tomorrow's temperature.

My unfair comment is probably more like "The Plural of Anecdote Is Not Data".

Sounds like you are a hammer and see nails everywhere.

Failed in what sense? Completely off, or close enough (like weather forecasting)?

Is it possible a school district couldn't afford the best services? Or that their goal wasn't necessarily the population number but rather an analyst of the causative forces and directional trends?

Excellent points, since school funding is student count dependent it was critical for it to be close. They mostly predicted increases but year after year it was flat or when decreasing. Maybe politics were involved, and like I said my comment probably isnt fair.

Yeah, "failing" for 15 years is a pretty good indication that those demographers are actually succeeding at what their employer wants... :)

someone in the system is probably paying for data that shows growth to protect their interests (eg coverying up insolvency of teacher pensions or justifying municipal debt/bonds)

It is difficult to make predictions, especially about the future.

I would imagine they were only looking at past trends of the school district population - but the reality is the most information would be USPS address changes, data that various companies like Zillow have, and new building permits.

The whole goal of science is the ability to make predictions, however small they are, about the future. Maybe with an uncertainty range.

Is the implication here that if they can't get that right, they wouldn't be able to predict world population either? If so, how do you know this is an issue with demography and not just being the ones your school hired not being particularly good?

Yup they might have not been good but they kept getting hired. In my mind I am thinking if they cant reliably predict a popualtion of around 5000, his could they predict the world. I get that they are apples and oranges.

I guess I just I wouldn't expect a local school district to be paying very much for a service like that, so I wouldn't expect that the results of who they hired would represent the best of the field.

I would politely suggest your school district got ripped off.

I am sure they were not getting their moneys worth.

Economic issues combine with demographics for school enrollment predictions, because houses become unaffordable for families pushing them out of the school district as the multi bedroom houses fill up with retirees.

Reply on news.ycombinator.com