Hacker News new | ask | show | jobs
by themafia 142 days ago
So if you give people money they'll endeavor to tell you precisely what they think you want to hear. I would be much more impressed if this wasn't based solely on self-reported results.
5 comments

I agree, further study is warranted. They should do it again with a larger cohort for longer and with more objective measures.
We already know just based on the limited amount from the annual payment to all residents of Alaska from their state wide oil and mineral fund over the past 40 some odd years, there are a lot of studies - off top of my head some study found people who were unemployed were in aggregate slightly more likely to get work, because the money enabled them to look farther for jobs instead of just surviving, it also increased demand for goods and services provided by other Alaskans. Here's one of many that breaks things down from economist point of view https://home.uchicago.edu/~j1s/Jones_Alaska_2022.pdf
Larger and larger until we're just flat out testing UBI is fine by me. If you want to test UBI, there's no better way than implementing UBI.
> “Just one year after completing (the program), I’m in my own place, halfway through a business degree, focused on building a stable, secure foundation for my daughter and myself, and working toward becoming a nonprofit leader who supports her community.”

I know, right? She did all that just so she could give her social workers the feedback they wanted to hear! Those liberals are so dastardly!

Those sneaky freeloaders, materially improving their lives and futures just to “prove” that social programs work.
And then they went back to homelessness. Those social workers running the survey won't even notice how poisoned the data set is.
self reported results that may be confirmed, vs subjective self reports such as, i feel happy and motivated now.
Yeah, couldn't possibly be that there were any actually good results coming from this. That's just an impossibility.
I don't think the parent is even saying that, their point is pretty reasonable: having some objective measure for before and after in any study is more reliable than self-reporting, especially when the subject might be incentivized to lie.

The self reports might be totally true, but the study isn't as good as it might be.

Maybe I misunderstood the article, but... the participants would have had no incentive to lie?

They were going to get the money for the fixed period unconditionally. That was the point.

TBF the welfare queen trope is well trodden ground. I'm actually surprised to see it brought up more than once in a supposedly sophisticated forum.
I don't think that's what GP is saying. This report would be more believable and more objective if it would have other types of metrics than just self-reporting ones.

There is a kind of people that function by finding edge-cases, questioning the results and posing uneasy questions when presented with a situation. Some might call them "haters", or nit-pickers, but I think their way of thinking is useful to make sure we're not just being fed feel-good make-believe.

A lot of existing social assistance is wildly inefficient as it is. With proper calibration of expectations, I think most people would be thrilled to see even 1/3 of the target population meaningfully helped. The rest of cash giveaway is not "waste" in that scenario, it's the cost of helping the ones that do end up homed, working, and paying taxes... which then contribute back to lowering the net cost of the "waste".
If you pay people $1k to kill snakes, you'll end up with a lot of dead snakes, but you'll also end up with more snakes.

It's not good enough to prove that the solution to the problem works for one side. It could create a problem elsewhere, and easily a bigger problem than you had before.

It's definitely not a good enough answer to give people $1k and essentially ask them: did you like getting $1k?

That's not what happened. This is what they did:

> Oregon’s results confirm what we saw in New York: When you cover the real cost of shared housing directly for two years — and pair it with support — young people stay housed

That's very light on details.

I would hope we can assume with a non-trivial sample size that you will find at least some success cases.

That should not surprise anyone. It matters: how often did it pay off (not answered), how much did it pay off (housed after is a start, for how long, what other improvements would be good to know), was it worth it (presumably we could've given them $10M per month and got similar results, which clearly would not have been worth it), and how can you prove it doesn't create a worse problem elsewhere (the hard part).

People like to just assume that if you give people money there's no hidden side effects elsewhere. Giving money is good. Plain and simple. There can't be any bad involved. Well, there can.

Hold on - what exactly do you think they're lying about? Whether they're currently housed?

Because, imo, that's the headline result - 94% is a great success rate.

What I got from the article is that 94% reported being housed at the end of the period where they were receiving money.

But the real question for me is what happens 6 months to a year after the funding ends.

yes. This is a good data point, but a follow-on study is needed to see the regression rate and whether such aid is needed continuously or if doing so for a period of time is sufficient. Likely would depend on the reasons why they became homeless in the first place.
It could be. How did the control group do?