Announcement

Collapse
No announcement yet.

Study: UBI "doesn't work".

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Study: UBI "doesn't work".

    So where is the author and the study wrong?

    https://www.theguardian.com/commenti...rty-inequality

    http://www.world-psi.org/sites/defau...eport_2019.pdf

  • #2
    First of its not A study but a report on the effectiveness of many studies.

    I really don't want to get into the weeds on this because it requires allot explanation of statistical systems including difficult ideas like 'Discreet Math' and 'Regression Toward the Mean'.

    So all your going to get is the tl:dr summery.

    Studies listed in paper looked at the 'micro economy' effects. Author(s) are trying to use those conclusions to predict a 'macro economy' results. Mostly by using bad statistical match to boost results to build some kind of statistical mean to jump off from.

    What makes it even more blatant is that each of the listed studies do not even have the same or even close to similar testing parameters. So that makes their control groups wildly different. Hence the issue of 'Regression Toward the Mean' in extrapolating a combined conclusion.

    Comment


    • #3
      Ok, but what is the mean, there? If the report on the studies is anywhere near accurate (and based upon your comments, I think you believe it's inaccurate), then what is that mean?

      Wouldn't there be a negative ROI and/or a negative utility at some point with UBI?

      Comment


      • #4
        1. Most of the smaller studies have different goals and the results from said studies should not be used to extrapolate out a larger effect. UBI in a single towns results would only be good for that town. You will need to redo the experiment the same way in many different areas to build a true mean for each of the specific criteria you are testing.

        2. Regression towards the mean becomes a bigger problem until it is statistically removed. When any experiment that is done once and has any potential randomness needs to be redone many times. If I was the first person to do an experiment on chance of flipping a quarter to heads, and out of 20 flips all come up heads. I would then support an erroneous conclusion that coin flips always come up heads.

        3. Most of the smaller experiments were to test where the money went in the conditions set by each study. Most of the studies do no match each other, with a few having overlapping goal. Long story short, they can cherry pick results.

        Comment


        • #5
          Even UBI "experiments" in towns are being canceled because they don't appear to be working. One country even did it for a short while, and then canceled it because (I believe) ROI wasn't what they expected. It "didn't work" like they thought it would.

          Comment


          • #6
            Originally posted by mjr View Post
            Even UBI "experiments" in towns are being canceled because they don't appear to be working. One country even did it for a short while, and then canceled it because (I believe) ROI wasn't what they expected. It "didn't work" like they thought it would.

            Finland Just did not renew the program, as far as I know its still running to 2020.

            Ontario did cancel early.

            France has been doing it since 1988

            Cyprus since 2013

            Neither have released any findings other then some broad strokes and anecdotal.

            There is much political pressure on both sides of the programs that did interfere with the studies. Both studies were also tests of particular ways to deploy such a program.

            I think you are under an impression I don't have a strong opinion on the subject right now. I don't, welfare is part of our society now. No mater what lables you place on it. It solves allot of issues, even if a minority (29%) of people hate the idea of 'money for nothing'.

            The truth of any situation resists simplicity. I am however, intrigued by the idea from a sociological aspect.

            And from that aspect I can tell you this idea is going to require ALLOT of testing. Anything that deals with human behavior has so many random variables that need to be accounted for. Tweaking the study, try again.

            Comment

            Working...
            X