4 ms·

facepalm Ugh, this is a horrible article. 45% of people found housing, but there was zero difference between the treatment group (given $1,000/month), a group

facepalm Ugh, this is a horrible article. 45% of people found housing, but there was zero difference between the treatment group (given $1,000/month), a group given a large lump sum, and a control group given $50 a month! https://www.denverbasicincomeproject.org/research https://www.denverbasicincomeproject.org/research A lot of people rotate in and out of homelessness, so we should expect some percentage to find housing after a year, even if nothing is done. Indeed, this seems like pretty strong evidence that giving more money does not help people find housing (at least in the range from $50/month to $1,000/month). They cite a decrease in public service costs, but again, there is no difference between treatment and control! See page 6 in this PDF: https://static1.squarespace.com/static/64f507a995b636019ef8853a/t/6671a2285cf8513e1383ed9d/1718723132992/FINAL_DBIP+Year+One+Executive+Summary.pdf https://static1.squarespace.com/static/64f507a995b636019ef88... They calculate "saved costs" by estimating the difference in costs between the start of the program (a year ago) and a year into the program (now). However, since they are tracking a particular group of people, some of them will find homes and jobs on their own, so we should expect costs for that cohort to trend down no matter what. (This wouldn't lower costs overall, because in equilibrium, they would be replaced with newly homeless people with higher costs. But this is a cohort study, so those people wouldn't be included here.)

AlbertCory 2y ago
Indeed, "compared to what?" is always the question you should ask. In this case, it's "compared to $50/month" or "compared to doing nothing."
Atotalnoob 2y ago
45% are housed, however 12% of the control already had housing compared to 6% of the 1000/mo people before the study started This means the delta is higher for the people receiving 1000/mo.
- roflulz 2y ago
  then this wasn't a good experiment at all then
  - lupire 2y ago
    It was a great experiment. It got the predetermined outcome that the investigators wanted, and it got headlines and major news media, and the rebuttals won't have nearly as wide audience.
- xigoi 2y ago
  That means the results are meaningless because the control group is not representative of the whole sample.
dghlsakjg 2y ago
The article is really badly written. There is no "control" group (per the actual paper). Just a group that was given a lesser intervention. The actual research paper is available, and while it suffers from a lack of a control group, it shows that there were indeed some pretty remarkable changes in the lives of participants, and that there were statistically significant differences between the low pay group and the high pay group (which... duh, $11,500 extra is going to make a huge difference). Of course, without a control group, it is pretty hard to say what the effect of the income is vs. other programs, self-help, external factors, etc. https://static1.squarespace.com/static/64f507a995b636019ef8853a/t/6671a15eec7a812dee108e7c/1718722914185/FINAL_DBIP+Year+One+Quantitative+Research+Report.pdf https://static1.squarespace.com/static/64f507a995b636019ef88...

AlbertCory 2y ago

Indeed, "compared to what?" is always the question you should ask. In this case, it's "compared to $50/month" or "compared to doing nothing."

Atotalnoob 2y ago

45% are housed, however 12% of the control already had housing compared to 6% of the 1000/mo people before the study started This means the delta is higher for the people receiving 1000/mo.

roflulz 2y ago
then this wasn't a good experiment at all then
- lupire 2y ago
  It was a great experiment. It got the predetermined outcome that the investigators wanted, and it got headlines and major news media, and the rebuttals won't have nearly as wide audience.
xigoi 2y ago
That means the results are meaningless because the control group is not representative of the whole sample.

dghlsakjg 2y ago

The article is really badly written. There is no "control" group (per the actual paper). Just a group that was given a lesser intervention. The actual research paper is available, and while it suffers from a lack of a control group, it shows that there were indeed some pretty remarkable changes in the lives of participants, and that there were statistically significant differences between the low pay group and the high pay group (which... duh, $11,500 extra is going to make a huge difference). Of course, without a control group, it is pretty hard to say what the effect of the income is vs. other programs, self-help, external factors, etc. https://static1.squarespace.com/static/64f507a995b636019ef8853a/t/6671a15eec7a812dee108e7c/1718722914185/FINAL_DBIP+Year+One+Quantitative+Research+Report.pdf https://static1.squarespace.com/static/64f507a995b636019ef88...