I’m pretty sure this is wrong, what they cited is wrong.
Kaiser and Oswald basically show that there’s a monotonic relationship between subjective feelings and real world outcomes, and this is robust.
First of all, monotonicity and cardinality are completely different things and this is pivotal to what HLI claims are. I’m not entirely sure they know the difference—this is bad (!).
A second more subtle issue is that Kaiser and Oswald are based on pretty well defined, moderately severe life events in the UK and wealthy countries. As broad, nonspecific wisdom, we should have much less confidence taking this to other contexts.
I only had time to read the second article, but my guess is that I can make decisive complaints about the first too (80%).
I started clicking on more links, and it’s No es buena mis amigas.
So there’s links to a happiness think tank, and a link to a UK government program for wellbeing.
It’s great and excellent that people are using wellbeing as metrics as part of complex interventions...which is normal and has been done for decades everywhere.
But:
These don’t “use this approach”. That these are used as metrics for healthcare work and other policy outcomes, is not at all sufficient evidence that we can focus on interventions solely targeted to these metrics of wellbeing.
Also, these links take time to get anywhere, this is not a good smell.
This isn’t dispositive, but there is a major presentation issue in how would-be authoritative content is being cited. What do we think this is, AI safety? There should be clearer standards of evidence and argument.
I’m pretty sure this is wrong, what they cited is wrong.
Kaiser and Oswald basically show that there’s a monotonic relationship between subjective feelings and real world outcomes, and this is robust.
First of all, monotonicity and cardinality are completely different things and this is pivotal to what HLI claims are. I’m not entirely sure they know the difference—this is bad (!).
See https://www.pnas.org/doi/10.1073/pnas.2210412119#sec-2
A second more subtle issue is that Kaiser and Oswald are based on pretty well defined, moderately severe life events in the UK and wealthy countries. As broad, nonspecific wisdom, we should have much less confidence taking this to other contexts.
I only had time to read the second article, but my guess is that I can make decisive complaints about the first too (80%).
Both comments by this author seemed in bad faith and I’m not going to engage with them.
I started clicking on more links, and it’s No es buena mis amigas.
So there’s links to a happiness think tank, and a link to a UK government program for wellbeing.
It’s great and excellent that people are using wellbeing as metrics as part of complex interventions...which is normal and has been done for decades everywhere.
But:
These don’t “use this approach”. That these are used as metrics for healthcare work and other policy outcomes, is not at all sufficient evidence that we can focus on interventions solely targeted to these metrics of wellbeing.
Also, these links take time to get anywhere, this is not a good smell.
This isn’t dispositive, but there is a major presentation issue in how would-be authoritative content is being cited.
What do we think this is, AI safety?There should be clearer standards of evidence and argument.