Latest / This Week In Wellbeing Measurement / Mental Health, Machine Learning, Social Media, Higher Education, Quality Of Life

Transcript

Mental Health, Machine Learning, Social Media, Higher Education, Quality Of Life · This Week In Wellbeing Measurement

• 0:00
  Davis: If a report card says you’re doing great but everyday life feels harder, do you trust the report card?
• 0:06
  Jenny: I don't, at least not yet, because a clean dashboard can make messy lives look solved when rent, pollution, or local services are moving the other way.
• 0:16
  Davis: I admit I love a clean number, but only if it doesn't flatten the thing it's supposed to describe, because institutions fund what they can see.
• 0:24
  Jenny: So this week we're asking what better wellbeing measures reveal, starting with the old problem that growth alone can miss inequality, environmental damage, and huge differences from place to place... welcome to This Week In Wellbeing Measurement on paperboy.fm.
• 0:54
  Davis: This week the feed is smaller but sharper. We looked at 461 research hits, qualified 91 papers, and saw 417 unique authors across 35 countries.
• 1:06
  Jenny: That qualified count only edged up from 89 to 91, about 2 percent, but the raw hits fell from 672 to 461. So the story isn't more noise; it's a better yield, from about 13 percent qualified last week to about 20 percent this week.
• 1:24
  Davis: The map widened fast too. Country coverage went from 22 to 35 countries, up about 59 percent, and the leaders are still modest counts: the UK has 6 papers, China has 4, and the U.S. has 2.
• 1:38
  Jenny: Which makes me ask what changed. Is this broader indexing, a more global week in wellbeing measurement, or just better country metadata? With only one institution showing in the metadata, I wouldn't over-read institutional spread.
• 1:54
  Davis: Method-wise, this is a measurement week in the literal sense. Qualitative work leads with 16 papers, meaning interviews or close reading of experience; quantitative work and surveys have 11 each; machine learning and case studies have 6 each.
• 2:09
  Jenny: And the people mix matters. Of 417 authors, 70 are first-time authors, meaning first-ever paper in the metadata, 207 are emerging, and 140 are experienced; meanwhile the top themes are mental health, machine learning, and wellbeing, which fits the through-line: better metrics change who institutions can actually see.
• 2:32
  Davis: Alright, let's get into the papers with a big frame-setter: Desenvolvimento além do crescimento, a 2026 Revista ft paper by Thyago Vinícius Marques Oliveira and Luisa Cabral Santos. The title means development beyond growth, and the argument is that a place can produce more income while still getting worse on inequality, environmental damage, or regional imbalance.
• 2:55
  Davis: Their central point is simple but pretty consequential: GDP, or Gross Domestic Product, counts the value of goods and services produced, but it doesn't tell you who benefits, what gets depleted, or which territories are left behind. So if public policy treats GDP as the scoreboard, the authors argue it can miss social, environmental, and territorial impacts that are part of development itself.
• 3:19
  Jenny: So what would actually have to be measured before we could say a place is developing well, not just growing?
• 3:25
  Davis: They don't build a new index here. The method is qualitative and theoretical-analytical, meaning they review and interpret existing arguments rather than test new data, drawing on development and sustainability authors including Cardoso and Faletto from 1970, Escobar from 2005, and work on territoriality and dependency. That gives the critique some conceptual weight, but the support is moderate because it's not an empirical test of a replacement dashboard.
• 3:52
  Jenny: That's the right caveat, but it's also a useful opening move for this whole week. This sits squarely in the beyond-single-indicators thread: GDP can stay in the room, but if it's the only number on the wall, then public decisions can look tidy while the real trade-offs stay hidden.
• 4:10
  Jenny: That line about one number on the wall carries straight into this Italian paper, Italian regional variation in amenable mortality 2015/2019, because here the dashboard number is deaths that good, timely healthcare should often be able to prevent.
• 4:26
  Jenny: Conrado, Minutiello, Lenzi, Marraffa, and Gianino look across Italian regions from 2015 to 2019, and they use age-standardized death rates, meaning they adjust for regions having older or younger populations before comparing them. The striking finding is that each annual one percentage-point rise in outsourcing to for-profit private providers, poverty rate, and ethnic minority rate was associated with an annual rise in amenable mortality, while higher household net income was linked to a slight fall.
• 4:59
  Davis: If amenable mortality is the performance scorecard, how do we know whether it's measuring care quality, or just mapping wider inequality onto the health system?
• 5:08
  Jenny: That's the right pressure test, because they ran a multivariate regression, which means they put several predictors into one model at the same time, including income, poverty, education, minority share, public and private spending, and outsourcing for inpatient and outpatient services. That gives the regional pattern real weight across five years, but it's still observational and regional, so it doesn't prove outsourcing itself caused the mortality changes.
• 5:36
  Davis: So the practical takeaway isn't, privatization bad, case closed; it's that a health-system metric can steer policy only if it's read beside poverty, income, and who lives in the region. That's the metrics-steer-systems thread in a very concrete form, because a region could look like it has a hospital performance problem when part of the signal is also deprivation, access, and uneven social risk.
• 5:59
  Davis: That same warning about not mistaking a dashboard for a cause shows up in a totally different place: buildings. In Metrics for Adaptation and Resilience Risks for Non-Domestic Buildings, J. Palmer and colleagues look at England, Scotland, and Wales, where the building stock is old and the weather is getting less temperate.
• 6:17
  Davis: Plain version: they’re not just drawing flood maps or heat maps; they’re trying to count who is inside the risky buildings. They propose five metrics across overheating, flooding, and air-conditioned safe space, meaning places people could retreat to during a dangerous heat event, and they assess flooding risk both now and in 2080.
• 6:36
  Jenny: When the unit of analysis is a building, how do they get from floor area to people at risk? A school, a hospital, and a warehouse can have the same square metres and wildly different occupancy.
• 6:48
  Davis: They use the National Buildings Database for Great Britain, which pulls together records for non-domestic buildings, meaning offices, schools, hospitals, shops, and industrial sites rather than homes. Then they combine building type, floor area, typical occupancy, overheating vulnerability, flood exposure, and air-conditioning availability to estimate usual users at risk; the caveat is that this is strongest for non-domestic buildings, so it doesn’t tell the whole residential climate-risk story.
• 7:17
  Jenny: That’s the beyond-single-indicators thread at building scale. A flood zone alone doesn’t tell you whether it contains an empty storage unit, a packed primary school, or the only cooled public space nearby, and Great Britain’s unusually good mapping data gives planners a sturdier count than a climate map with scary colours.
• 7:38
  Paperboy.fm: This is the free version of the podcast. Subscribe at paperboy.fm to access a dozen different paper review podcasts for five dollars a month.