Kim Doell
@kimdoell.bsky.social
Social and environmental psychologist/neuroscientist. Environmental Collective Behaviour (ECo) Group Leader, Uni Konstanz. Banjolele enthusiast. Failed painter.
This unconference is presented by
October 1, 2025 at 2:20 PM
This unconference is presented by
Where: online!
When: 7th of Oct. at 3pm CEST/ 9am EDT
Want to join? Sign up here (registration is by donation): bigteamscienceconference.github.io
When: 7th of Oct. at 3pm CEST/ 9am EDT
Want to join? Sign up here (registration is by donation): bigteamscienceconference.github.io
Big Team Science Conference
The fourth annual Big Team Science Conference will be held virtually via Zoom October 6-8, 2025. The goal of this three-day virtual conference is to bring together a multidisciplinary group of researc...
bigteamscienceconference.github.io
October 1, 2025 at 2:20 PM
Where: online!
When: 7th of Oct. at 3pm CEST/ 9am EDT
Want to join? Sign up here (registration is by donation): bigteamscienceconference.github.io
When: 7th of Oct. at 3pm CEST/ 9am EDT
Want to join? Sign up here (registration is by donation): bigteamscienceconference.github.io
With @epronizius.bsky.social, @monabielig.bsky.social, @clauslamm.bsky.social, @protzko.bsky.social, Olena Vitkovska, and Celina Kacperski, we'll explore the ethical tensions, institutional constraints, and political risks of doing science across borders—especially in times of war or crisis.
October 1, 2025 at 2:20 PM
With @epronizius.bsky.social, @monabielig.bsky.social, @clauslamm.bsky.social, @protzko.bsky.social, Olena Vitkovska, and Celina Kacperski, we'll explore the ethical tensions, institutional constraints, and political risks of doing science across borders—especially in times of war or crisis.
BIG thanks to the coauthors: Lukas Lengersdorff,
@shawnrhoadsphd.bsky.social,
@todorova.bsky.social,
@jonasnitschke.bsky.social, Jamie Druckman,
@madalina.bsky.social, The Many Labs Climate Consortium (i.e., the academic expert forecasters),
@clauslamm.bsky.social, and @jayvanbavel.bsky.social
@shawnrhoadsphd.bsky.social,
@todorova.bsky.social,
@jonasnitschke.bsky.social, Jamie Druckman,
@madalina.bsky.social, The Many Labs Climate Consortium (i.e., the academic expert forecasters),
@clauslamm.bsky.social, and @jayvanbavel.bsky.social
May 16, 2025 at 12:46 PM
BIG thanks to the coauthors: Lukas Lengersdorff,
@shawnrhoadsphd.bsky.social,
@todorova.bsky.social,
@jonasnitschke.bsky.social, Jamie Druckman,
@madalina.bsky.social, The Many Labs Climate Consortium (i.e., the academic expert forecasters),
@clauslamm.bsky.social, and @jayvanbavel.bsky.social
@shawnrhoadsphd.bsky.social,
@todorova.bsky.social,
@jonasnitschke.bsky.social, Jamie Druckman,
@madalina.bsky.social, The Many Labs Climate Consortium (i.e., the academic expert forecasters),
@clauslamm.bsky.social, and @jayvanbavel.bsky.social
📄 Full preprint here: osf.io/preprints/ps...
Would love your thoughts and feedback! #openscience #climatepsych #forecasting
Would love your thoughts and feedback! #openscience #climatepsych #forecasting
OSF
osf.io
May 16, 2025 at 12:46 PM
📄 Full preprint here: osf.io/preprints/ps...
Would love your thoughts and feedback! #openscience #climatepsych #forecasting
Would love your thoughts and feedback! #openscience #climatepsych #forecasting
Takeaway: If we want to improve behavioral science and intervention design, we need better ways of evaluating expert judgment—and clearer benchmarks.
Forecasting experiments also give unique insights into how experts (and nonexperts) think and act!
Forecasting experiments also give unique insights into how experts (and nonexperts) think and act!
May 16, 2025 at 12:46 PM
Takeaway: If we want to improve behavioral science and intervention design, we need better ways of evaluating expert judgment—and clearer benchmarks.
Forecasting experiments also give unique insights into how experts (and nonexperts) think and act!
Forecasting experiments also give unique insights into how experts (and nonexperts) think and act!
So what does this mean?
➡️ Being an expert helps—but it doesn’t guarantee accuracy.
➡️ Predicting behavioral outcomes is especially hard.
➡️ And heuristics can be more useful than expected.
➡️ Being an expert helps—but it doesn’t guarantee accuracy.
➡️ Predicting behavioral outcomes is especially hard.
➡️ And heuristics can be more useful than expected.
May 16, 2025 at 12:46 PM
So what does this mean?
➡️ Being an expert helps—but it doesn’t guarantee accuracy.
➡️ Predicting behavioral outcomes is especially hard.
➡️ And heuristics can be more useful than expected.
➡️ Being an expert helps—but it doesn’t guarantee accuracy.
➡️ Predicting behavioral outcomes is especially hard.
➡️ And heuristics can be more useful than expected.
We also looked at who tends to be a better forecaster.
The only consistent predictor across outcomes? Age.
Older participants were more accurate.
Other traits (e.g., open-mindedness, political orientation) mattered for beliefs and policy—but not behavior.
The only consistent predictor across outcomes? Age.
Older participants were more accurate.
Other traits (e.g., open-mindedness, political orientation) mattered for beliefs and policy—but not behavior.
May 16, 2025 at 12:46 PM
We also looked at who tends to be a better forecaster.
The only consistent predictor across outcomes? Age.
Older participants were more accurate.
Other traits (e.g., open-mindedness, political orientation) mattered for beliefs and policy—but not behavior.
The only consistent predictor across outcomes? Age.
Older participants were more accurate.
Other traits (e.g., open-mindedness, political orientation) mattered for beliefs and policy—but not behavior.
That heuristic?
Just assume the interventions do nothing. No effect.
It turns out this "do nothing" model was surprisingly hard to beat—especially when predicting real behavior.
Just assume the interventions do nothing. No effect.
It turns out this "do nothing" model was surprisingly hard to beat—especially when predicting real behavior.
May 16, 2025 at 12:46 PM
That heuristic?
Just assume the interventions do nothing. No effect.
It turns out this "do nothing" model was surprisingly hard to beat—especially when predicting real behavior.
Just assume the interventions do nothing. No effect.
It turns out this "do nothing" model was surprisingly hard to beat—especially when predicting real behavior.
How did they do?
▶️ Academics were more accurate than the public—especially for belief and policy outcomes.
▶️ But their predictions were less accurate for behavior.
▶️ However, nobody outperformed a simple heuristic model.
▶️ Academics were more accurate than the public—especially for belief and policy outcomes.
▶️ But their predictions were less accurate for behavior.
▶️ However, nobody outperformed a simple heuristic model.
May 16, 2025 at 12:46 PM
How did they do?
▶️ Academics were more accurate than the public—especially for belief and policy outcomes.
▶️ But their predictions were less accurate for behavior.
▶️ However, nobody outperformed a simple heuristic model.
▶️ Academics were more accurate than the public—especially for belief and policy outcomes.
▶️ But their predictions were less accurate for behavior.
▶️ However, nobody outperformed a simple heuristic model.
We tested four groups:
Academics (N = 242)
Government officials (N = 23)
Climate communicators (N = 23)
General public (N = 574)
We then compared their predictions to actual results from a nationally representative U.S. sample (N = 6,954).
Academics (N = 242)
Government officials (N = 23)
Climate communicators (N = 23)
General public (N = 574)
We then compared their predictions to actual results from a nationally representative U.S. sample (N = 6,954).
May 16, 2025 at 12:46 PM
We tested four groups:
Academics (N = 242)
Government officials (N = 23)
Climate communicators (N = 23)
General public (N = 574)
We then compared their predictions to actual results from a nationally representative U.S. sample (N = 6,954).
Academics (N = 242)
Government officials (N = 23)
Climate communicators (N = 23)
General public (N = 574)
We then compared their predictions to actual results from a nationally representative U.S. sample (N = 6,954).
Forecasters were asked to predict how 11 climate interventions would impact:
✅ Beliefs about climate change
✅ Support for climate policy
✅ A costly pro-environmental behavior
These weren’t hypotheticals—these were real interventions, with real data.
✅ Beliefs about climate change
✅ Support for climate policy
✅ A costly pro-environmental behavior
These weren’t hypotheticals—these were real interventions, with real data.
May 16, 2025 at 12:46 PM
Forecasters were asked to predict how 11 climate interventions would impact:
✅ Beliefs about climate change
✅ Support for climate policy
✅ A costly pro-environmental behavior
These weren’t hypotheticals—these were real interventions, with real data.
✅ Beliefs about climate change
✅ Support for climate policy
✅ A costly pro-environmental behavior
These weren’t hypotheticals—these were real interventions, with real data.
BIG thanks to all coauthors @todorova.bsky.social, David Steyrl, Matthew Hornsey, @cameronbrick.bsky.social Florian Lange, @jayvanbavel.bsky.social @madalina.bsky.social
May 9, 2025 at 8:25 AM
BIG thanks to all coauthors @todorova.bsky.social, David Steyrl, Matthew Hornsey, @cameronbrick.bsky.social Florian Lange, @jayvanbavel.bsky.social @madalina.bsky.social
📣 We hope this work helps refine climate models and guide global interventions by:
🔹 Prioritizing modifiable psychological research targets
🔹 Accounting for national context
🔹 Emphasizing outcome specificity
🔹 Prioritizing modifiable psychological research targets
🔹 Accounting for national context
🔹 Emphasizing outcome specificity
May 9, 2025 at 8:25 AM
📣 We hope this work helps refine climate models and guide global interventions by:
🔹 Prioritizing modifiable psychological research targets
🔹 Accounting for national context
🔹 Emphasizing outcome specificity
🔹 Prioritizing modifiable psychological research targets
🔹 Accounting for national context
🔹 Emphasizing outcome specificity
🧩 One size doesn’t fit all.
Public vs private, easy vs effortful behaviors are driven by different factors.
Designing effective interventions means targeting the right outcome with the right lever.
Public vs private, easy vs effortful behaviors are driven by different factors.
Designing effective interventions means targeting the right outcome with the right lever.
May 9, 2025 at 8:25 AM
🧩 One size doesn’t fit all.
Public vs private, easy vs effortful behaviors are driven by different factors.
Designing effective interventions means targeting the right outcome with the right lever.
Public vs private, easy vs effortful behaviors are driven by different factors.
Designing effective interventions means targeting the right outcome with the right lever.
⚠️ One of the most striking findings:
Political orientation strongly predicts beliefs and policy support,
but not actual behavior—and even predicts less info sharing.
Is polarization appears more psychological than behavioral?
Political orientation strongly predicts beliefs and policy support,
but not actual behavior—and even predicts less info sharing.
Is polarization appears more psychological than behavioral?
May 9, 2025 at 8:25 AM
⚠️ One of the most striking findings:
Political orientation strongly predicts beliefs and policy support,
but not actual behavior—and even predicts less info sharing.
Is polarization appears more psychological than behavioral?
Political orientation strongly predicts beliefs and policy support,
but not actual behavior—and even predicts less info sharing.
Is polarization appears more psychological than behavioral?
Explained variance ranged widely:
🔹 Belief: 57% 🥳
🔹 Policy support: 46%🍾
🔹 Info sharing: 74% accuracy🎉
🔹 Actual behavior: just 10%🫣
Private, effortful actions are harder to predict—likely influenced by unmeasured situational factors.
🔹 Belief: 57% 🥳
🔹 Policy support: 46%🍾
🔹 Info sharing: 74% accuracy🎉
🔹 Actual behavior: just 10%🫣
Private, effortful actions are harder to predict—likely influenced by unmeasured situational factors.
May 9, 2025 at 8:25 AM
Explained variance ranged widely:
🔹 Belief: 57% 🥳
🔹 Policy support: 46%🍾
🔹 Info sharing: 74% accuracy🎉
🔹 Actual behavior: just 10%🫣
Private, effortful actions are harder to predict—likely influenced by unmeasured situational factors.
🔹 Belief: 57% 🥳
🔹 Policy support: 46%🍾
🔹 Info sharing: 74% accuracy🎉
🔹 Actual behavior: just 10%🫣
Private, effortful actions are harder to predict—likely influenced by unmeasured situational factors.
People in lower-HDI countries showed stronger climate beliefs and behaviors—supporting the precarity hypothesis that less affluent nations, with fewer resources to buffer climate impacts, are more attuned to the need for action.
May 9, 2025 at 8:25 AM
People in lower-HDI countries showed stronger climate beliefs and behaviors—supporting the precarity hypothesis that less affluent nations, with fewer resources to buffer climate impacts, are more attuned to the need for action.
📊 Top 4 predictors consistently mattered across all outcomes:
✅ Environmentalist identity
✅ Trust in climate science
✅ Internal environmental motivation
✅ HDI
Most other predictors had inconsistent or even opposing effects (positive relationship with one outcome, neg with another).
✅ Environmentalist identity
✅ Trust in climate science
✅ Internal environmental motivation
✅ HDI
Most other predictors had inconsistent or even opposing effects (positive relationship with one outcome, neg with another).
May 9, 2025 at 8:25 AM
📊 Top 4 predictors consistently mattered across all outcomes:
✅ Environmentalist identity
✅ Trust in climate science
✅ Internal environmental motivation
✅ HDI
Most other predictors had inconsistent or even opposing effects (positive relationship with one outcome, neg with another).
✅ Environmentalist identity
✅ Trust in climate science
✅ Internal environmental motivation
✅ HDI
Most other predictors had inconsistent or even opposing effects (positive relationship with one outcome, neg with another).