Watch in slide show mode to observe (modest) animation. comments questions: dan.kahan@yale.edu papers, etc: www.culturalcognition.net
Communicating Science in a Polluted Science Communication Environment www.culturalcognition.net Communicating Science in a Polluted Science Communication Environment Dan M. Kahan Yale University & many x 103 others
“Skin cream experiment”
“Skin cream experiment”
Two conditions
Correct interpretation of data rash decreases rash increases incorrect Numeracy score Lowess regression line.
Covariance & Numeracy Derived via logistic regression. Bars denote 0.95 CIs.
“Gun ban experiment”
“Gun control experiment”
Four conditions
Correct interpretation of data skin treatment Numeracy score Gun ban Lowess regression line.
Correct interpretation of data skin treatment Numeracy score Gun ban Numeracy score Lowess regression line.
Correct interpretation of data Liberal Democrats (< 0 on Conservrepub) Conserv Republicans (> 0 on Conservrepub) skin treatment Numeracy score Gun ban
Correct interpretation of data Liberal Democrats (< 0 on Conservrepub) Conserv Republicans (> 0 on Conservrepub) skin treatment Numeracy score Numeracy Score Gun ban
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases skin treatment rash increases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases 35%, ± 10 skin treatment rash increases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases rash increases rash decreases skin treatment rash increases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime increases crime increases crime decreases crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy 5%, ± 6 rash increases rash increases rash decreases rash decreases rash decreases rash decrease rash decreases rash increases rash increases skin treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime increases crime increases crime decreases crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases rash increases rash decreases rash decreases rash decreases rash decreases rash decreases rash increases rash increases skin treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime increases crime increases crime decreases crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases rash increases rash decreases rash decreases rash decreases rash decreases rash decreases rash increases rash increases skin treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime increases crime increases crime decreases crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases rash increases rash decreases rash decreases rash decreases rash decreases rash decreases rash increases rash increases skin treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime increases crime increases crime increases crime increases crime increases crime decreases crime decreases Gun ban crime increases crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases rash increases rash decreases rash decreases rash decreases rash decreases rash decreases rash increases rash increases skin treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime decreases crime increases crime increases crime decreases crime decreases crime decreases crime decreases crime decreases Gun ban crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy rash increases rash increases rash decreases rash decreases rash decreases rash decreases rash decreases rash increases rash increases skin treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% crime increases 80% 90% 100% crime decreases crime decreases crime increases crime increases crime increases crime increases crime increases crime decreases crime decreases crime decreases crime decreases crime decreases Gun ban crime increases crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy Avg. “polarization” on crime data for low numeracy partisans 25% (± 9%) Avg. “polarization” on crime data for high numeracy partisans 46% (± 17%) crime decreases crime decreases crime increases crime increases crime increases crime increases crime increases crime decreases crime decreases crime decreases crime decreases crime decreases Gun ban crime increases crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Bounded Rationality Thesis (BRT)
Conservative Republican “How much risk do you believe global warming pose to human health, safety, or prosperity?” BRT prediction Reality Conservative Republican Adapted from Kahan, D.M. Climate-Science Communication and the Measurement Problem. Advances in Political Psychology 36, 1-43 (2015).
Conservative Republican “How much risk do you believe global warming pose to human health, safety, or prosperity?” BRT prediction Reality Conservative Republican Adapted from Kahan, D.M. Climate-Science Communication and the Measurement Problem. Advances in Political Psychology 36, 1-43 (2015).
Conservative Republican “How much risk do you believe global warming pose to human health, safety, or prosperity?” BRT prediction Reality Conservative Republican shaded area denotes 0.95 CI Adapted from Kahan, D.M. Climate-Science Communication and the Measurement Problem. Advances in Political Psychology 36, 1-43 (2015).
Conservative Republican “How much risk do you believe global warming pose to human health, safety, or prosperity?” BRT prediction Reality Conservative Republican shaded area denotes 0.95 CI Adapted from Kahan, D.M. Climate-Science Communication and the Measurement Problem. Advances in Political Psychology 36, 1-43 (2015).
Conservative Republican BRT prediction Reality Conservative Republican shaded area denotes 0.95 CI Adapted from Kahan, D.M. Climate-Science Communication and the Measurement Problem. Advances in Political Psychology 36, 1-43 (2015).
Conservative Republican Conservative Republican BRT prediction Reality Liberal Democrat Conservative Republican Conservative Republican Actively Open-minded Thinking (z-score) shaded area denotes 0.95 CI Adapted from Kahan, D. M., Corbin, Res. & Politics (2016), doi: 10.1177/2053168016676705
Conservative Republican BRT prediction Reality Liberal Democrat Conservative Republican conservative Republican shaded area denotes 0.95 CI Adapted from Kahan, D. M., Advances in Pol. Psych, 36, 1-43 (2015), doi: 10.1111/pops.12244.
Conservative Republican BRT prediction Reality Conservative Republican shaded area denotes 0.95 CI Adapted from Kahan, D. M., Advances in Pol. Psych, 36, 1-43 (2015), doi: 10.1111/pops.12244.
Conservative Republican BRT prediction Reality Conservative Republican shaded area denotes 0.95 CI Adapted from Kahan, D. M., Advances in Pol. Psych, 36, 1-43 (2015), doi: 10.1111/pops.12244.
Monte carlo simulations Liberal Democrat (-1 SD on Conservrepub) high numeracy = 7 correct low numeracy = 3 correct Conserv Republican (+1 SD on Conservrepub) Low numeracy High numeracy Avg. “polarization” on crime data for low numeracy partisans 25% (± 9%) Avg. “polarization” on crime data for high numeracy partisans 46% (± 17%) crime decreases crime decreases crime increases crime increases crime increases crime increases crime increases crime decreases crime decreases crime decreases crime decreases crime decreases Gun ban crime increases crime increases crime decreases 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% probabilility of correct interpretation of data probabilility of correct interpretation of data
Bounded rationality thesis
not too little rationality . . .
not too little rationality . . . but too much
tragedy of the science communications commons
not too little rationality . . .
not too little rationality . . . but too much
tragedy of the science communications commons is not normal
tragedy of the science communications commons is pathological
tragedy of the science communications commons is pathological
= polluted science communication environment tragedy of the science communications commons = polluted science communication environment
Two strategies for communicating in a polluted science communication environment:
Two strategies for communicating in a polluted science communication environment: 1. Detoxification
Two strategies for communicating in a polluted science communication environment: 1. Detoxification 2. Adaptation
Two studies . . .
Two studies . . .
Two studies . . .
Two Channel Communication Strategy channel 1: content Risk Perception Information channel 2: meaning
Cultural Cognition Worldviews Hierarchy Individualism Communitarianism Egalitarianism
Cultural Cognition Worldviews Risk Perception Key Low Risk High Risk Hierarchy Environment: climate, nuclear Gays military/gay parenting hierarchical individualists hierarchical communitarians Marijuana legalization Guns/Gun Control cats/annoying varmints HPV Vaccination Individualism Communitarianism Gays military/gay parenting Environment: climate, nuclear Marijuana legalization Guns/Gun Control egalitarian individualists egalitarian communitarians HPV Vaccination cats/annoying varmints Egalitarianism
study_dismiss scale (α = 0.85)
Cultural Cognition Worldviews Risk Perception Key Low Risk High Risk Hierarchy dismiss study Individualism Communitarianism credit study Egalitarianism
Control Condition
Anti-pollution Condition
Geoengineering Condition
Two Channel Communication Strategy channel 1: content Risk Perception Information channel 2: meaning
Anti-pollution Condition
Study dismissiveness Study dismissiveness Study dismiss Study dismiss Hierarch Individ Study dismiss Hierarch Individ Egal Commun Egal Commun Credit Credit anti-pollution anti-pollution
Study dismissiveness Study dismissiveness Study dismiss Study dismiss Hierarch Individ Study dismiss Hierarch Individ Egal Commun Egal Commun Credit Credit anti-pollution anti-pollution
Geoengineering Condition
Study dismissiveness Study dismissiveness Study dismiss Study dismiss Hierarch Individ Study dismiss Hierarch Individ Egal Commun Egal Commun Credit Credit anti-pollution anti-pollution
Study dismissiveness Study dismissiveness Study dismiss Study dismiss Hierarch Individ Study dismiss Hierarch Individ Egal Commun Egal Commun Credit Credit anti-pollution anti-pollution
Polarization more polarization Study dismiss (z-score) less anti-pollution
Two Channel Communication Strategy channel 1: content Risk Perception Information channel 2: meaning
Two studies . . .
Two studies . . .
www.culturalcognition.net
State of the art “Science Curiosity”/“curiosity” Test of Science Related Attitudes (TOSRA) I am curious about the world in which we live I find it boring to hear about new ideas I would enjoy visiting a science museum at the weekend I would like to be given a science book as a present I get bored when watching science programs on TV
Sample self-report item
Performance measure
“Science Curiosity Scale” (SCS_1.0 ) (item response theory 2PL) Science Curiosity Scale (SCSI) Science Curiosity Scale (SCS)
YIF Clip: Origins of color vision
Robustness of SCS Darwin’s Dangerous Idea Mass Extinction: Life at the Brink
SCS, percentile SCS, percentile
Robustness of SCS Darwin’s Dangerous Idea Mass Extinction: Life at the Brink Daily Hollywood Rundown
SCS, percentile SCS, percentile
SCS, percentile SCS, percentile
Q. Are “curious” partisans more likely to examine surprising contrary evidence?
Q. Are “curious” partisans more likely to examine surprising contrary evidence? 1. believer unsurprising vs. skeptical surprising 2. believer surprising vs. skeptical unsurprising
Q. Are “curious” partisans more likely to examine surprising contrary evidence? 1. believer unsurprising vs. skeptical surprising 2. believer surprising vs. skeptical unsurprising
Probability of selecting surprising skeptical vs Probability of selecting surprising skeptical vs. unsurprising believer story .25 .5 .75 1
below avg. science curiosity Probability of selecting surprising skeptical vs. unsurprising believer story 24% (± 18%) Liberal Dem. below avg. science curiosity
above avg. science curiosity below avg. science curiosity Probability of selecting surprising skeptical vs. unsurprising believer story 24% (± 18%) 68% (±20%) Liberal Dem. above avg. science curiosity Liberal Dem. below avg. science curiosity
below avg. science curiosity above avg. science curiosity Probability of selecting surprising believer vs. unsurprising skeptical story .25 .5 .75 1 42% (± 13) 62% (± 12%) Conser. Repub. below avg. science curiosity Conser. Repub. above avg. science curiosity
How can the two-channel model and science curiosity be used to fight a polluted science communication environment?
Science curiosity as adaptation strategy. . . shifting the curve Science Curiosity Scale
How can the two-channel model and science curiosity be used to fight a polluted science communication environment? You tell me!