“Statistical Notes § Dealing With All-Or-Nothing Unreliability of Data”, Gwern2014-07-17 (Haskell, JS, R, genetics, IQ, Bayes, causality, decision theory, order statistics, power analysis, survey; similar)⁠:

Miscellaneous statistical stuff

Given two disagreeing polls, one small & imprecise but taken at face-value, and the other large & precise but with a high chance of being totally mistaken, what is the right Bayesian model to update on these two datapoints? I give ABC and MCMC implementations of Bayesian inference on this problem and find that the posterior is bimodal with a mean estimate close to the large unreliable poll’s estimate but with wide credible intervals to cover the mode based on the small reliable poll’s estimate.

• Critiques
  • Failed Facebook Critiques
  • Correlation=Causation in Cancer Research
  • Aerobic vs Weightlifting
  • Moxibustion Mouse Study
• “Someone Should Do Something”: Wishlist of Miscellaneous Project Ideas
• Estimating Censored Test Scores
• The Traveling Gerontologist Problem
• Bayes Nets
  • Daily Weight Data Graph
  • Zeo Sleep Data
• Genome Sequencing Costs
• Proposal: Hand-Counting Mobile App for More Fluid Group Discussions
• Air Conditioner Replacement
  • Parameters
  • Cost-Benefit
    • Discounting
  • Sensitivity Analysis
• Some Ways of Dealing With Measurement Error
• Value of Information: Clinical Prediction Instruments for Suicide
• Bayesian Model Averaging
• Dealing With All-Or-Nothing Unreliability of Data
  • Binomial
    • Binomial With Binary Unreliability
      • ABC
      • Mixture
      • Weakening Heuristic?
• Dysgenics Power Analysis
  • Selection on SNPs
  • Mutation Load
  • Weaknesses
  • Genetic Data Availability
    • Proprietary
    • Public
• Power Analysis for Racial Admixture Studies of Continuous Variables
  • Sibling Power Analysis
  • Adoption Power Analysis
    • Mean Population European Ancestry & Population Standard Deviation
    • Power Simulation
• Operating on an Aneurysm
  • Risk
  • Expected Loss
    • QALY/DALY Adjustment
  • Cost-Benefit
• The Power of Twins: Revisiting Student’s Scottish Milk Experiment Example
• RNN Metadata For Mimicking Individual Author Style
• MCTS
• Candy Japan A/B Test
• DeFries-Fulker Power Analysis
• Inferring Mean IQs From SMPY/TIP Elite Samples
• Genius Revisited: On the Value of High IQ Elementary Schools
• Great Scott! Personal Name Collisions and the Birthday Paradox
• Detecting Fake (human) Markov Chain Bots
• Optimal Existential Risk Reduction Investment
• Model Criticism via Machine Learning
• Proportion of Important Thinkers by Global Region Over Time in Charles Murray’s Human Accomplishment
• Program for Non-Spaced-Repetition Review of past Written Materials for Serendipity & Rediscovery: Archive Revisiter
• On the Value of New Statistical Methods
• Bayesian Power Analysis: Probability of Exact Replication
• Expectations Are Not Expected Deviations and Large Number of Variables Are Not Large Samples
  • Founder Effects
• Oh Deer: Could Deer Evolve to Avoid Car Accidents?
• Evolution As Backstop for Reinforcement Learning
• Acne: a Good Quantified Self Topic
• Fermi Calculations
• Selective Emigration and Personality Trait Change
  • See Also
• The Most Abandoned Books on GoodReads
• Best Student Ever!
  • Record Value Problem
    • Doubling Interval
    • Harmonic Series
• Little’s Law in the Wild
• Tail Collapse

Similar Links: