How Not to be Wrong Excerpt Public Opinion Doesn't Exist 🌱

VotingPref

  • If we ask: should we cut spending or raise taxes? Then cut spending wins.
  • If we ask: should we cut medicare? 2/3 will say no.

The majority rules system seems fair and simple, but it really doesn’t make sense when deciding between more than 2 options

Look at Obamacare

In an October 2010 poll of likely voters, 52% of respondents said they opposed Obamacare, while only 41% supported it. Bad news for Obama? Not once you break down the numbers.

  • Outright repeal of the Obamacare reform was favored by 37%
  • 10% said the law should be weakened
  • 15% preferred to leave it as is
  • 36% said Obamacare should be expanded to change the current health care system more than it currently does

This suggests that many of the law’s opponents are to Obama’s left, not his right (36% said it should be expanded although 37%+10% think the policy is too far left).

Which of these news headlines are true:

  1. Majority of Americans oppose Obamacare!
  2. Majority of Americans want to preserve or strengthen Obamacare!

Looking at a Real Election

In Burlington, Vermont (also Australia) they use a different voting system (instant runoff voting)

  1. where you mark off your preferences in order (not just picking a single candidate).
  2. Start by counting the first-place votes, and eliminate the candidate who got the fewest.
  3. Then, using the new preferences count the remaining votes and decide the winner
Vote Count
Montroll, Kiss, Wright 1332
Montroll, Wright, Kiss 767
Montroll 455
Kiss, Montroll, Wright 2043
Kiss, Wright, Montroll 371
Kiss 568
Wright, Montroll, Kiss 1513
Wright, Kiss, Montroll 495
Wright 1289
  • Wright, the Republican, gets 3297 first-place votes in all
  • Kiss (democrat) gets 2982
  • Montroll (democrat) gets 2554.

This city in Vermont is pretty liberal, but if they followed the conventional voting scheme Wright the republican would have won.

What happened:

  1. Montroll, the Democrat, had the fewest first-place votes, so he was eliminated.
  2. Recounted vote: Kiss 4314, Wright 4064, and Kiss is reelected.

What’s wrong?

  • You can check that 4067 voters liked Montroll better than Kiss
    • 3477 liked Kiss better than Montroll.
  • 4597 voters preferred Montroll to Wright
    • only 3668 preferred Wright to Montroll

In other words, a majority of voters liked the centrist candidate Montroll better than Kiss but a centrist is very unlikely to win in this setup

Interesting Take Away

  • Traditional American voting method: Wright wins
  • Instant-runoff method: Kiss wins
  • Head-to-head matchups: Montroll wins

Now Think About This

  • Suppose those 495 voters who wrote β€œWright, Kiss, Montroll” had decided to vote for Kiss instead, leaving the other two candidates off their ballot
  • And let’s say 300 of the Wright-only voters switch to Kiss too
  1. Wright now has 2502 first place votes so he gets eliminated instead of Montroll
  2. The election then goes down to Montroll vs. Kiss, and Montroll wins, 4067–3777

Basically if we give Kiss more votes, he will end up losing the election.

Notes mentioning this note

There are no notes linking to this note.


Here are all the notes in this garden, along with their links, visualized as a graph.

5G and WiFiAWS Step FunctionsAnalyzing Reddit Post on the Dollar StandardAsync, Await, and PromisesBayesian AverageBias Variance DecompositionBlockchain PresentationBreakpoint Debugging in VSCodeBrief Look into Measure TheoryC4 Model for Software ArchitectureCache vs Session StoreCant compare mean and median from different setsClient vs Server Side RenderingCode Production in an AI CompanyComparing Client Side Storage MethodsComputational Perception HighlightsConfidence Intervals for Known Distributions and...Cool Stocks ListCrazy Meeting with Obama, McCain, and Bush Post...Curse of DimentionalityDatabase vs Data Warehouse vs Data LakeDifferent Git AddsDocker (containerization) vs Vagrant (virtual...Explaining Decision Boundary of a Support Vector...Exporting Databricks Files to GithubFloyds Tortoise and Hare AlgorithmFresh Mac Setup Installation EssentialsGraphical Model IndependenciesHighlights from Bad SamaritansHighlights from Good Economics for Hard TimesHighlights from The Righteous MindHow Does Chromosomal Heredity WorkHow Does Light Influence the Rate of Capture in a...How Does Sweating WorkHow Does Version Naming WorkHow Not To Be Wrong Excerpt Self Selecting BiasHow Not to be Wrong Excerpt Public Opinion Doesn't...How Quantum Computers Could Quickly Break...How Someone Made a Spectral Lamp that Can Emit all...How are images compressed and stored in a computerHow do SPACs WorkHow does Hypothesis Testing WorkHow does air slow objects downHow is Neural Network a Universal ApproximatorHow is Unit Testing DoneHow to Access a Previous Commit with GitHow to Add to Your System Path Variable for MacHow to Build a Full Stack ApplicationHow to Clear Unused Docker ContainersHow to Convert from Celsius to FahrenheitHow to Delete a Branch GithubHow to Export Pandas DataFrame to CSV ProperlyHow to Force Pull and Overwrite GitHow to Get the Bootstrapped Standard Error for a...How to handle violations in positivityHow to Properly Explain Technical ToolsHow to Push Code for ProductionHow to Read a Path in S3How to Set Up Python Aliasing In the Command LineHow to Set a Specific Branch to Track a Specific...How to Store and Access SQL Queries in DatabricksHow to Take a Weighted AverageHow to Temporarily Stash Changes with Git StashHow to Untrack Committed Files from GitHow to Use PyenvHow to Use Sample Splitting for Doubly Robust...How to Write Output to Text FileHow to edit Obsidian themes with CSSHow to make copies of DNA with PCRHow to use Bounds and Sensitivity Analysis in...How to use Scipy Optimize to solve for values when...Info on Stock OptionsInspirational Computer PioneersIntuition Behind the Doubly Robust EstimatorInverting Hypothesis TestsInvesting LessonsJupyter Widgets ExistML CheatsheetsMaking Sense of a Betting Market with...Managing Ruby Versions with rbenvMarket Makers and Quant TradingMarket Making PresentationMatching IntuitionMethodology for Managing Web AppsMicroservices vs Monolithic ArchitectureModeling Advice and Lessons Learned Working at a...Multinomial to Binomial Stick Breaking...Music Theory NotesNotes from Michael Nielsen Effective Research PostNotes from the Martian by Andy WeirNotes on Bayesian OptimizationNotes on Exon Skipping with ASOsNotes on Options SpreadsNotes on Quantum CountryOne Persons Perspective About Why We Shouldnt Read...Presentation on the Kronovet Family Clothing...Python Dataclasses UpdatePython Package Reference InstructionsRandom CMU Course WebpagesRandom Facts from What If by Randall MunroeReading about Internet ServicesRock Thrust ExplainedSSHing into AWS and Running ThingsSome Bash Commands to Find Redundant Files and...Some Cool Python FeaturesSome Notes on Exploding Gradient ProblemStats BlogsStock Options in a CompanyTesting Code on GithubThoughts after Reading Hillbilly ElegyThoughts on Andy Matuschak Article on Teaching...Thoughts on Approaching Infinite KnowledgeThoughts on Maria Konnikova Knowledge Project...Thoughts on the End of Natural SelectionTor Network and .Onion DomainsUsing nonparametric models in doubly robust...Various Treatment Effects and their...Virtual Environment in AWSWhat Database do I useWhat are Git Pull and Push RequestsWhat are Information CriteriaWhat are Javascript WorkersWhat are MakefilesWhat are Moment Generating FunctionsWhat are Multiple CPU CoresWhat are Progressive Web AppsWhat are Wasserstein and Earth Movers DistancesWhat are the Four Fundamental Forces in Our...What is Apache SparkWhat is Bootstrapping in StatisticsWhat is Cryptocurrency StakingWhat is ElasticsearchWhat is Express.jsWhat is GLUEWhat is GraphQLWhat is HTTPSWhat is IV CrushWhat is Integration ReallyJAMStackWhat is KubernetesWhat is Mahalanobis DistanceWhat is MakerDAO CryptoWhat is Markov Chain Monte Carlo SamplingWhat is Nested Cross ValidationWhat is Next.jsWhat is PAC LearningWhat is R SquaredWhat is RedisWhat is ShrinkageWhat is Spearman CorrelationWhat is SvelteWhat is TerraformWhat is The Graph (Blockchain)What is Variational InferenceWhat is Vue.jsWhat is WebAssemblyWhat is a Credible IntervalWhat is a Fourier transformWhat is a Gaussian Mixture ModelWhat is a Gaussian ProcessWhat is a Object Relational MapperWhat is a Qini CurveWhat is a Sufficient StatisticWhat is independent component analysisWhat is the C-Statistic for BenefitWhat is the Dirichlet ProcessWhat is the EM AlgorithmWhat is the Hidden Markov ModelWhat is the Indian Buffet ProcessWhat is the Naive Bayes algorithmWhat is the Negative Binomial DistributionWhat is the Runtime of a LanguageWhat is the Studentized BootstrapWhat is the Wake Sleep AlgorithmWhat is the hypergeometric distributionWhy are Conjugate Priors UsefulWhy are there 12 Notes in Western MusicWhy is Cross Fitting Useful for Estimating...Why is a room hotter when you leave the fridge...Working with ClientsWorking with Terminaldata science overviewhighlights from Debt The First 5000 Yearshighlights from Enlightenment Nowhighlights from Hacking Darwinhighlights from How Not to be Wronghighlights from Leonardo da Vincihighlights from Open an Autobiographyhighlights from Range Why Generalists Triumph in a...highlights from Salt, Fat, Acid, HeatSapiens a Brief History of Humankindhighlights from Stumbling on Happinesshighlights from The Genehighlights from Thinking Fast and Slowhighlights from Trick Mirror