MegaMaester

SUBJECTUnderstanding Data

Statistics for Everyday Life

Statistics is one of the most misunderstood subjects, because it is usually taught as a collection of formulas rather than as a way of understanding uncertainty.

7 of 7 modules48 lessons~12-16h

Not started yet

Prefer the whole picture first? See the Statistics for Everyday Life study guide — full lesson outline, key terms, and practice on one page.

Understanding Data — Lessons

Modules in this subject

Concept map

How the core concepts in Statistics for Everyday Life relate to one another.

DataVariablePopulationSampleCategorical DataOrdinal DataNumerical DataDistributionHistogramBox PlotScatter PlotMeanMedianModeSkewOutlierRangeVarianceStandard DeviationRegression to the MeanProbabilityRandomnessIndependenceConditional ProbabilityBase RateGambler's FallacyConjunction FallacyRandom SamplingSelection BiasNon-response BiasMargin of ErrorCorrelationRegressionExtrapolationCausationRiskAbsolute RiskRelative RiskExpected ValueUncertaintyConfidence IntervalStatistical Significancep-valueEffect SizePractical ImportanceSurvivorship BiasCherry-picking
  • Populationsampled bySample
  • SampleestimatesPopulation
  • Datasummarized byStatistics
  • Distributiondescribed byMean
  • Distributiondescribed byMedian
  • Distributiondescribed byStandard Deviation
  • ProbabilitymodelsUncertainty
  • Correlationdoes not implyCausation
  • Statistical Significancediffers fromPractical Importance
  • Confidence IntervalexpressesUncertainty
  • Effect SizemeasuresPractical Importance
  • Variableis part ofData
  • Datais aCategorical Data
  • Datais aOrdinal Data
  • Datais aNumerical Data
  • HistogramshowsDistribution
  • Box PlotshowsDistribution
  • Scatter PlotshowsCorrelation
  • Modedescribed byDistribution
  • SkewaffectsMean
  • OutlieraffectsMean
  • Rangedescribed byDistribution
  • Variancerelates toStandard Deviation
  • Regression to the MeanexplainsOutlier
  • Randomnessis part ofProbability
  • Independenceis part ofProbability
  • Conditional ProbabilityinvolvesBase Rate
  • Gambler's FallacymisunderstandsIndependence
  • Conjunction FallacymisunderstandsProbability
  • Random SamplingproducesSample
  • Selection BiasunderminesSample
  • Non-response BiasunderminesSample
  • Margin of ErrorexpressesSample
  • RegressionextendsCorrelation
  • ExtrapolationunderminesRegression
  • Relative Riskcontrasts withAbsolute Risk
  • Riskmeasured byAbsolute Risk
  • Expected ValueinvolvesRisk
  • p-valuedefinesStatistical Significance
  • Survivorship BiasunderminesSample
  • Cherry-pickingunderminesData
  • Statisticsrelates toCritical Thinking
  • Statisticsrelates toScientific Thinking

Statistics for Everyday Life: frequently asked questions

Does the average always tell you what's typical?
Not always. The mean gets pulled by extreme values, so on a skewed distribution it can sit far from anyone's real experience. The median, the middle value, often describes the typical case better. Always ask to see the distribution behind a single average.
Will a bigger sample size fix a biased survey?
No. If some people can't appear in your sample, collecting more responses just measures the same skewed group more precisely. Size reduces random noise, not bias. A small representative sample beats a huge one that systematically leaves certain people out.
If a test is 99% accurate, does a positive result mean I'm 99% likely to have the condition?
No. When a condition is rare, most positives are false alarms. If only 1 in 1,000 people have it, a 99% accurate test flags roughly ten healthy people for every true case, so a positive can still be unlikely to be real. Base rates matter.
Does 'doubles your risk' mean the risk is now high?
Not necessarily. Doubling a tiny risk still leaves a tiny risk: 2 in a million instead of 1. Relative figures like 'twice as likely' are meaningless without the absolute baseline. Always convert them to plain numbers, such as chances per thousand people.
After a coin lands heads five times, are tails more likely next?
No. A fair coin has no memory, so each flip stays fifty-fifty regardless of the streak. Expecting independent events to self-correct after a run is the gambler's fallacy. Past results simply don't change the odds of the next independent event.
Does statistically significant mean the effect is big or important?
No. Significance only suggests an effect probably isn't pure chance; it says nothing about size. With a large enough sample, even a trivial difference can be significant. Ask for the effect size and confidence interval to judge whether the finding actually matters in practice.
What's the difference between percent and percentage points?
They're not the same. If support rises from 40% to 44%, that's a 4 percentage-point increase but a 10% relative increase. Confusing the two makes changes sound far bigger or smaller than they really are, so always check which one a statistic actually means.