How to Lie with Statistics Summary Quiz

Study Notes

Introduction

Biased samples can distort crime data, such as inferring crime rates from media coverage.
Anti-histamine claims about curing colds are misleading since colds naturally resolve themselves.

Chapter 1: Built in Bias

Yale Alumni survey data can misrepresent the population due to selective responses and small sample sizes.
People often lie about their magazine subscriptions, affecting reported sales.
Sample reliability is critical; poor sampling leads to inaccurate results.
Cancer survival data based on guesswork can mislead due to population movement affecting statistics.
Individuals who seek psychiatric help may not represent general neuroticism levels in the wider population.
Telephone and subscription access skew voting results towards wealthier and typically Republican demographics.
People tend to provide socially desirable responses in surveys, introducing bias.

Chapter 2: Well Chosen Average

The mean income is often inflated by high earners; median income can provide a more realistic picture.
Sample characteristics, like unemployment status, can significantly skew average wage figures.

Chapter 3: The Little Figures That Are Not There

Small sample sizes, such as 12 in an toothpaste efficacy study, can yield misleading results.
Rare conditions like polio can result in inconclusive vaccine efficacy studies due to insufficient cases.
Definitions of household size can be vague, affecting average occupants reported.
Statistics can be misleading based on presentation; emphasizing unavailable power supply can create negative connotations.
Graphs lacking numerical scales can exaggerate perceptions of financial increases.

Chapter 4

IQ scores have a margin of error that can misplace individuals within average ranges; context is key for interpretation.

Chapter 5: The Gee-Whiz Graph

Graph scales can misrepresent data trends, making small changes appear dramatic.
Selective data presentation, such as truncating y-axes, can mislead viewers about growth trends.

Chapter 6: One Dimensional Picture

Bar chart visualizations can distort data by creating unbalanced representations.
Publications have used exaggerated visuals to emphasize perceived increases in metrics like life expectancy.

Chapter 7: Semiattached Figure

Misinterpretations can arise from associative claims that fail to prove causation.
Terms like "26% more juice" lack context, making comparisons vague.
Observational statistics can be misleading; higher accident rates in certain conditions may reflect higher traffic instead of risk levels.
Contextual differences can skew comparative statistics, such as contrasting death rates in the navy versus the general population.
The term "semiattached figure" describes misleading statistics that misrepresent relationships.

Chapter 8: Post Hoc Fallacy

Correlation does not imply causation; deteriorating school performance may correlate with smoking but isn’t a definitive cause.
Post hoc fallacy refers to erroneous causal associations based on inadequate data.
Misleading correlations can arise from chance or third-factor influences, such as income and stock ownership.
Assumptions about higher income from college attendees overlook factors like socioeconomic background.
Cancer rates linked to milk consumption demonstrate inadequacies in sample representation and confounding variables.

Chapter 9: How to Statisticulate

Statisticulation involves misrepresentation through selective use of data.
Poor population representations in maps can exaggerate areas of importance.
Self-reported data, like sleep duration, can be inaccurate and subject to individual bias.
Percentage manipulation creates confusion regarding stat accuracy.
Misleading retail discounts can misconstrue true savings.
Productivity comparisons may not reveal true economic health if based on inflated revenue rather than hourly wages.
Composite graphs can obscure the independent movement of datasets, misleading interpretations of profits versus wages.

Chapter 10: How to Talk Back to a Statistic

Analyze potential biases—who conducted the study and their motivations can influence results.
Critical evaluation of response rates and sample selection is vital for understanding significance.
Missing context in statistics can lead to skewed interpretations.
Be wary of shifts in subject matter within reported figures; the conclusions may differ from the data.
Ensure statistical claims make logical sense—averages must be contextualized for meaningful insights.

Description

Test your understanding of the key concepts from 'How to Lie with Statistics' through this comprehensive chapter summary quiz. Explore biases in data and the misinterpretation of statistics using real-world examples. Perfect for anyone looking to grasp the subtleties of statistics in everyday life.