How to Read Charts and Statistics Without Being Fooled

You will learn to spot common tricks used in data visualization and statistical reporting. This course covers how scales can distort reality, why correlation is not causation, and how to identify biased samples. By the end, you will be able to look at a news graphic or a scientific claim and ask the right questions to find the truth. You will gain the tools to see through misleading numbers in everyday life.

How scales change the story

The learner will be able to identify when a chart axis is manipulated to make small changes look large.

Key takeaways

  • A truncated Y-axis starts at a number higher than zero to make small differences look much larger than they are.
  • Bar charts must always start at zero because the human brain uses the height of the bar to judge its total value.
  • Dual-axis charts can be manipulated by scaling the two axes differently to create a fake relationship between two sets of data.
  • The Lie Factor is a mathematical way to show how much a chart exaggerates the actual data.

Sources

The truth about averages

The learner will be able to choose the right type of average to describe a set of data.

Key takeaways

  • The mean is the arithmetic average and is best for symmetric data without extreme outliers.
  • The median is the middle value of a sorted list and is the most reliable choice for skewed data or data with outliers.
  • The mode is the most frequent value and is the only measure of central tendency that works for categorical data.
  • Skewed distributions pull the mean away from the center, making the median a better representation of the typical value.

Sources

Finding the hidden bias in samples

The learner will be able to explain how a small or non-random group can ruin a study.

Key takeaways

  • Selection bias happens when the way you pick a group makes it different from the general population.
  • Adding more people to a biased group makes the results more precise but does not make them more accurate.
  • Attrition bias occurs when people who disagree or struggle leave a group, making the remaining members look more successful or unified than they are.
  • Convenience sampling is useful for small tests and early ideas but cannot be used to make broad claims about everyone.

Sources

Why correlation is not causation

The learner will be able to distinguish between things that happen together and things that cause each other.

Key takeaways

  • Correlation measures how two things move together but does not explain why they do so.
  • A hidden third factor can make two unrelated things appear to have a connection.
  • Large datasets often contain accidental patterns that look like real relationships but are just coincidences.
  • Proving a cause requires controlled experiments and evidence that the trigger happens before the result.

Sources

Understanding risk and probability

The learner will be able to calculate the difference between absolute and relative risk in health news.

Key takeaways

  • Absolute risk is the actual number of people affected out of the total group.
  • Relative risk is a ratio that compares the risk of one group to another.
  • News headlines often use relative risk because large percentages sound more dramatic.
  • You calculate absolute risk reduction by subtracting the final risk from the starting risk.

Sources

Checking the latest data claims

The learner will be able to apply data literacy skills to evaluate recent news and social media graphics.

Key takeaways

  • Data is not neutral because human choices during collection and design introduce bias into the final numbers.
  • Visual manipulations like truncated axes and 3D effects can make small changes look much larger than they really are.
  • Complex mathematical models are not always more accurate and can fail to predict the future if they are overfitted to past noise.
  • Reliable data claims should include citations to primary sources and use data that is ideally less than one year old.

Sources