← Back to Insights

Module: Data Literacy Basics

By SAUFEX Consortium 23 January 2026

[screen 1]

A chart shows crime rates “skyrocketing.” You share it, concerned. Later, you learn the vertical axis was manipulated to exaggerate the increase - actual change was minimal.

Data seems objective, but how it’s presented profoundly affects meaning. Basic data literacy helps you evaluate statistical claims critically.

[screen 2]

Why Data Literacy Matters

Modern discourse is saturated with data: charts, statistics, polls, studies. These seem authoritative - harder to dispute than opinions.

This makes data a powerful manipulation tool:

  • Statistics are perceived as objective facts
  • Most people struggle to evaluate data critically
  • Visualizations can mislead even when numbers are accurate
  • Small sample studies are presented as definitive
  • Correlation is confused with causation

[screen 3]

Reading Data Visualizations

Charts and graphs translate numbers into visual patterns. Design choices profoundly affect interpretation:

Y-axis manipulation: Starting the vertical axis at a number other than zero can exaggerate changes

Selective timeframes: Choosing start/end dates can hide or emphasize trends

Scale distortion: Using non-linear scales without clear indication

Missing context: Showing raw numbers without population adjustments or baselines

Always check axes, scales, and timeframes before accepting visual data.

[screen 4]

Common Chart Deceptions

Truncated axes: Chart shows dramatic spike, but axis starts at 95 instead of 0 - actual change is small

Cherry-picked timeframes: “Sales doubled!” - true, but only over one unusual week

Misleading categories: Grouping data in ways that support desired narrative

3D effects: Making bars appear different sizes through visual distortion

Dual axes: Two different scales on one chart, making unrelated things seem correlated

[screen 5]

Understanding Surveys and Polls

Not all surveys are equal. Key questions:

  • Sample size: How many people? (Larger is generally better)
  • Sampling method: Random selection or self-selected respondents?
  • Margin of error: How much statistical uncertainty?
  • Question wording: Leading questions produce biased results
  • Response rate: What percentage completed the survey?

“A poll shows…” means little without this context.

[screen 6]

Survey Red Flags

Self-selection bias: Online polls where anyone can respond aren’t representative

Small samples: “4 out of 5 dentists” might be just 5 dentists total

Loaded questions: “Do you support protecting children from dangerous content?” vs “Do you support internet censorship?”

Missing methodology: Credible polls explain their methods

Legitimate survey organizations (Pew, Gallup, national statistics offices) publish detailed methodology.

[screen 7]

Correlation vs. Causation

One of the most common statistical manipulation tactics: implying that because two things are correlated, one causes the other.

Example: “Countries with more chocolate consumption win more Nobel Prizes.” True correlation, but absurd to conclude chocolate creates Nobel laureates.

Many factors correlate without causal relationships. Always ask: could something else explain both? Is this just coincidence?

[screen 8]

Statistical Significance

Research findings often cite “statistical significance” - but what does this mean?

It means the result is unlikely due to random chance alone. It does NOT mean:

  • The effect is large or important
  • The study proves causation
  • The finding will replicate in other studies
  • The result has practical real-world significance

“Statistically significant” has a specific technical meaning often misunderstood by media.

[screen 9]

Evaluating Research Claims

When media reports “a study shows…”:

  • Who funded it? Tobacco-funded studies downplayed smoking risks
  • Peer-reviewed? Published in credible journals or just press released?
  • Sample size and duration? Small, short studies are preliminary
  • Replicated? Single studies are just starting points
  • Conflicts of interest? Researchers with financial stakes in results

Treat single studies as suggestive, not definitive.

[screen 10]

Data in the Wild: Real-World Tactics

How manipulators misuse data:

Absolute vs. relative numbers: “Crime increased 50%!” (from 2 incidents to 3)

Base rate neglect: Ignoring how common something is overall

Percentage point confusion: “Increased from 1% to 2%” is 1 percentage point but 100% relative increase

Survivor bias: Only counting successes, ignoring failures

Simpson’s Paradox: Trend reverses when data is disaggregated

Understanding these tactics helps you spot manipulation.

[screen 11]

Practicing Data Skepticism

Before accepting data claims:

  1. Check the source - who collected this data?

  2. Look for methodology - how was it gathered?

  3. Consider sample size and selection

  4. Check axes and scales on visualizations

  5. Ask what’s not shown - what data is omitted?

  6. Look for conflicts of interest

  7. Seek independent verification

  8. Remember: data can be accurate but misleading

[screen 12]

Building Data Literacy

You don’t need advanced statistics to evaluate data claims. Basic critical questions suffice:

  • Does the visualization fairly represent the numbers?
  • Is the sample representative and large enough?
  • Are alternative explanations possible?
  • Who benefits from this interpretation?
  • What context is missing?

As data becomes central to public discourse, these questions become essential citizenship skills. Data literacy isn’t about math - it’s about critical thinking applied to numbers.