Skip to main content

Section Understanding Uncertainty and Limitations

All data has limitations, and all conclusions involve some degree of uncertainty. Teaching students to identify and honestly communicate these limitations is crucial for developing scientific integrity and critical thinking skills.
Professor Kat Werner discusses the limitations and challenges that can arise when collecting data using complex social issues as examples.

Exploration 24. Try This Week: Limitations Detective.

Time needed: 15 minutes with any data investigation
The Process: After students have drawn conclusions from data, have them systematically consider limitations:
Sample Limitations: Who did we ask? Who didn’t we ask? Does our sample represent everyone we want to know about?
Method Limitations: How did we collect this data? What might have influenced people’s responses? What did we not measure?
Context Limitations: When did we collect this? Where? What was happening that might have affected the results?
Interpretation Limitations: What other explanations could there be for what we found? What don’t we know?
Elementary Application: After surveying favorite pizza toppings: “We only asked kids in our class, so this might not be true for other classes. We also only gave three choices, so we don’t know about other toppings people might like.”
Secondary Application: After analyzing homework time data: “This is self-reported data, so students might have over or under-estimated. We collected it on a Tuesday, which might not represent typical homework loads. We also didn’t account for different class schedules or extracurricular activities.”

Checkpoint 63.

Students survey 30 students in their lunch period about favorite music genres. They want to make claims about “teenagers” in general. What should they understand about this generalization?
Hint.
Think about how representative their sample is of the broader population they want to understand.
Solution.
Students should understand that their findings apply most directly to students similar to those they surveyed—same school, same area, same time period. To make claims about teenagers in general, they would need data from different schools, regions, and backgrounds. This teaches them to be precise about what their data can and cannot tell them.

Checkpoint 64.

When students work with small datasets (like their class of 25 students), what’s the most important thing for them to understand about their conclusions?
Hint.
Consider how much confidence we can have in patterns found in small samples.
Solution.
Students should understand that small samples can reveal interesting patterns worth investigating further, but conclusions should be tentative. A pattern found in 25 students might or might not appear in 250 students or in a different classroom. This teaches appropriate humility about what conclusions can be drawn from limited data.

Exploration 25. Building Uncertainty Vocabulary.

Help students develop precise language for expressing different levels of confidence in their conclusions:
High Confidence: “Our data strongly suggests...” “We found consistent evidence that...” “The pattern was clear...”
Moderate Confidence: “Our data suggests...” “There appears to be...” “We found some evidence that...”
Low Confidence: “Our data hints that...” “It’s possible that...” “We’re not sure, but...”
Teaching Tip: Have students justify their confidence level by referencing sample size, data quality, consistency of patterns, and potential alternative explanations.

Checkpoint 65.

Students find that classes with more windows have higher average test scores. Before concluding that natural light improves learning, what should they consider?
Hint.
Think about other differences between classrooms that might affect test scores.
Solution.
Students should brainstorm alternative explanations: Maybe windowed classrooms are newer and have better facilities overall, or are located in quieter parts of the building, or house smaller classes, or have more experienced teachers assigned to them. Teaching students to generate alternative explanations develops critical thinking and prevents overconfident conclusions.