# How can I apply statistical thinking to investigate claims?

###### Author

Statistics are powerful tools, but they can sometimes be misleading or misunderstood. By understanding a few key concepts and asking the right questions, you can become a savvy consumer of statistical information.

1) Sample size: Check the sample size whenever you see a claim based on a study or survey. A larger sample size usually means more reliable results. For example, a survey claiming “90% of people love a new product” might sound impressive, but you should be skeptical if it’s based on only 10 respondents.

Ask yourself: How large is the sample size? Is it representative of the population it claims to represent?

2) Correlation vs. causation: A classic pitfall is a confusing correlation (when two things happen together) with causation (when one thing causes another). For example, a news report might claim that “people who eat more ice cream have a higher risk of sunburn.” While these two events might be correlated, eating ice cream doesn’t cause sunburns – they both happen more frequently in hot, sunny weather.

Ask yourself: Does the study establish a causal relationship, or is it merely observing a correlation? Are there alternative explanations for the observed correlation?

3) Confirmation bias: People tend to seek out and remember information that supports their existing beliefs. For example, if you believe a certain diet is the key to weight loss, you might be more likely to notice and share articles that support that belief. Be aware of this tendency and actively look for information that challenges your assumptions.

Ask yourself: Am I only noticing statistics that support my existing beliefs? What evidence is there that contradicts or questions this claim?

4) Misleading visuals: Graphs and charts can sometimes be manipulated to emphasize or downplay certain information. For example, a bar chart might show a small difference between two groups as a large visual gap by changing the scale of the y-axis.

Ask yourself: Are the visuals accurately representing the data or distorting the information? Does the scale of the graph make sense?

5) Cherry-picking data: When people want to prove a point, they might selectively present data supporting their argument while ignoring data contradicting it. For example, a company might highlight one positive review of their product while ignoring the many negative ones.

Ask yourself: Is the data being presented selectively, or is it comprehensive? Are there other data points that might tell a different story?

6) Uncertainty and error: All measurements have some degree of error or uncertainty. Be cautious of claims that present data as absolute or definitive. For example, a study might find that “60% of people prefer chocolate over vanilla,” but with a margin of error of 5%, the true percentage could be anywhere from 55% to 65%.

Ask yourself: Is there a margin of error or uncertainty in the data? How does that affect the conclusions drawn from the data?

By considering these questions when you encounter statistical claims, you’ll be better equipped to critically evaluate the information and make more informed decisions. Statistics can be a powerful tool, but it’s essential to understand the context and limitations of the data presented.