The article by Cumming, Fidler, and Vaux discusses the proper use and interpretation of error bars in experimental biology. Error bars are commonly used in scientific publications to communicate data and assist in correct interpretation. They can represent confidence intervals, standard errors, standard deviations, or other quantities, each providing different types of information. The authors emphasize the importance of distinguishing between descriptive and inferential error bars. Descriptive error bars, such as range and standard deviation, show how data are spread, while inferential error bars, like standard error (SE) and confidence intervals (CIs), provide information about the precision of estimates and the likelihood of differences between groups. Key points include: 1. **Descriptive Error Bars**: Range and standard deviation (SD) show how data are spread. SD is calculated as the square root of the sum of squared deviations from the mean divided by the number of samples minus one. 2. **Inferential Error Bars**: SE and CIs are used to infer whether differences between groups are statistically significant. SE is defined as SD divided by the square root of the sample size (n), and CIs are adjusted by a factor (t) to capture the true mean μ on 95% of occasions. 3. **Rule 2**: The value of n (the number of independently performed experiments) must be stated in the figure legend. 4. **Rule 3**: Error bars and statistics should only be shown for independently repeated experiments, not for replicates. 5. **Rule 4**: For comparing experimental results with controls, inferential error bars (SE or CI) are usually appropriate, but for small n (e.g., n = 3), plotting individual data points is better. 6. **Rule 5**: 95% CIs capture μ on 95% of occasions, and SE bars can be doubled in width to get approximate 95% CIs if n ≥ 10. 7. **Rule 6**: When comparing SE bars or 95% CIs for two sets of results, the smaller the overlap of bars, the stronger the evidence for a true difference. 8. **Rule 7**: With 95% CIs and n = 3, overlap of one full arm indicates P ≈ 0.05, and overlap of half an arm indicates P ≈ 0.01. 9. **Rule 8**: In repeated measurements of the same group, CIs or SE bars are irrelevant to comparisons within the same group. The authors conclude that while error bars can be valuable, they should be used critically and interpreted with biological understanding.The article by Cumming, Fidler, and Vaux discusses the proper use and interpretation of error bars in experimental biology. Error bars are commonly used in scientific publications to communicate data and assist in correct interpretation. They can represent confidence intervals, standard errors, standard deviations, or other quantities, each providing different types of information. The authors emphasize the importance of distinguishing between descriptive and inferential error bars. Descriptive error bars, such as range and standard deviation, show how data are spread, while inferential error bars, like standard error (SE) and confidence intervals (CIs), provide information about the precision of estimates and the likelihood of differences between groups. Key points include: 1. **Descriptive Error Bars**: Range and standard deviation (SD) show how data are spread. SD is calculated as the square root of the sum of squared deviations from the mean divided by the number of samples minus one. 2. **Inferential Error Bars**: SE and CIs are used to infer whether differences between groups are statistically significant. SE is defined as SD divided by the square root of the sample size (n), and CIs are adjusted by a factor (t) to capture the true mean μ on 95% of occasions. 3. **Rule 2**: The value of n (the number of independently performed experiments) must be stated in the figure legend. 4. **Rule 3**: Error bars and statistics should only be shown for independently repeated experiments, not for replicates. 5. **Rule 4**: For comparing experimental results with controls, inferential error bars (SE or CI) are usually appropriate, but for small n (e.g., n = 3), plotting individual data points is better. 6. **Rule 5**: 95% CIs capture μ on 95% of occasions, and SE bars can be doubled in width to get approximate 95% CIs if n ≥ 10. 7. **Rule 6**: When comparing SE bars or 95% CIs for two sets of results, the smaller the overlap of bars, the stronger the evidence for a true difference. 8. **Rule 7**: With 95% CIs and n = 3, overlap of one full arm indicates P ≈ 0.05, and overlap of half an arm indicates P ≈ 0.01. 9. **Rule 8**: In repeated measurements of the same group, CIs or SE bars are irrelevant to comparisons within the same group. The authors conclude that while error bars can be valuable, they should be used critically and interpreted with biological understanding.