Key Points from the Appendix 289 too small or the difference was too modest, we didn't observe a statistically significant difference in our test. Chapters 6 and 7 contain a thorough discussion of power and computing sample sizes to control Type II errors. A discussion about the importance of balancing Type I and Type II errors for applied research appears in Chapter 9. If you need more background and exposure to statistics, we've put together interactive lessons with many visualizations and examples on the www.measuringusability.com website. KEY POINTS FROM THE APPENDIX · · · · · Pay attention to the type of data you're collecting. This can affect the statistical procedures you use and your interpretation of the results. You almost never know the characteristics of the populations of interesting data, so you must infer the population characteristics from the statistics you calculate from a sample of data. Two of the most important types of statistics are measures of central tendency (e.g., the mean, median, and geometric mean) and variation (e.g., the variance, standard deviation, and standard error). Many metrics tend to be normally distributed. Normal distributions follow the Empirical Rule-- that 68% of values fall within one standard deviation, 95% within two, and 99.7% within three. As predicted by the Central Limit Theorem, even for distributions that are not normally