 # confidence interval

## Terms from Statistics for HCI: Making Sense of Quantitative Data

Confidence intervals give a lower and upper bound for the true value of an effect. They allow you to be able to answer questions such as "What is the maximum likely value of this?", or "How close to no difference is there between these two conditions?". They use the same information and mathematics as are used to generate the p-values in a significance test. You can choose to set higher or lower confidence levels, for example the 95% confidence interval for a value might be [0.3,1.1] and the 99% confidence interval [-0.1,1.5]. In general the higher the level of confidence you ask for, the wider the range. Like all statistics these bounds are uncertain. However, if you do lots of studies and calculate 95% confidence intervals for them all, then over time the proportion of true values that lie within the 95% confidence intervals will be at least 95% ... although you can't tell which they are!

Defined on page 65

Used on pages 12, 15, 57, 60, 61, 65, 66, 67, 68, 77, 82, 85, 86, 93, 94, 100, 102