For example, a consultant for a large company uses a 1-sample sign test to determine whether the company's median salary differs from the industry average of $45,000. If the median differs from the target, the analyst uses the confidence interval to determine how large the difference is likely to be and whether that difference has practical significance.
The 1-sample sign test does not make assumptions about population symmetry.
To perform a 1-sample sign test, choose
.If your data come from a symmetric distribution, use 1-Sample Wilcoxon.
If you have more than 20 observations, or if your data aren't severely skewed, use 1-Sample t because the test has more power.