Bootstrap resampling is a procedure helpful in constructing confidence interval estimates of parameters when the traditional methods should not be used (such as data sets with distributions that are far from normal).
First enter the sample values in column 1 of the STATDISK data window.
For the number of resamplings, enter the number (between 1 and 10000) of new "bootstrap" samples to be generated by randomly selecting (with replacement) values from the original data set.
Results: Column 2 will list the means of the generated samples. Column 3 will list the standard deviations of the generated samples.
Suggestion: After obtaining the results, sort the means and standard deviations by clicking on "Data Tools" and "Sort a column" in the STATDISK data window. The sorted statistics can then be used to construct confidence interval estimates of the parameters.
EXAMPLE: If you enter 1000 for the number of resamplings, then sort the 1000 sample means, a 95% confidence interval estimate of the mean can be found by finding the 2.5 percentile value and the 97.5 percentile value. The lower confidence interval limit is the 2.5 percentile value (which is the mean of the 25th and 26th values in the sorted list of 1000 values), and the upper confidence interval limit is the 97.5 percentile value (which is the mean of the 975th and 976th values in the sorted list of 1000 values). Such a confidence interval can also be used as a tool for hypothesis testing.