Date of Award
Master of Science (MS)
Mathematics and Statistics
Ronald V. Canfield
This report compares the bootstrapping to jacknifing statistical procedures in terms in bias, confidence interval and estimation of median. Related literature have been reviewed. A bootstrap allows a researcher to get an approximation to the distribution of possibly complicated statistical summaries. It is based on random sampling with replacement from experimental units. Jacknife has also been in operation prior to bootstrapping statistical procedure. The jacknife divides the data into subgroups and obtains partial estimates of these subgroups by omitting one subgroup at a time. When both of these statistical resampling procedures are compared the bootstrap has less bias, more accurate confidence interval and better estimation of the median than the jacknife.
Gobena, Amanuel, "Comparison of Bootstrap and Jacknife Statistical Procedures" (1985). All Graduate Plan B and other Reports. 1251.