An exploratory investigation on the distribution of Miller's jackknife test statistics for small samples

• Main Authors: Badiola, Donna G., Cruzado, Edward C.
• Format: text
• Language: English
• Published: Animo Repository 1992
• Subjects: Sampling (Statistics); Distribution (Probability theory); Mathematical statistics
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/15966
• Institution: De La Salle University

The accuracy of obtaining critical points for Miller's Jackknife test for small samples by way of large sample approximation is investigated through small sample Monte Carlo experiment. With the aid of a computer program in Turbo Pascal Language Version 5.5, random values from a uniform distribution are generated. From these random values, 26 sets of 75 Q-statistics coming from equal sample sizes and another 98 sets of 75 Q-statistics coming from unequal sample sizes are computed. Sample sizes range from 5,5 to 30,30. These Q-statistics are tested for normality using Chi-square Goodness-of-fit-test. The results show that the Q-statistics from 26 cases of equal sample sizes are normally distributed while the Q-statistics from the 98 cases of unequal sample sizes were not normally distributed. Thus, obtaining critical points by way of large sample approximation is accurate for Miller's Jackknife test statistics for small and equal sample sizes.