ERIC Number: ED235193
Record Type: RIE
Publication Date: 1983-Apr
Reference Count: 0
Adventures in Uncertainty: An Empirical Investigation of the Use of a Taylor's Series Approximation for the Assessment of Sampling Errors in Educational Research.
This study investigates the accuracy of the Woodruff-Causey technique for estimating sampling errors for complex statistics. The technique may be applied when data are collected by using multistage clustered samples. The technique was chosen for study because of its relevance to the correct use of multivariate analyses in educational survey research. To apply the technique the researcher must be able to write Fortran subroutines and must be able to ascertain a sampling error formula for a mean for whatever sampling situation is to be used (i.e. look up one of the standard texts). In return the technique will provide an estimate of the sampling error for any statistic which can be expressed in terms of a Fortran subroutine. Guides to numerical differentiation for the technique, and use of the technique and the writing of the Fortran subroutines are provided as appendixes to this paper. (PN)
Descriptors: Data Collection, Educational Research, Estimation (Mathematics), Guidelines, Mathematical Formulas, Multivariate Analysis, Research Methodology, Sampling, School Surveys, Statistics
The Australian Council for Educational Research, Frederick St., Hawthorn, Victoria, Australia 3122.
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: Australian Council for Educational Research, Hawthorn.
Identifiers: Approximation (Statistics); FORTRAN Programing Language; Sampling Error; Taylors Theorem
Note: Document may be marginally legible due to small print.