Xin Liang

Date of Award

1-3-2003

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Richard Landry

Second Advisor

Mary Laycock

Abstract

The purpose of this study was to conduct statistical tests to examine the reliability and effectiveness of the sampling approach applied which is different from the conventional matrix sampling method in the instrument design and sampling procedures. The instrument (The Spring 2001 Protective and Risk Factor Survey) was created to represent a cross-section of faculty and staff from the specialties of Health, Life Skills, Physical Education, administration of the district, and other personnel, such as police and professors from the university. The instrument design and sampling procedure was conducted via a modified matrix sampling model. The data was collected and reported by the school district in 2001. The most salient differences of the modified matrix sampling model lay in the incorporation of core items across the different assessment instruments, while dividing the remainder of the items in the entire item pool into several survey forms. The outcome of this design satisfied the requirements of the school district in two ways: (1) All the items generated by the school district instrument development committee were included in the survey forms. (2) All the students from the 4th to the 12th grades participated in the study. Chi square tests, factor analyses, correlation coefficient analyses, and coefficient of congruency analyses results all support the modified matrix sampling model. There was no significant difference among the students in responding to the different survey forms. The core items formed similar factor patterns, and there existed a strong positive relationship among the common factors across the forms. These findings indicated that this modified matrix sampling model is reliable and effective.

Download

Share

COinS