WRIGHT, WILLIAM FRANCIS

This study sought to accomplish three things:
1) to develop a model which would use semantic differential
techniques to measure the intensity of teacher attitudes toward public school supervisory concepts; 2) to
determine if teachers' attitudes differ significantly
in the subgroups of the study's population when analysis
of variance is applied to the data supplied by the model
instrument; and 3) to determine which rotation, orthogonal
or oblique, more closely approximates the simple
structure when factor analysis is applied to the data.
Analysis of the data was based on the responses
of 217 public school teachers who were enrolled in graduate
courses located at the University of Central
Florida and Florida Atlantic University. Each subject
responded to eighteen bipolar scales, evaluating seven supervisory concepts. Analysis of the data was accomplished
by three computer programs, Kaiser's "Little
Jiffy, Mark IV," to measure sampling adequacy, reliability,
and degree of empirical confirmation of the
model; analysis of variance to determine whether the
subgroups were statistically significant and evaluate
the relationship between the variables in the groups; and
factor analysis applied to the data which incorporated
two methods of rotation, orthogonal and oblique, to
determine which rotation more closely approximates the
simple structure. Findings: The findings may be summarized as follows:
1. Teachers' responses clustered into four
principal factors, here presented in rank order of
extraction--evaluation, potency, activity, and stability.
The evaluation factor accounted for over 55 percent of
the common variance.
2. The concept instrument, bipolar scales, and
semantic differential techniques can be combined to
differentiate teacher attitudes concerning specific concepts.
3. A comparison of the mean variable scores
for concept instruments revealed that responses by this
study's teachers rated the positions of principal, director and superintendent very low.
4. Data comparing such groups as university
attending, school districts of employment, and age groups
by analysis of variance techniques revealed statistical
differences that were substantive when evaluating the concepts
of dean, principal, director and superintendent.
5. Teaching level and male/female teacher group
accounted for statistical differences that did not prove
to be substantive over all concepts.
6. A comparison of the quartimax rotated factor
matrix and the oblique pattern matrix revealed that
clusters of variables are more clearly defined on the
pattern matrix. Near zero loadings are minimized and
high loadings are maximized. However, on several of the
concept data runs the differences were not clearly
defined. Conclusions: 1. The model developed for measuring teacher
attitudes was reliable and demonstrated a high degree
of empirical confirmation.
2. Supervisory job does appear to be a factor
influencing teachers' attitudinal responses on the
bipolar variables.
3. University of attendance, school district
of employment, and age group are factors which produce significant differences in teacher attitudes.
4. Factor analysis involving either orthogonal
or oblique should be selected based on the objectives
of the research project. Both methods may be utilized,
results compared, and the solution which provides the
best answer for the study involved should be selected.