Date of Award


Degree Name

Doctor of Education


Educational Leadership

First Advisor

Dr. Uldis Smidchens

Second Advisor

Dr. Larry Schlack

Third Advisor

Dr. Richard Hogancamp


In this study the relationship and differences in student achievement and class size were investigated. The population of the study was the 111,199 students who completed the fall 1988 Michigan Educational Assessment Program (MEAP, Michigan State Board of Education, 1988) fourth grade test. The MEAP is a criterion referenced test and is the state of Michigan's only achievement test.

The independent variable of the study was class size. The dependent variables of the study were student achievement as measured by the students' fall 1988 MEAP reading, mathematics, and combined reading and mathematics test scores. The hypothesis of the study, supported by the findings of Glass and Smith (1978; M. L. Smith & Glass, 1979) was that there is an inverse relationship between class size and student achievement.

Data for completing the research design and statistical analysis were secured from National Computer Systems (NCS), the firm which completes all MEAP computer reporting services for the Michigan Department of Education. A listing of 400 students' (200 males and 200 females), class size, MEAP test scores for reading and mathematics, and sex of each student was provided by NCS and identified as the study's first sample. A second and third sample were derived from the original listing of 400 students that allowed the testing of two operational hypotheses.

The results of the hypotheses testing did not support the presence of an inverse relationship between the independent variable class size and the dependent variables of student achievement as measured by the fourth grade MEAP reading, mathematics, or combined reading and mathematics test scores. This was the case when a Pearson product-moment correlation coefficient was used with all the data as well as when the class sizes of 16 to 20 students were compared with class sizes of 30 to 34 students.

The relatively high test scores, absence of a normal distribution of test scores, and the resulting limited variation of tests scores, all of which can be expected with a criterion referenced test were identified as suspected explanations for the findings of the study. Suggestions were made for further research.

Access Setting

Dissertation-Open Access