Date of Award


Document Type


Degree Name

Master of Arts in Education



First Advisor

Gene Henderson

Second Advisor

Diane Denney


This project was an exercise in assessment of student writing in the Francis Howell School District. This district began a writing program in the fall of 1986, emphasizing 45 minutes of daily writing instruction and practice each school day . After approximately 44 weeks of exposure to the program, the assessment was conducted in the winter of 1988.

A published narrative essay test, the CAT Writing Assessment System, was given to 280 randomly selected students in grades three through six. Of these, 80 were used in the training of a rater team which then inter-scored the remaining 200 essays. They read each essay, gaining an overall impression of writing proficiency, and assigned a rank score of one (unacceptable) to four (good) . These raters were guided in their scoring by the test publisher's provision of a scoring manual and anchor papers (sample essays chosen to illustrate the level of writing expected at each of the four categories , derived from experience in scoring tests from schools nation-wide). This type of grading is called holistic scoring.

The results of this writing assessment would mean little without the ability to compare them with either a pretest or the performance of a control group. Since no pretest was systematically given before the writing program was begun, and all students in the district were treated in the writing program, it became necessary to create a theoretical control group, using the Kolmogorov-Smirnov one sample test. This is a statistical test of goodness of fit, which determines whether the scores in a sample can reasonably be thought to have come from a population having some specified theoretical distribution.

The 1986 achievement test stanines for language expression were recorded for each student in the sample, and used as an indicator of rank in written language proficiency. These rank-scores were compiled to develop a theoretical cumulative frequency distribution of student ability before treatment. That distribution was compared to the distribution of scores on the writing test using the K-S one sample test. Significant differences in distributions were found for all grades sampled except grade six (at the .05 level), indicating writing performances above expectations for grades three through five.

This project also demonstrated that holistic scoring can be done reliably and efficiently. This method of scoring was being practiced (to varying degrees) by most teachers in the district regularly in evaluating student writing; yet, without proper training , the majority found it laborious and intimidating. A lack of district standards and a unified understanding of the relative importance to place on the various aspects of writing may have contributed to their insecurity with the system. The training methods and use of anchor papers described herein, along with the practice of inter-scoring essays for an averaged score, may provide a model for a more successful alternative in writing evaluation which could be used district-wide .

Creative Commons License

Creative Commons Attribution-NonCommercial 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License

Included in

Education Commons