Annika Stark and Barbara Culatta, Communication Disorders
Children are exposed to a variety of narrative texts from a young age. However, young children are not given this same amount of exposure to expository texts. Expository texts inform and describe as they expose a child to factual information. There is a current need in early childhood education for an assessment tool that helps teachers gauge young students’ expository text comprehension skills. In 2005, Dr. Kendra Hall-Kenyon, Dr. Janet Markham, and Dr. Barbara Culatta collaborated to develop an assessment tool to address this problem. The tool was titled the Early Expository Comprehension Assessment (EECA) and was found to be a reliable measure. Since 2005, the test has gone through two revisions. Through this study, the reliability and validity of the revised assessment tool, the EECA-R2, has been tested. This testing will help place greater emphasis in assessing expository text structures and will be beneficial to children’s future academic successes.
Students from 7 Title I preschools in the Alpine School District and students from the Brigham Young University lab school were assessed. Students were given three rounds of assessments. The first round was an assessment of the Narrative Language Measures (NLM). This is a previously validated preschool comprehension assessment. The second and third rounds of assessments included a Form A and Form B of the EECA-R2. Students were randomly assigned to two different groups. One group received Form A first. The other group received Form B first. All students were given the NLM and both forms of the EECA-R2. The 3 rounds of assessments were administered by 9 trained undergraduate students from Brigham Young University. The NLM assessment included 2 parts: a test of narrative retelling (TNR) and a test of story comprehension (TSC). The EECA-R2 addressed types of texts, graphical representations, labeling, verbal retelling, and presenting the information through matrices.
Each form of the EECA-R2 consisted of: 4 questions regarding the concepts of fiction versus non-fiction writing; 3 questions regarding graphics; 7 labeling questions; 6 narrative retelling tasks; and 3 matrix activities. The results from each of these areas of the test were scored by 8 trained undergraduate students through three training sessions.
To test the reliability of the EECA-R2, a Many Facets Rasch Model (FACETS) was used. The model accounts for the individual data from the student and the rater for each item of the test. This model is able to factor in the variability of each of the 8 different test scorers.
To test the validity of the EECA-R2, the NLM scores of each preschooler were compared with the child’s test scores on the EECA-R2.
Reliability of the EECA-R2
If the test were to be administered again, it would be found reliable. An estimated rater reliability of .00 was found through FACETS analysis, showing the raters to not be reliably different in the way they scored the tests. The reliability estimate of .96 was found for the test’s form. A reliability estimate of .99 was found for the items on the test.
Through infit mean square statistics, the difficulty of each item on the test was found. Items with lower measure numbers were found to be easier. Items with higher measure numbers are considered more difficult. Results showed Item 3, a graphic question, to be the easiest with a measure of -1.46. Item 13, which asked the child to compare and contrast a short passage, was found to be the most difficult task with a measure of 1.83. All other items on the test received measures between -1.29 and 1.23, suggesting the items of the test to be reliable measures of the data.
The reliability of each form of the EECA-R2 was also found through infit mean square statistics. The results showed Form A to be slightly more difficult than Form B. Form A received a measure of .05 logits. Form B received a measure of -.05 logits. These results indicate both forms of the test to be good measures of the desired measure.
Validity of the EECA-R2
The EECA-R2 results were compared with the results of the test of story comprehension (TSC) section of the NLM. A positive correlation was found with the following results:
These results suggest the EECA-R2 to be a valid measure.
In discussing the results, the nearly perfect rater reliability score and the lack of comparison in the TNR section of the NLM will both be discussed:
The exceptional rater reliability score may be due to the number of non-ambiguous items on the EECA-R2. It should be noted that of the 27 test items, 13 were considered “non-ambiguous”. In scoring and analyzing the results, the data from the narrative retelling (TNR) portion of the NLM was difficult to accurately correlate with the EECA-R2 scores. A possible reason for the poor scores on the TNR portion of the NLM could be the sequencing of the tests, as administrators gave the students the TNR section first.
The revised Early Expository Comprehension Assessment (EECA-R2) was found to be a reliable and valid measure. The results showed the rater reliability, item reliability, and form reliability to demonstrate the EECA-R2 to be a reliable assessment tool. In comparing the EECA-R2 results to the TSC section of the standardized NLM test, the results demonstrated validity.
In the development of a reliable and valid assessment tool that addresses expository comprehension in preschool-age children, the EECA-R2 will accurately assess expository comprehension in early childhood education. This effective assessment tool will benefit children from a young age and throughout their future years of education.