The effect of intensive reading instruction on the reading achievement of low socio-economic elementary students in a rural Tennessee school system : the development of an evaluation model

The purpose of this study was the development of an evaluation model to assess the effectiveness of an intensive reading program for rural, low socio-economic elementary students involved in a Special Emphasis Reading Project. To measure the effectiveness of one or two years of intensive instruction, the evaluation model utilized intact classes, normally collected school data, and a prescribed treatment. There was an attempt to determine with which grade level(s), sex and/or race, if any, the reading instruction was most effective.

Initially there was an analysis of the treatment and control schools in the Special Emphasis Reading Project which determined comparability in reading achievement and other significant characteristics, such as IQ. The determination that there were no substantial differences on pretest comparisons between the treatment and control school further strengthened the evaluation model. The evaluation model utilized an adaptation of a quasi-experimental design. It was a measured equivalent-control group model. This was possible due to the selection of schools based on socioeconomic indicators and the comparison of grade achievement scores.

In addition to the establishment of a rigorous system to help insure the authenticity of the equivalency of the two groups, this model was also strengthened by the testing of longitudinal repetitions of the basic model and by measuring the projected progress of groups of students versus their actual gain, using a test of multiple regression.

Another method of examining data for determining success of the reading program was by comparing, using an analysis of covariance, selected pre- and posttest Total Reading scores of the Metropolitan Achievement Test.

In terms of mean grade equivalent on the Metropolitan Achievement Test, the data indicated that the experimental group showed higher mean values than the control group. The differences ranged from less than one month to more than nine months with the gains leaning in the direction of the experimental group. In the fourth grade, after two years in the project, there was a mean gain of l.19 for the experimental group and also a gain (3.69 to 4.25) over the summer.

Upon examination of the variables used in the regression equations for predicting achievement, it was found that the fourth-grade Total Reading score was the best predictor for both fifth- and sixth-grade reading in the experimental school; in the control school this was not the case.

Using the analysis of covariance, the data showed a significant (6.9l) between control and experimental groups when the second-grade Total Reading score was compared with the fifth-grade score using the former score as the covariate.

Since school systems are usually interested in an evaluation that would indicate the grade levels at which funds could best be spent, this study should allow some conclusions in that regard. Utilizing data from regression equations to predict gains, it appeared that the best concentration of effort would be at the fourth-grade level. If funds were available the results of this study would indicate that a summer program at the end of the third grade would be highly desirable.

The project school was, at the time of this study, participating in ESEA, Title I programming. It had been involved with a "pull-out" type of reading program for ten years and still had more than 70 percent of the students reading one or more years below grade level. After a year and a half of Special Emphasis programming, this percentage dropped to less than fifty percent. Thus, it seems fairly safe to say that an intensive in-classroom reading approach seemed to be more effective than a "pull-out" type of activity.

This study revealed that the Special Emphasis Reading Project worked quite well with children whose mean IQ scores (85) fell approximately two standard deviations below the mean IQ scores of the total school district.

Of the three measures to predict performance: pre-post achievement test comparisons, regression equations, and analyses of covariance, it appeared that the regression equation was most adequate for this evaluation model.