http://pocketknowledge.tc.columbia.edu/home.php/browse/18831/5292?thumbs=no This is a file listing from PocketKnowledge This is a file listing from PocketKnowledge en-us Thu, 23 May 2013 13:09:09 EDT Various: PocketKnowledge Digital Archiving http://pocketknowledge.tc.columbia.edu/home.php/viewfile/12497 <br />ABSTRACT<br />A STUDY OF RELATIONSHIPS BETWEEN CERTAIN MEASURES OF CLASS SIZE IN ELEMENTARY SCHOOLS AND A CRITERION OF PUPIL ACHIEVEMENT<br />Marshall Scott Woodson<br />Relationships between several measures of class size policy of 95 school districts (ranging in size from small to medium-large) and criteria of academic achievement of the fourth and sixth grade pupils in these districts were examined in this study.<br />Criteria of academic achievement were computed as residuals based upon pupil scores on achievement and intelligence tests. This was done by predicting the achievement test score from the intelligence test score through the use of a standard regression equation and by converting these to standard scores. Differences between the predicted score and the actual score on the achievement test (the residual) were used as the criteria against which to examine class size data.<br />Three sets of these criterion (residual) scores were obtained--those based upon the total achievement test battery and those based upon the reading and the arithmetic sub-tests.<br />School district scores for the fourth and sixth grades were then obtained by averaging the residuals (individual criteria) of pupils in each grade for each district. Pupils were further classified into a high ability group (I.Q. above 116 on the Otis scale), a middle ability group (I.Q. 85 to 116) and a low ability group (I.Q. below 85).<br />In addition to the twelve criteria computed for each grade, eight size variables were computed. Two of these represented measures of school system size. The others were direct or indirect measures of school district class size practice. These measures reflect a system-wide condition, rather than an analysis of a one-for-one relationship between individual studentÆs scores on the class scale and the achievement test scale.<br />Thus, 20 variables (12 criteria plus 8 size variables) were generated for each grade and for the averages of the two grades. Through the assistance of the College Computer Center, a correlation matrix 60 by 60 was run. In addition, differences of means tests were run, comparing the mean of each criterion value of those districts in the upper and lower thirds of the distribution of a size variable. This procedure was followed for each of the criteria and then repeated, but with the roles of the criteria and size variables reversed. Both correlation coefficients and differences of means were tested for levels of significance.<br />For the most part, the size and significance level of the individual statistical findings were very modest. However, the consistency of the patterns formed by these findings from one variable to another was impressive. The conclusions were based more upon consistency of statistical pattern than upon the absolute value of discrete findings.<br />It was concluded from the weight of the evidence that there was a small inverse relationship between the academic achievement of pupils and class size; but:<br />This relationship tended to be smaller for pupils of higher scholastic potential than pupils of lower scholastic potential.<br />This relationship tended to be smaller for criteria based upon total achievement test batteries or arithmetic sub-tests than criteria based upon reading sub-tests.<br />This relationship tended to be more certain of measurement at the sixth grade level than at the fourth grade level.<br />There was little evidence that district size, per se, reflected itself in the magnitudes of the achievement criteria, although the relationship between scholastic achievement and class size was essentially obliterated in the case of a group of small, relatively sparsely populated, school districts.<br />All of these conclusions were subject to the kinds of class size measures used. The findings from this study raised the possibility that the practice of using ôaverage class sizeö as the lone measure of class size tended to oversimplify the study of the relationship with pupil achievement.<br /> <br />ABSTRACT<br />A STUDY OF RELATIONSHIPS BETWEEN CERTAIN MEASURES OF CLASS SIZE IN ELEMENTARY SCHOOLS AND A CRITERION OF PUPIL ACHIEVEMENT<br />Marshall Scott Woodson<br />Relationships between several measures of class size policy of 95 school districts (ranging in size from small to medium-large) and criteria of academic achievement of the fourth and sixth grade pupils in these districts were examined in this study.<br />Criteria of academic achievement were computed as residuals based upon pupil scores on achievement and intelligence tests. This was done by predicting the achievement test score from the intelligence test score through the use of a standard regression equation and by converting these to standard scores. Differences between the predicted score and the actual score on the achievement test (the residual) were used as the criteria against which to examine class size data.<br />Three sets of these criterion (residual) scores were obtained--those based upon the total achievement test battery and those based upon the reading and the arithmetic sub-tests.<br />School district scores for the fourth and sixth grades were then obtained by averaging the residuals (individual criteria) of pupils in each grade for each district. Pupils were further classified into a high ability group (I.Q. above 116 on the Otis scale), a middle ability group (I.Q. 85 to 116) and a low ability group (I.Q. below 85).<br />In addition to the twelve criteria computed for each grade, eight size variables were computed. Two of these represented measures of school system size. The others were direct or indirect measures of school district class size practice. These measures reflect a system-wide condition, rather than an analysis of a one-for-one relationship between individual studentÆs scores on the class scale and the achievement test scale.<br />Thus, 20 variables (12 criteria plus 8 size variables) were generated for each grade and for the averages of the two grades. Through the assistance of the College Computer Center, a correlation matrix 60 by 60 was run. In addition, differences of means tests were run, comparing the mean of each criterion value of those districts in the upper and lower thirds of the distribution of a size variable. This procedure was followed for each of the criteria and then repeated, but with the roles of the criteria and size variables reversed. Both correlation coefficients and differences of means were tested for levels of significance.<br />For the most part, the size and significance level of the individual statistical findings were very modest. However, the consistency of the patterns formed by these findings from one variable to another was impressive. The conclusions were based more upon consistency of statistical pattern than upon the absolute value of discrete findings.<br />It was concluded from the weight of the evidence that there was a small inverse relationship between the academic achievement of pupils and class size; but:<br />This relationship tended to be smaller for pupils of higher scholastic potential than pupils of lower scholastic potential.<br />This relationship tended to be smaller for criteria based upon total achievement test batteries or arithmetic sub-tests than criteria based upon reading sub-tests.<br />This relationship tended to be more certain of measurement at the sixth grade level than at the fourth grade level.<br />There was little evidence that district size, per se, reflected itself in the magnitudes of the achievement criteria, although the relationship between scholastic achievement and class size was essentially obliterated in the case of a group of small, relatively sparsely populated, school districts.<br />All of these conclusions were subject to the kinds of class size measures used. The findings from this study raised the possibility that the practice of using ôaverage class sizeö as the lone measure of class size tended to oversimplify the study of the relationship with pupil achievement.<br /> Marshall Scott Woodson Marshall Scott Woodson Thu, 23 May 1968 13:09:09 -0400 Class size http://pocketknowledge.tc.columbia.edu/home.php/viewfile/12497 http://pocketknowledge.tc.columbia.edu/home.php/browse/18831/5292?thumbs=no