In a message dated 11/20/2001 9:13:41 PM Central Standard Time, [EMAIL PROTECTED] writes:
> n the extract from "What Matter Most" cited above, Linda Darling-Hammond > says, > > "In an analysis of 900 Texas school districts, Ronald Ferguson found that > teachers’ expertise—as measured by scores on a licensing examination, > masters degrees, and experience—accounted for about 40% of the measured > variance in students’ reading and mathematics achievement at grades 1 > through > 11, more than any other single factor. The Minneapolis School District did a study, reportedly using a regression procedure to analyze the data, which attributed 44% of test score variability to "teacher efficacy." It is important to understand that Ferguson's analysis of 900 Texas school district is also a study of "teacher efficacy." In a study of teacher efficacy, measures of teacher expertise, such as years of experience, are compared to measures of student performance, with controls for other variables, such as eligibility for free or reduced-price lunch programs, student attendance, mobility (transfers from one school to another), and "parent involvement." There are school policies and practices that influence the quality of instruction and academic achievement, but may not be accounted for in studies of teacher efficacy, including the use (and abuse) of computers, ability-grouping, choice of curriculum products, special education practices, teacher assignment policies, and so forth. One way to control for the effect of school policies and practices on the quality of instruction is to analyze data from schools where differences in school policies and practices are kept to a minimum. One controls for student and family background characteristics by comparing data from schools with similar student and family background characteristics. Ditto for class size, years of teaching experience and other factors you are trying to evaluate. The measured variance in academic achievement related to teacher expertise, class size, and student and family background characteristics does not account for all of the variance in student performance. As Michael Atherton noted in a posting on 21 November 2001 (Re: [Mpls] Re: Schools: Financial Concerns), "...this article [by Linda Darling-Hammond] never reports what amount of the influence on student MATH achievement was not accounted for by their model. This is really important because without knowing the influence of the unknown factors you can't truly calculate the influence of the known factors." Either an estimate of the effects of unknown factors were left out of the report or they were not taken into account in the study. In the either case, what you get is an inflated estimate of the importance of teacher expertise, class size, and student and family background characteristics. After all, factors that explain some of the variance in measures of student achievement do not neatly fit into any of those categories. For example, developmental patterns and learning styles are influenced by genetic inheritance (parent involvement?) and are best accommodated by a having teacher with good assessment skills and a large repertoire of learning strategies (teacher expertise). Factors that influence the quality of instruction and variability in academic achievement attributable to school policies and practices, such as curriculum choices and ability-grouping, may be erroneously attributed to student and family background characteristic and "parent involvement." It is certainly misleading to say "...research shows that student achievement can be accurately measured as follows..." on the basis of information provided in Linda Darling-Hammond's article. Policy decisions by the board that are based upon (or justified by) limited and distorted information are certain to do the school system more harm than good. -Doug Mann, King Field <http://educationright.tripod.com> _______________________________________ Minneapolis Issues Forum - A Civil City Civic Discussion - Mn E-Democracy Post messages to: [EMAIL PROTECTED] Subscribe, Unsubscribe, Digest option, and more: http://e-democracy.org/mpls
