|
Evaluation of the National Examination's Impact on the Quality of Learning in Russian Schools
|
| Presenter(s):
|
| Zvonnikov Victor,
State University of Management,
zvonnikov@mail.ru
|
| Marina Chelyshkova,
State University of Management,
mchelyshkova@mail.ru
|
| Abstract:
This year the Law about National Examination was entered in Russia. Despite of the Law and six years of experiment National Examination has numerous opponents who are afraid of it's negative impact on traditions of the Russian education system.
In paper we present the results of the research program focused on evaluation of the National Examination's impact on the quality of leaning. This research was conducted on different directions: the analysis of quality's change by comparative measurement of achievements with using the results of National Examination, the development scaling methods of achievements and test design for comparison of interval estimates of students, the creating methods providing correct interpretation of achievement's scores for management in education, the application of Hierarchical Linear Models for prediction of quality's changes of leaning and another.
|
|
Multi-level Longitudinal Analysis as a Method for Evaluating Reading First
|
| Presenter(s):
|
| Bruce Randel,
Mid-continent Research for Education and Learning,
brandel@mcrel.org
|
| Abstract:
This study uses longitudinal growth modeling to examine changes in reading proficiency for students in schools participating in Reading First. Data were available from two mid-western states; each state included approximately 15 schools and approximately 500 students. All students were administered the state-wide test of reading comprehension at three time points, one year apart. Scores from both state reading tests are on a vertical scale but analyses were run separately by state because the tests do not share the same measurement scale. Analyses were conducted to model the growth in reading comprehension during and after participation in Reading First programs. Each analysis estimated individual growth trajectories at the student level and also estimated the unique contribution of student demographic characteristics and school characteristics in explaining growth.
|
|
The Application of Multi-level Modeling in the Evaluation of After-school Programs: Linking Academic Success to Attendance
|
| Presenter(s):
|
| Jeremy Lingle,
Georgia State University,
jlingle1@gsu.edu
|
| Carolyn Furlow,
Georgia State University,
cfurlow@gsu.edu
|
| Sheryl Gowen,
Georgia State University,
sgowen@gsu.edu
|
| Syreeta Skelton,
Georgia State University,
snskelton@gsu.edu
|
| Abstract:
Multilevel modeling provides evaluators with a powerful tool to isolate the individual-level factors that may contribute to program effectiveness as well as to identify the impact of program-level factors and the interaction of variables across levels. These models also allow for evaluation of the effects of social programs which are often limited to quasi-experimental designs. This presentation arises from a state-wide evaluation of federally-funded after school programs. The purpose of this presentation is two-fold: (1) to enumerate the challenges provided by use of state standardized assessment scores and (2) to discuss the hierarchical linear models (HLM) that we used to analyze these data. Findings from our analyses support that attendance in after-school programs has positive effects upon certain, but not all, academic outcomes.
|
|
Comparing Urban and Suburban Schools: An Investigation of the Intervention Effects of Reading Recovery With Multi-level Growth Modeling
|
| Presenter(s):
|
| Jing Zhu,
The Ohio State University,
zhu.119@osu.edu
|
| Francisco Gómez-Bellengé,
Reading Recovery National Data Evaluation Center,
gomez-bellenge.1@osu.edu
|
| Abstract:
Usually, it is difficult to evaluate the effects of educational interventions. Nonrandom assignment of participants to different groups and various confounding factors cause concern for this kind of investigations. It is believed that multilevel models are effective in dealing with these issues. Because of No Child Left Behind legislation, a crucial question for many interventions is whether the program works equally well for different populations. In this study, multilevel growth modeling is applied to the national longitudinal assessment data from Reading Recovery (RR) during the 2005-2006 academic year. A particular interest for evaluation is comparing the intervention effects of RR on reading achievement in urban schools with suburban. Multilevel models will estimate the trajectories of student reading performance measured by the six tasks in the Observation Survey (OS). The estimates of average difference in mean OS scores between urban and suburban schools and the corresponding effect size will be reported.
|
| | | |