Dear all,

For the next Quant SIG Seminar meeting (Monday, 27th Feb), Ioulia Televantou  (Department of Education, Oxford University) will be presenting on the following topic:

Integrating Structural Equation models with Contextual Value Added models and Regression Discontinuity Designs: A progress report on two distinct studies.

This promises to be another excellent talk. As usual the Quant SIG will meet in Seminar Room J from 12:15pm-2pm, Department of Education, 28 Norham Gardens, Oxford OX2 7PY. If you do not have access to the building, please contact Patrick Alexander ([log in to unmask]) to arrange access.


Best wishes,

Patrick Alexander


**
In this session of the Quant-Sig I plan to give a presentation on two separate studies, the second being given in a progress report format:

(i)	Investigating Contextual Effects Using English Primary School Data
This study, a “methodological-substantive” synergy, applied the four models demonstrated by Marsh et al. (2009) to longitudinal attainment and attitudinal data for English primary school pupils to address two substantive issues: The first was concerned with the contextual effect of school average attainment that is being used in contextual value-added models assessing school effectiveness. The second involved analysis relating attainment in and attitudes to Maths, and tested whether school average attainment is negatively related to individual self concept (the Big Fish Little Pond Effect). Both effects were found negative and significant. The study has important implications for school policies such as tracking and ability grouping.

(ii)	Integrating Multilevel Structural Equation Models with Regression Discontinuity Designs: An application to TIMSS-95
The majority of school effectiveness studies focus on the relative differences between schools in their contribution to the student’s knowledge. It is nevertheless just as important to assess the absolute effect of schooling (schooling versus no schooling) on students’ development. One way in which this can be achieved by applying the Regression Discontinuity (RD) approach to educational data that meet the assumptions underpinning the approach.
The data used come from the Third International Mathematics and Science Study (TIMSS 95) and refer to students from two adjacent grades both in primary (grades 4 and 5) and secondary school (grades 8 and 9). This is – to my knowledge – the first study to use the new achievement data released in April 1999 that contain scaled scores not only for the overall mathematics attainment of the students but also for their achievement in the different content areas. The latter are used as multiple indicators to control for unreliability in the measured achievement of the students: The latent variable models recently developed by quantitative researchers (e.g. Marsh, Lüdtke, Robitzch, Trautwein, Asparouhov, Muthén and Nagengast, 2009) are demonstrated. Multilevel models are used throughout the analysis that make adjustments for the effect of age and grade level on students’ achievement as well as for a variety of different student and school level variables.
The absolute effect of schooling was found positive both for primary and secondary school. The effect of age was found much smaller in the higher grades than in the lower ones. Primary schools appeared to differ significantly in their estimated absolute effects –this was not the case for secondary schools. Controlling for background variables did not substantially alter the estimates while measurement error adjustments led to larger grade to age effect ratio.

You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.