11.07.2015 Views

Abstracts - Earli

Abstracts - Earli

Abstracts - Earli

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

processing, is an iterative process that is capable of discovering complex relationships andinteractions in the inputs and outcomes. The goal of NNs is often to maximize classificationaccuracy, regardless of understanding the relative strengths of predictors. However, recent workhas demonstrated that NNs in combination with more traditional statistical techniques can providenovel insights into complex relationships to identify which variables are driving the model(Costigan, 2003). We will present an introduction to NNs, including definitions, datarequirements, and parameter settings. Included in the presentation will be a discussion of relativestrengths and limitations, linking NNs in a sequential analytic chain to maximize modeling. Wewill also present a case example of this application in educational research.Fourth panelist: Peter NennigerPeter Nenniger, University of Koblenz-Landau (Campus Landau), GermanyAny assessment of self-regulated learning is faced with a highly complex system of interactingelements which have to be regarded as the dynamics in the individual’s autonomous regulation.Because of this complexity predictive assessment requires a description of the structuralcharacteristics as well as of the functions of structural transformations. Although a structuralapproach may help, a comprehensive evaluation of the assessment concepts that are used is hardlypossible, due to their diversity in elaboration and formalisation and due to the diversity and partialincomparableness of their underlying theoretical concepts. For this reason, I suggest that in anyfurther development of predictive systems approaches emphasis should be given to thedevelopment of actual "systems" and that the respective implications should be examined. Inaddition, I would suggest to frame the differential assessment approach in a mode thatencompasses the multiple facets of the phenomenon so that it can cope successfully withcomplexity.M 231 August 2007 14:35 - 15:55Room: 0.87 MarxPaper SessionResearch methodologyChair:Filip Dochy, K.U.Leuven, BelgiumMeta-analysis for repeated measures designs in educational research – options and challengesHans A. Pant, Humboldt-University Berlin, GermanyKai S. Cortina, University of Michigan, USAIn empirical educational research, meta-analysis is widely considered the most adequate statisticaltool in establishing efficacy of interventions. Meta-analysis uses the summary statistics (effectsizes) from individual studies as raw data and analyzes the heterogeneity of effect sizes acrossstudies. In educational research, key concepts like learning are time-dependent processes bydefinition. Hence, the evaluation of intervention effects or long-term trajectories inevitablyrequires longitudinal or repeated- measurement designs. However, until recently, the statisticalmodels to deal with within-subject designs within a meta-analytical framework were based onunrealistic model assumptions. We review the core statistical problems and discuss current– 687 –

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!