Eric Loken, Ph.D.

2001, Harvard University

Assistant Professor/Senior Research Associate

119-B Henderson
(814) 865-7091
loken@psu.edu

Research

My work is currently divided into two major areas - research methodology and adolescent academic achievement.

Recent advances in statistics have made sophisticated techniques available to researchers. When used responsibly, these advanced statistical techniques can support theoretical models and also lead to valuable new empirical insights.  I am working on latent variables models, both with continuous latent variables such as factor and SEM models, and mixture models which have categorical latent variables.  I also work in the area of item response theory, models typically used for educational measurement. In general I focus on the application of Bayesian methods to improve inference and inform model selection.

In terms of academic achievement, my current focus is on the early identification of academic talent in middle school children. The prevailing method of selecting participants for national "talent searches" is to test students with the SAT. Because a large number of academically talented young adolescents "max out" on age-appropriate tests, it is thought that an out of level test like the SAT can be a useful tool for identifying the most talented. Little research has been done on the properties of the SAT in this age-group. One major question is the degree of variance at the school and district level that is due to unequal educational opportunities for students up to the middle school years.  Another question regards the specific properties of the testing situation for such young students.  I am also interested in the longitudinal development of student ability as reflected in standardized tests, along with the personality and motivational factors that relate to performance.  Finally, I am interested in the development of large-scale web-based assessment and instruction.

Education

Research and Professional Experience

Honors and Awards

Selected Publications