Current research projects/outreach activities of QME faculty
Dr. Davenport has done quite a bit of work on mathematical
artifacts of statistical procedures and have several current
projects along those lines: geometrical representations of
statistical procedures, Simpson's Paradox (averages can sometimes be
misleading), and profile analysis (finding quantitative ways to
categorize individuals based on a set of predictor variables). He
also does work in academic achievement. Current projects there
consist of the relationship of mathematics course-taking patterns to
mathematics achievement and profiling high school dropouts who
return to complete their education.
Dr. Davenport is also the director of a free
ACT/SAT preparation course for at-risk students that has been in
existence since 1991. It is partly sponsored by the university with
other community groups. The program is conducted annually and last
year he served 156 students.
Dr. Davison has several research projects: 1) Measuring Average
Speed of Numerical Reasoning. In this project, his research team is
developing measures of a reasoning speed factor separate from a
reasoning accuracy factor. Students take two numerical reasoning
tests, one with unlimited time and the other with a time limit on
each item. For each item, they record response accuracy and response
time. 2) Identifying Patterns of Scores in Longitudinal and
Cross-sectional Data. This is a long term project which has
continued for several years. For cross-sectional data, his
colleagues Drs. Ernest Davenport of the University of Minnesota and
Se-Kang Kim of Fordham University, and he have developed (a)
multidimensional scaling techniques for identifying the major
patterns that appear in a battery of test scores, (b) a multiple
regression technique for identifying patterns of scores that predict
criterion variables or that distinguish among criterion groups, and
(c) structural equations techniques for testing hypotheses about
patterns of scores that account for associations among variables.
These analyses are being used in studies of noncognitive measures
that predict academic success in high school and college,
personality variables that predict interest and success in
management careers, and cognitive score patterns that distinguish
educational disability groups. 3) A colleague, Dr. Cody Ding of the
University of Missouri-St. Louis and he are now extending these
pattern techniques to longitudinal data. In longitudinal data, the
“patterns” are more typically called “growth trends” or “change
patterns.’’ They are developing the multidimensional scaling methods
for the identification of growth/change patterns that appear in
longitudinal data. These techniques are being designed as
exploratory hypothesis generation techniques to complement the
confirmatory hypothesis testing techniques of growth curve modeling
with structural equations modeling, hierarchical linear modeling,
and mixed effects modeling. They are using these techniques to study
the growth rates of initially low achieving children to see if they
seem to be catching up with initially higher achieving children;
that is, to see if initial achievement gaps tend to close over time.
Dr. Davison is also involved with the
Center for Applied Research and Educational
Improvement (CAREI) at the University of Minnesota on a
quasi-experimental evaluation of an intensive reading and
mathematics program for disadvantaged children. In the process, they
are employing new techniques that use educational accountability
data bases to draw a control group closely matched to their
experimental group on prior achievement and related variables. In
this project, they are trying to develop methods of drawing
quasi-experimental control groups that come as close as possible to
rivaling true experiments in the degree to which treatment and
control groups are matched on potentially confounding variables for
those situations in which random assignment may not be possible.
Dr. delMas is currently co-PI with Dr. Joan Garfield on two NSF
grants. The ARTIST project
(Assessment Resource Tools for Improving Statistical Thinking) has
been developing test items and research instruments to be used in
first college level statistics courses. The current phase of the
project is developing a Statistics Teaching Inventory that assesses
teachers instructional methods and beliefs about teaching as well as
institutional and course characteristics.
The second NSF project is AIMS (Adapting and
Implementing innovative materials in statistics). This project is
developing detailed, research-based lesson plans for a first course
in statistics that is aligned with ASA-endorsed guidelines for
teaching statistics.
A third research project, with Joan Garfield,
Andy Zieffler, and Rob Gould (UCLA) is a study of the development of
students’ reasoning about statistical inference.
Dr. Garfield is currently co-PI with Dr. Robert delMas on two NSF
grants. The ARTIST project
(Assessment Resource Tools for Improving Statistical Thinking) has
been developing test items and research instruments to be used in
first college level statistics courses.
The current phase of the project is developing a
Statistics Teaching Inventory that assesses teachers instructional
methods and beliefs about teaching as well as institutional and
course characteristics. The second NSF project is AIMS (Adapting and
Implementing innovative materials in statistics). This project is
developing detailed, research-based lesson plans for a first course
in statistics that is aligned with ASA-endorsed guidelines for
teaching statistics.
A third research project, with Bob delMas, Andy
Zieffler, and Rob Gould (UCLA) is a study of the development of
students’ reasoning about statistical inference.
Dr. Harwell is a Co-PI(with T. Post) in the project, “Impact of
Standards-Based, Traditional, and UCSMP High School Mathematics
Curricula on Post-Secondary Students’ Achievement, Course Taking
Patterns and Persistence in STEM Classes (3 years, National Science
Foundation grant). This investigation will provide evidence for the
effectiveness of NSF funded mathematics curriculum materials (CMIC,
IMP, MMOW) as reflected in the college STEM achievement, course
taking patterns, and persistence.
He is also the statistician for the project
“Exploring the Development of Beginning Secondary Science Teachers
in Various Induction Programs.” (5 years, National Science
Foundation grant). This project is a longitudinal study following
the development of beginning secondary science teachers during their
first three years in the classroom with a focus on how different
types of mentoring and induction programs influence their
development.
Dr. Lawrenz is involved with several projects related to
evaluation, mostly of science or mathematics programs. One recent
grant is a research project studying the effect of project level
involvement in the planning, implementation and dissemination of
national program evaluation on use of the program evaluation
processes and outcomes.
Another grant project is to promote the
development of collaborative evaluation communities in selected St.
Paul schools. These communities involve teachers, administrators and
others from the schools with graduate students and faculty from the
university to facilitate data based decision making.
A third grant is the evaluation of a national NSF
program, the Noyce Scholarship Program, designed to increase the
numbers of science and mathematics teachers in high need areas.
Dr. Long has two main areas of interest, ordinal data analysis
and longitudinal data analysis. His work in ordinal methods
concentrates on bivariate and multivariate regression based on
dominance scores (Long, 1999, 2005; Long, Feng, & Cliff, 2003). His
work with longitudinal methods focuses on application, especially
the use of linear mixed models and generalized linear mixed models
in developmental psychology and education (Long & Pellegrini, 2003;
Pellegrini & Long, in press; Webb, Long, & Nelson, 2005).
Dr. Long is statistical advisor for two centers
at the University of Minnesota, the
Research Institute on
Progress Monitoring (RIPM) and the Center for Neurobehavioral
Development (CNBD). The goal of RIPM is to develop a system of
progress monitoring to evaluate effects of individualized
instruction on access to and progress within the general education
curriculum. CBND is a research center housing many studies related
to children’s cognitive and neurobehavioral development and spans
multiple departments including Pediatrics, Psychology, Educational
Psychology, and the Institute of Child Development. In his role as
statistical advisor for these centers he works with faculty members
and graduate students from diverse backgrounds (e.g., MDs and
Ph.D.s) to design and carry out research projects. Collaborations
with CNBD colleagues have lead to a number of publications in top
journals (e.g., deRegnier, Long, Georgieff, & Nelson, in press;
Siddappa, Rao, Long, Widness, & Georgieff, in press; Webb, Long, &
Nelson, 2005).
Dr. Rodriguez is currently investigating recent advancements of
the application of meta-analysis to reliability coefficients in a
methodology referred to as Reliability Generalization. It also
includes additional work related to understanding the role of
heterogeneity and sampling variability in inferences related to
coefficient alpha. He works closely with the
Center on Reading Research in their efforts to model growth in
reading and literacy skills and the impact of school-based
interventions. This work is challenging in the context of
multi-level modeling of growth given student, teacher, and school
level characteristics—including cross-classified models of student
growth as they move from one teacher to the next over time.
Finally, he is also working on two projects which
include international assessment work. The first is the development
of a national standards-based achievement assessment system in
Guatemala. This is challenging in two respects, including the fact
that the assessments are annual assessments with a form of
matrix-sampling and the second being that the tests are given in 5
languages, including Spanish and four Mayan languages. The second
project involves the development of instruments for the IEA funded
first international study of teacher education in mathematics. Two
dozen countries are involved in this project, headed by faculty at
Michigan State University, to study how teachers learn mathematics,
what they believe about teaching mathematics, and their knowledge of
mathematics.
March 2007
|
