May
19, 20, and 21, 2008
Brady
West
This
workshop introduces the analysis of multilevel and
longitudinal data, emphasizing the use of hierarchical
linear models (HLMs). Participants
will be introduced to the use of the HLM software, and each
participant will have continuous access to a computer. The
workshop will consist of lectures including several hands-on
examples using the HLM software.
Many
studies in social sciences (e.g., education, human
development, public health, sociology) are multilevel,
longitudinal, or both. Multilevel data arise when
participants are clustered within social settings. The
variation and covariation within and between such settings
are often of interest substantively and should not be
ignored when assessing relationships between explanatory
variables and outcomes. In longitudinal research, we
repeatedly observe subjects. These repeated measures for
each participant will be correlated and explanatory
variables may be time-varying or time-invariant. This
workshop will consider the issues of analysis that arise in
multilevel and longitudinal research settings.
We
will first consider two-level cross-sectional studies in
which persons (level 1) are nested within organizations
(level 2). The level-1 model specifies a process within each
organization, and the level-2 model explains how these
processes are different between organizations. Next, we will
discuss two-level studies of individual growth and compare
the structures of these studies to multilevel studies. We
will also consider three-level models. We will focus on the
case in which repeated measures (level 1) are nested within
persons (level 2) who are themselves nested in organizations
(level 3).
All
of these studies will involve nearly continuous outcomes for
which the normality distribution is at least plausible. They
will also feature purely nested designs (e.g., persons
nested within organizations). The workshop will provide
participants with an overview of other types of applications
where hierarchical linear models or generalized hierarchical
linear models are appropriate (e.g., binary outcomes), and briefly discuss how the HLM software could be used to model such data.
Instructor:
Brady
West
is a senior statistician and statistical consultant at CSCAR. His CSCAR
responsibilities include consulting on linear
regression and related techniques, and his research interests revolve around
regression models for clustered and longitudinal data. He has co-authored a
book entitled Linear Mixed Models: A Practical Guide using Statistical
Software (Chapman & Hall / CRC Press) that features the use of the HLM
software to fit hierarchical linear models.
Audience:
The workshop is intended primarily for students, faculty, and researchers, with interests in human development, public health,
sociology, education, and related fields, but is appropriate for all interested persons who meet the prerequisites.
Prerequisite:
A working knowledge of applied multiple regression and analysis of variance is required.
Provisions:
Enrollees will receive substantial handouts, example
computer software output, and a bibliography.
Break time for off-site lunch (lunch will not be provided, but refreshments for
mid-session breaks will be included.)
When:
Monday,
Tuesday, Wednesday, May 19, 20, and 21, 2008,
9:00
a.m. - 5:00 p.m.
Location:
Rackham
Bldg, 3rd
floor, East Seminar Room.
Size:
Up to 20 participants with 1 person per computer
Fee:
Registration until
May 6, 2008:
$300 for University of Michigan affiliated
faculty, staff and students
$650 for others
Registration after May 6, 2008:
$360 for University of Michigan affiliated
faculty, staff and students
$780 for others
Registration:
Call CSCAR at 734-764-7828. Enrollment is
limited. Please make check payable to: CSCAR--University of
Michigan, or give the University
grant/project to be
billed. Send check to CSCAR, 3550 Rackham Bldg., University of Michigan, 915 E. Washington St. Ann Arbor, MI 48109-1070
|
