Research Agenda
Research Methodology
I. Analysis
of TDSB Data
II. Knowledge
Mobilization and Knowledge Transfer
I. Analysis of TDSB Data
The main source of data to be used in the proposed study is the year
2000 Grade 9 cohort of students who began high school in the TDSB in
September 2000 and were tracked through the TDSB until fall 2006. The
data includes more than 18,469 respondents. Of these, 2,220 students
transferred to another educational system outside the TDSB sometime
between Fall 2000 and Fall 2006, leaving a base of 16,249 students
tracked in the cohort until Fall 2006. The TDSB data will be made
available for the project by Robert Brown, Research Coordinator in the
Research and Information Services Department of the Toronto District
School Board.
The TDSB administrative data also contains a wealth of information
relating to the academic achievement of every student in the cohort.
For example, in addition to graduation status (i.e., whether the
student graduated by 2006), the TDSB data includes detailed information
on the courses each student has taken during each year of study
(including ESL), the students’ educational program of study (e.g.,
whether they are enrolled in academic, applied, or essentials programs
of study), course grades, and credit accumulation. We can also assess
whether the respondents started high school on time or a year late
(using the age variable), and if the student entered their high school
from within the TDSB. The data also allows us to track whether students
changed schools and residences during the period of investigation.
School-level variables available in the dataset include school size,
ethnic or language concentration within each school, and a variable
derived by the TDSB called the Learning Opportunities Index (LOI). The
LOI is used to rank schools in terms of the social and economic
characteristics of their school populations. It was developed by the
TDSB in 1999 to ensure that appropriate funding allocations would be
made to schools identified as “challenged.” Since we also have access
to the respondents’ residential information during each year of study,
it is possible to create neighbourhood-level variables (e.g., LICO;
income, ethnic composition etc.) by merging TDSB data with from the
2006 Canadian Census via the students’ residential postal codes.
The post-high school (PHS) outcomes of these students will be obtained
by matching data from the Ontario universities Applications Centre
(OUAC) and the Ontario College Application Centre (OCAS) with the TDSB
Student Information System for the Year 2000 Grade 9 cohort. The OUAC
and OCAS data from the 2004, 2005, and 2006 applications cycles will be
used to identify applications made to programs (e.g., arts, commerce,
engineering, social sciences, etc.) in Ontario universities or
community colleges or, in many cases, both, as well as confirmations of
an offer of acceptance from an Ontario university or community college.
It may also be possible to acquire registration data (i.e., if the
student is formally registered at a postsecondary institution the
following year) from OUAC and OCAS; however, we cannot confirm that we
will have access to that information at this time. The OUAC and OCAS
data will be made available through the TDSB and acquired for the
project through Robert Brown.
Data
Analysis
The TDSB dataset is valuable because its large sample size and the
availability of multiple nesting structures make it possible for us to
employ multi-level models. When multi-level data are available and
there is theoretical justification for employing a multi-level
framework, mixed models (models with fixed and random effects) should
be employed to account for correlated observations within higher-level
units (e.g., schools or neighbourhoods). Failure to account for the
correlated nesting structure can result in inflated Type I error rates
(i.e., rejecting a “correct” null hypothesis). Thus, in addition to
providing
descriptive statistics (means, proportions, cross-tabulations) our
statistical analyses will employ multi-level (mixed) models with
cross-classified nesting structures (e.g., students nested within
schools, and also separately nested within neighbourhoods) for
generalized linear models (i.e., regression models with discrete
response variables). Fortunately, itt is possible to estimate these
models (including models with multinomial-responses) using the most
recent version of Stata (Version 10).
The data permit us to create response (dependent) variables that
capture whether the respondents applied to college and/or university,
and/or whether they were accepted into college and/or university. Thus,
we intend to employ a series of mixed logit, mixed multinomial logit,
or mixed nested logit models, where applicable. The response variables,
and hence the models we employ, will depend on the specific pathways
examined and the objectives outlined for each set of analyses. The
explanatory variables in the regression models can be divided into
separate groups, which will consist of the key variables in the
proposed project (e.g., immigration status and home language),
individual-level variables relating to
other sociodemographic characteristics and school achievement,
neighborhood characteristics, school characteristics, and finally
interactions (e.g., immigrant status by gender) where appropriate. The
specific variables employed in each analysis, and the order that they
are included in the models will be
determined by the objectives set out for each paper. As our research is
intended for policy makers, academic audiences and school board
personnel, all of our statistical results (e.g., regression
coefficients) will be converted into meaningful quantities (e.g.,
predicted probabilities) and presented in
easy-to-understand graphical displays.
II. Knowledge
Mobilization and Knowledge Transfer: The Participation of School Boards
in Ontario
The use of administrative secondary school data for the purposes of
tracking post-secondary participation among immigrant and native born
youth need not be limited to the TDSB. Building on the research
procedures elaborated in the previous section, we suggest that school
boards from across Ontario be invited to participate. Given that at
this time few school boards have developed procedures or made links to
post-secondary data relating to confirmations/acceptances, the TDSB
process (a specific process linking OUAC and OCAS records to existing
board data) will be used as a template. It is the belief of the
research team that effective */knowledge transfer/* of
policy/practice-relevant research findings requires the building of
partnership networks and infrastructure that permit, facilitate and
support ongoing, systematic, timely exchange of social science
knowledge between academic and non-academic stakeholders.
Furthermore, we see the overall knowledge exchange/ /process as an
active, two-way exchange of information and people between knowledge
creators, knowledge brokers and knowledge users. For these reasons we
will work closely with our two partners—AERO and HEQCO-- in building an
infrastructure that will facilitate the exchange of information among
school boards with regard to application and use of the TDSB template.
This process will be further facilitated through the creation of a
website and the assignment of a graduate student who will work with
boards on knowledge mobilization and transfer activities in conjunction
with the PI. In addition, existing websites for AERO and HEQCO will be
employed to disseminate information regarding project developments. We
will work closely with school boards to address issues of data
security, confidentiality and identification. For example, schools
boards will not be obligated to provided data sets to anyone outside
their board; school boards will be expected only to share their process
documentation.
The Knowledge Mobilization component would be developed with AERO. AERO
is a unique partner for several reasons. It is the provincial
association most closely associated with school board research and
researchers. Furthermore, AERO worked with OUAC in putting together a
format for transferring university application information that can be
used by all school boards in Ontario; a next step (during the lifespan
of this research project) would be to do the same with OCAS for
community college information. School board members with an interest in
looking at the information have formed an AERO sub-committee that met
in June 2008 for the first time. This research project therefore
dovetails with an authentic knowledge transfer process in its very
beginning stages.
It is anticipated that there would be four stages over 3 years. In
Stage 1, AERO and this subcommittee would act as the main Ontario-wide
knowledge facilitators in details of the matching process, such as
downloading university applications information from OUAC, matching it
with board information, and looking at key patterns such as who applies
to different programs and who does not apply to any program. Stage 2
would continue this process with community college (OCAS) applications
as well as university applications. Stage 3 would involve matching this
combined information with cohort studies that would be developed by the
boards. Stage 4 would involve looking at the ‘indirect’ transition of
board students into post-secondary as adult students.
|
|
|
|
