Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Doctoral Award - Frederick Banting and Charles Best Canada Graduate Scholarships

University of Toronto

$105,000.00

A modified regression approach for modeling small sample and non-normal distributed longitudinal outcome

FY 2016-17

Summary
The relationship between the growth trajectories, rates of change of a longitudinal measurement and a future outcome are usually of great interest in epidemiology and clinical research. Many statistical modelling approaches have been proposed over the past decades to model the growth trajectories, such as regression calibration methods, bivariate growth methods and structural equation modellings. However, these existing methods require the normality assumptions to be made to estimate the coefficients and standard errors, and may only be appropriate for data with large sample size. Misuses of these approaches may result in bias estimates of parameters of interest and thus lead to misleading conclusions. My research aims to investigate the properties of different existing modeling approaches under the scenarios when the sample size is small and the outcome is not distributed normally and to develop a novel regression approach to overcome the aforementioned issues which usually encounters in real data. The statistical methods evaluated and the new proposed method will be illustrated using the real data from an ongoing general population longitudinal study, the Canadian Healthy Infant Longitudinal Development (CHILD) Study. The findings from my research will provide accurate and stable estimates when modeling longitudinal data with small sample size and non-normal distributed outcomes and aid in the discovery of lung protective and early diagnostic strategies.
Program
Doctoral Award - Frederick Banting and Charles Best Canada Graduate Scholarships
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
365478
Name
Lu Zihang
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Population and Public Health
Theme2
Social / Cultural / Environmental / Population Health
Prcname2
Doctoral Research Awards - B
Projecttitle
A modified regression approach for modeling small sample and non-normal distributed longitudinal outcome
Programtype2
Studentship Programs
Orgnameinp2
University of Toronto
Cihramount2
$105,000
Cihrequipment2
$0
Programname2
Doctoral Award - Frederick Banting and Charles Best Canada Graduate Scholarships
Keyworddelim
Biostatistics; Growth Trajectories Modeling; Longitudinal Data Analysis; Lung Function Development; Non-Normal Distributed Outcomes; Simulation Studies; Small Sample Size
Abstract
The relationship between the growth trajectories, rates of change of a longitudinal measurement and a future outcome are usually of great interest in epidemiology and clinical research. Many statistical modelling approaches have been proposed over the past decades to model the growth trajectories, such as regression calibration methods, bivariate growth methods and structural equation modellings. However, these existing methods require the normality assumptions to be made to estimate the coefficients and standard errors, and may only be appropriate for data with large sample size. Misuses of these approaches may result in bias estimates of parameters of interest and thus lead to misleading conclusions. My research aims to investigate the properties of different existing modeling approaches under the scenarios when the sample size is small and the outcome is not distributed normally and to develop a novel regression approach to overcome the aforementioned issues which usually encounters in real data. The statistical methods evaluated and the new proposed method will be illustrated using the real data from an ongoing general population longitudinal study, the Canadian Healthy Infant Longitudinal Development (CHILD) Study. The findings from my research will provide accurate and stable estimates when modeling longitudinal data with small sample size and non-normal distributed outcomes and aid in the discovery of lung protective and early diagnostic strategies.
Competitiondate
201610
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Lu, Zihang
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $105,000~ $0~