Online Program Home
My Program

Abstract Details

Activity Number: 621 - Beyond Linear Regression: Nonlinear Association, Quantile Regression and Generalized Linear Models
Type: Contributed
Date/Time: Thursday, August 1, 2019 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics in Epidemiology
Abstract #304392
Title: Look at the Whole Picture: Quantile Regression in Developmental Disabilities Research
Author(s): Lin Tian*
Companies: CDC
Keywords: quantile regression; developmental disabilities; fine motor ; heterogeneous; linear regression

Developmental disabilities data, such as fine motor (FM) scores, typically are not normally distributed and can be highly heterogeneous. Commonly used statistical methods such as linear regression (LR) and ANOVA assume a normal distribution and are population-averaged models that do not fully specify the entire conditional distribution of the population. This talk demonstrates the advantages of quantile regression (QR) by comparing QR to LR in analyzing the relationship between dysmorphology severity (DS) and FM scores among 514 children with a developmental disability. I analyzed the relationship at quantile levels (QL) of 0.1, 0.3, 0.5, 0.7, and 0.9 of FM scores using QR, and at the mean using LR. Even though the FM data were not normally distributed, the findings from both QR and LR analyses indicate that the FM is negatively associated to DS. However, QR analysis further showed that the negative relationship tended to be stronger among children with lower FM scores (< 0.7 QL) and was not significant among children with a high FM score (< 0.9 QL). The results demonstrate that QR is a valuable tool and can provide a more comprehensive picture of the relationship.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program