Presentation Title

Logistic Regression Analysis of Onset Puberty Growth Spurt Data

Faculty Mentor

Jessica Jaynes

Start Date

23-11-2019 10:45 AM

End Date

23-11-2019 11:30 AM

Location

254

Session

poster 4

Type of Presentation

Poster

Subject Area

physical_mathematical_sciences

Abstract

Height and body mass index (BMI) have been shown to be correlated to earlier rates of childhood pubescence. Children who experience precocious, or early puberty, are seemingly prone to develop diseases such as breast cancer, ovarian cancer, and cardiovascular disease. Based on a previous published study on Polish children, a multiple linear regression model was used to determine that height and BMI have an effect on the age of pubescence. In an effort to determine whether Polish children experienced on-time or early puberty, we propose the use of a logistic regression model by creating a binary categorization for the age of pubescence. We consider the effect of BMI standard deviation scores (SDS) and height gap SDS to model the log-odds of on-time and early puberty and compare the impact of height gap SDS and BMI SDS by gender. With this research we found that height gap SDS had a significant effect on the predicted probability of early puberty for both boys and girls, however, BMI SDS only had a significant effect on the predicted probability of early puberty for girls.

This document is currently not available here.

Share

COinS
 
Nov 23rd, 10:45 AM Nov 23rd, 11:30 AM

Logistic Regression Analysis of Onset Puberty Growth Spurt Data

254

Height and body mass index (BMI) have been shown to be correlated to earlier rates of childhood pubescence. Children who experience precocious, or early puberty, are seemingly prone to develop diseases such as breast cancer, ovarian cancer, and cardiovascular disease. Based on a previous published study on Polish children, a multiple linear regression model was used to determine that height and BMI have an effect on the age of pubescence. In an effort to determine whether Polish children experienced on-time or early puberty, we propose the use of a logistic regression model by creating a binary categorization for the age of pubescence. We consider the effect of BMI standard deviation scores (SDS) and height gap SDS to model the log-odds of on-time and early puberty and compare the impact of height gap SDS and BMI SDS by gender. With this research we found that height gap SDS had a significant effect on the predicted probability of early puberty for both boys and girls, however, BMI SDS only had a significant effect on the predicted probability of early puberty for girls.