A Multiple Regression Model for Weeks of Gestation at Delivery I ntroduction The main objective of the

A Multiple Regression Model for Weeks of Gestation at Delivery

I ntroduction

The main objective of the project is to use the provided Department of Public Health data set provided to carry out various statistical analyses, including Multiple Regression Analysis, Contingency Table Analysis and Factorial ANOVA. One of the objectives of this paper is to build a meaningful model to predict Weeks of Gestation at Delivery (response variable) in terms of various variables included in the dataset, such as MAT_AGE, MARRIED, MAT_WIC, MM_DIAB, MM_HBP and MOMSMOKE. The second objective of the analysis is to assess through a contingency table analysis (using a Chi-Square test) whether maternal diabetes is related to maternal high blood pressure or not. The final objective of the analysis is to assess whether or not Infant Weight at birth is affected by whether or not the mother is on WIC and by maternal history of smoking.

For the purpose of the analysis the Department of Public Health data set will used, which is a dataset that consists of 38 variables and 1137 cases. The variables that will be used for all the analyses required are (some variables have been recoded for the purpose of the analysis):

GEST_WK: Weeks Gestation at Delivery (measured at the ratio level)
GRAM: Infant Weight in grams (measured at the ratio level)
MARRIED: Mother Married (1 = Yes, 0 = No, measured at the nominal level)
MAT_AGE: Maternal Age (measured at the ratio level)
MAT_WIC: Mother on WIC (1 = Yes, 0 = No, measured at the nominal level)
MM_DIAB: Maternal Diabetes (1 = Yes, 0 = No, measured at the nominal level)
MM_HBP: Maternal High Blood Pressure (1 = Yes, 0 = No, measured at the nominal level)
MOMSMOKE: Maternal History of Smoking (1 = Yes, 0 = No, measured at the nominal level)