Dangers of Smoking ACM 270-1: Introduction to Statistical Inference Lecture 2b, Jan 6, 2016 Agenda • Maternal Smoking and Infant Health • Background • Data • Summarizing Data 01/06/2016 ACM 270-1, Lecture 2b: Dangers of Smoking 2 Background: Fetal Development 37w preterm 42w 40w typical gestation period brain damage is likely due to deterioration of the placenta • Placenta is a special organ that develops during pregnancy • Transfers • Nutrients and oxygen from the mother to the fetus • Waste products and carbon dioxide back to the maternal blood • Fetus is attached to the placenta by its umbilical cord • Gestational age is a measure of the baby’s maturity • Another measure is the birth weight • Smaller babies and babies born early have lower survival rates Source: http://cdn.lifemartini.com/wp-content/uploads/2013/05/Placenta-Previa.jpg 01/06/2016 ACM 270-1, Lecture 2b: Dangers of Smoking 3 Birth Weight Babies born at term with weight < 2.5kg are considered small for their gestational age 4kg Sophia, Hiking in New England 3.8kg • Many pregnant women still smoke • National Center for Health Statistics: • In 1996: 15% mothers smoked during pregnancy • In 2013: 8% mothers smoked during pregnancy 2.5kg 2.3kg 1.8kg ~90g/w 28w 01/06/2016 32w birth ACM 270-1, Lecture 2b: Dangers of Smoking 4 Source: http://images.tutorvista.com/cms/im ages/38/molecular-orbital-theory.jpg Smoke • Contains many chemical agents • We focus on: carbon monoxide CO • colorless, tasteless gas, slightly less dense than air • toxic to humans (in concentrations above about 35ppm) • in smoke, reduces the oxygen supplied to the fetus • A steady oxygen supply is critical for the developing baby • Merkatz & Thompson (1990): • smoking responsible for a 150-250 gram reduction in birth weight • smoking mothers are twice as likely as nonsmoking mothers to have a low-birth-weight baby (<2.5kg) Source: http://pngimg.com/upload/smoke_PNG964.png 01/06/2016 ACM 270-1, Lecture 2b: Dangers of Smoking 5 Data: birth.txt • Subset of CHDS: Child Health and Development Studies • Pregnancies occurred between 1960-1967 in Oakland, CA • birth.txt: data for 1236 babies 01/06/2016 ACM 270-1, Lecture 2b: Dangers of Smoking 6 Histograms • histogram(data,nbins)creates a not normalized histogram • histogram(data,nbins,’Normalization’,’pdf’) creates a normalized (total area is 1) histogram of data small babies outliers One mode, symmetric, approximately bell-shaped 01/06/2016 One mode, not symmetric, right-skewed ACM 270-1, Lecture 2b: Dangers of Smoking 7 Numerical Summaries Location • mean(x) returns the sample mean of x • median(x) gives the median of x • trimmean(x,200α) is the α-trimmed mean Spread • range(x) is range of x • std(x,flag) is the standard deviation • flag=0 ⇒ = 3.40 = 3.43 = 0.52 • flag=1 ⇒ • iqr(x) computes the interquartile range 01/06/2016 = 3.39 ACM 270-1, Lecture 2b: Dangers of Smoking = 0.64 8 Boxplot, Empirical CDF, QQ-plot boxplot(x) 01/06/2016 ecdf(x) ACM 270-1, Lecture 2b: Dangers of Smoking qqplot(x) 9 Smokers vs Nonsmokers • Sample mean difference: ≈ 0.25 𝑘𝑘𝑘𝑘 (8.8 𝑜𝑜𝑜𝑜) 01/06/2016 ACM 270-1, Lecture 2b: Dangers of Smoking 10