Chapter11 {DanielBiostatistics10th} R Documentation

Chapter 11: Regression Analysis: Some Additional Techniques

Description

Examples in Chapter 11, Regression Analysis: Some Additional Techniques.

Value

No function defined for Chapter 11.

Examples

library(DanielBiostatistics10th)
library(car) # ?car::scatterplot
library(DescTools)

# Example 11.1.1; Page 540,
head(log(EXA_C11_S01_01$conc, base = 10)) head(EXA_C11_S01_01$logConc)

# Example 11.1.2; Page 542,
cor.test(~ sbp + weight, data = EXA_C11_S01_02)
cor.test(~ sbp + bmi, data = EXA_C11_S01_02)

# Example 11.2.1; Page 545,
d1121 = within(EXA_C11_S02_01, expr = {
SMOKE = as.logical(SMOKE)
})
xlab1121 = 'Length of gestation (weeks)'; ylab1121 = 'Birth weight (grams)'
scatterplot(GRAMS ~ WEEKS | SMOKE, data = d1121, regLine = FALSE, smooth = FALSE,
xlab = xlab1121, ylab = ylab1121, main = 'Figure 11.2.1')
summary(m1121_main <- lm(GRAMS ~ WEEKS + SMOKE, data = d1121)) # Figure 11.2.2
confint(m1121_main)
scatterplot(GRAMS ~ WEEKS | SMOKE, data = d1121, regLine = FALSE, smooth = FALSE,
xlab = xlab1121, ylab = ylab1121, main = 'Figure 11.2.3')
(cf_main = m1121_main$coefficients) abline(a = cf_main[1L], b = cf_main[2L], col = 'blue') # regression line for non-smoking mothers abline(a = cf_main[1L] + cf_main[3L], b = cf_main[2L], col = 'magenta') # Example 11.2.3; Page 551, d1123 = within(EXA_C11_S02_03, expr = { METHOD = factor(METHOD, levels = c('C', 'A', 'B')) # textbook designated 'C' as reference level }) summary(mod_1123 <- lm(EFFECT ~ AGE * METHOD, data = d1123)) # Figure 11.2.5 confint(mod_1123) scatterplot(EFFECT ~ AGE | METHOD, data = d1123, smooth = FALSE, xlab = 'Age', ylab = 'Treatment effectiveness', main = 'Figure 11.2.6') # Example 11.3.1; Page 561, head(EXA_C11_S03_01) names(EXA_C11_S03_01) = c('JOBPER', 'ASRV', 'ENTH', 'AMB', 'COMM', 'PROB', 'INIT') summary(mod_1131_raw <- lm(JOBPER ~ ASRV + ENTH + AMB + COMM + PROB + INIT, data = EXA_C11_S03_01)) # summary(mod_1131 <- MASS::stepAIC(mod_1131_raw, direction = 'backward')) # the stepwise selection criterion used in MINITAB is not necessarily AIC # Example 11.4.1; Page 572, addmargins(d1141 <- array(c(92L, 21L, 15L, 20L), dim = c(2L, 2L), dimnames = list( OCAD = c('Present', 'Absent'), Sex = c('Male', 'Female')))) # Table 11.4.2 (d1141a = within(as.data.frame(as.table(d1141)), expr = { OCAD = (OCAD == 'Present') Sex = factor(Sex, levels = c('Female', 'Male')) })) summary(m1141 <- glm(OCAD ~ Sex, family = binomial, weights = Freq, data = d1141a)) # Figure 11.4.1 exp(m1141$coefficients[2L]) # exp(beta_M)
exp(confint(m1141)) # confidence interval of exp(beta)
predict(m1141, newdata = data.frame(Sex = setNames(nm = c('Male', 'Female'))), type = 'response')

# Example 11.4.2; Page 573,
summary(mod_1142 <- glm(ATT ~ AGE, family = binomial, data = EXA_C11_S04_02)) # Figure 11.4.2
exp(mod_1142\$coefficients[2L])
exp(confint(mod_1142))
car::Anova(mod_1142) # Optional

# Example 11.4.3; Page 576,