case2002: Birdkeeping and Lung Cancer
In Sleuth3: Data Sets from Ramsey and Schafer's "Statistical Sleuth (3rd Ed)"
Description
Usage
Format
Source
References
Examples
Description
A 1972–1981 health survey in The Hague, Netherlands, discovered an
association between keeping pet birds and increased risk of lung
cancer. To investigate birdkeeping as a risk factor, researchers
conducted a case–control study of patients in 1985 at four
hospitals in The Hague (population 450,000). They identified 49 cases
of lung cancer among the patients who were registered with a general
practice, who were age 65 or younger and who had resided in the city
since 1965. They also selected 98 controls from a population of
residents having the same general age structure.
Usage
1
Format
A data frame with 147 observations on the following 7 variables.
- LC
Whether subject has lung cancer
- FM
Sex of subject
- SS
Socioeconomic status, determined by occupation of the
household's principal wage earner
- BK
Indicator for birdkeeping (caged birds in the home for
more that 6 consecutive months from 5 to 14 years before diagnosis
(cases) or examination (control))
- AG
Age of subject (in years)
- YR
Years of smoking prior to diagnosis or examination
- CD
Average rate of smoking (in cigarettes per day)
Source
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A
Course in Methods of Data Analysis (3rd ed), Cengage Learning.
References
Holst, P.A., Kromhout, D. and Brand, R. (1988). For Debate: Pet Birds
as an Independent Risk Factor for Lung Cancer, British Medical
Journal 297: 13–21.
Examples
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
str(case2002)
attach(case2002)
## EXPLORATION AND MODEL BUILDING
myCode <- ifelse(BK=="Bird" & LC=="LungCancer","Bird & Cancer",
ifelse(BK=="Bird" & LC=="NoCancer","Bird & No Cancer",
ifelse(BK=="NoBird" & LC=="LungCancer","No Bird & Cancer", "No Bird & No Cancer")))
table(myCode)
if(require(car)){ # Use the car library
scatterplotMatrix(cbind(AG,YR,CD), groups=myCode, diagonal="none",reg.line=FALSE,
pch=c(15,21,15,21), col=c("dark green","dark green","purple","purple"),
var.labels=c("Age","Years Smoked","Cigarettes per Day"), cex=1.5)
}
# Reorder the levels so that the model is for log odds of cancer
LC <- factor(LC, levels=c("NoCancer","LungCancer"))
myGlm <- glm(LC ~ FM + SS + AG + YR + CD + BK, family=binomial)
if(require(car)){ # Use the car library
crPlots(myGlm) }
# It appears that there's an effect of Years of Smoking and of Bird Keeping
# after accounting for other variables; no obvious effects of other variables
# Logistic regression model building using backward elimination (witholding BK)
myGlm1 <- glm(LC ~ FM + SS + AG + YR + CD, family=binomial)
summary(myGlm1)
myGlm2 <- update(myGlm1, ~ . - SS)
summary(myGlm2)
myGlm3 <- update(myGlm2, ~ . - CD)
summary(myGlm3)
myGlm4 <- update(myGlm3, ~ . - FM)
summary(myGlm4) # Everything left has a small p-value (retain the intercept)
## INFERENCE AND INTERPRETATION
myGlm5 <- update(myGlm4, ~ . + BK) # Now add bird keeping
summary(myGlm5)
myGlm6 <- update(myGlm5, ~ . + BK:YR + AG:YR) # Try interaction terms
anova(myGlm6,myGlm5) # Drop-in-deviance = 1.61 on 2 d.f.
1 - pchisq(1.61,2) # p-value = .45: no evidence of interaction
anova(myGlm4,myGlm5) # Test for bird keeping effect
(1 - pchisq(12.612,1))/2 # 1-sided p-value: 0.0001916391
BK <- factor(BK, levels=c("NoBird", "Bird")) # Make "no bird" the ref level
myGlm5b <- glm(LC ~ AG + YR + BK, family=binomial)
beta <- myGlm5b$coef # Extract estimated coefficients
exp(beta[4]) # 3.961248
exp(confint(myGlm5b,4)) # 1.836764 8.900840
# Interpretation: The odds of lung cancer for people who kept birds were
# estimated to be 4 times the odds of lung cancer for people of similar age, sex,
# smoking history, and socio-economic status who didn't keep birds
# (95% confidence interval for this adjusted odds ratio: 1.8 times to 8.9 times).
# See bestglm library for an alternative variable selection technique.
detach(case2002)
Example output
'data.frame': 147 obs. of 7 variables:
$ LC: Factor w/ 2 levels "LungCancer","NoCancer": 1 1 1 1 1 1 1 1 1 1 ...
$ FM: Factor w/ 2 levels "Female","Male": 2 2 2 2 2 2 2 2 2 2 ...
$ SS: Factor w/ 2 levels "High","Low": 2 2 1 2 2 1 1 2 2 1 ...
$ BK: Factor w/ 2 levels "Bird","NoBird": 1 1 2 1 1 2 1 2 1 2 ...
$ AG: int 37 41 43 46 49 51 52 53 56 56 ...
$ YR: int 19 22 19 24 31 24 31 33 33 26 ...
$ CD: int 12 15 15 15 20 15 20 20 10 25 ...
myCode
Bird & Cancer Bird & No Cancer No Bird & Cancer No Bird & No Cancer
33 34 16 64
Loading required package: car
Call:
glm(formula = LC ~ FM + SS + AG + YR + CD, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.3910 -0.9718 -0.5519 1.1733 2.5020
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.37895 1.67206 0.227 0.82070
FMMale -0.74923 0.50501 -1.484 0.13792
SSLow 0.07303 0.43893 0.166 0.86785
AG -0.05799 0.03432 -1.690 0.09112 .
YR 0.07955 0.02636 3.018 0.00255 **
CD 0.01978 0.02422 0.817 0.41421
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 165.87 on 141 degrees of freedom
AIC: 177.87
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ FM + AG + YR + CD, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.4134 -0.9744 -0.5430 1.1749 2.5123
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.46101 1.59688 0.289 0.77282
FMMale -0.76832 0.49178 -1.562 0.11821
AG -0.05858 0.03415 -1.715 0.08628 .
YR 0.08027 0.02603 3.083 0.00205 **
CD 0.01959 0.02420 0.810 0.41820
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 165.90 on 142 degrees of freedom
AIC: 175.9
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ FM + AG + YR, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.2597 -0.9794 -0.5462 1.1718 2.4894
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.82886 1.52662 0.543 0.587172
FMMale -0.73638 0.48914 -1.505 0.132210
AG -0.06363 0.03359 -1.894 0.058195 .
YR 0.08776 0.02452 3.579 0.000344 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 166.55 on 143 degrees of freedom
AIC: 174.55
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ AG + YR, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.2933 -0.9869 -0.5682 1.2448 2.5943
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.67653 1.49597 0.452 0.651100
AG -0.06568 0.03291 -1.996 0.045976 *
YR 0.07815 0.02321 3.368 0.000758 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 168.83 on 144 degrees of freedom
AIC: 174.83
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ AG + YR + BK, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.5466 -0.8649 -0.4911 0.9763 2.2584
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.34296 1.58002 0.217 0.828159
AG -0.04610 0.03430 -1.344 0.178952
YR 0.07485 0.02296 3.261 0.001111 **
BKNoBird -1.37656 0.40073 -3.435 0.000592 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 156.22 on 143 degrees of freedom
AIC: 164.22
Number of Fisher Scoring iterations: 5
Analysis of Deviance Table
Model 1: LC ~ AG + YR + BK + YR:BK + AG:YR
Model 2: LC ~ AG + YR + BK
Resid. Df Resid. Dev Df Deviance
1 141 154.60
2 143 156.22 -2 -1.6163
[1] 0.4470879
Analysis of Deviance Table
Model 1: LC ~ AG + YR
Model 2: LC ~ AG + YR + BK
Resid. Df Resid. Dev Df Deviance
1 144 168.83
2 143 156.22 1 12.612
[1] 0.0001916391
BKBird
3.961248
Waiting for profiling to be done...
2.5 % 97.5 %
1.836764 8.900840
Sleuth3 documentation built on May 2, 2019, 6:41 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Description Usage Format Source References Examples
Description
A 1972–1981 health survey in The Hague, Netherlands, discovered an association between keeping pet birds and increased risk of lung cancer. To investigate birdkeeping as a risk factor, researchers conducted a case–control study of patients in 1985 at four hospitals in The Hague (population 450,000). They identified 49 cases of lung cancer among the patients who were registered with a general practice, who were age 65 or younger and who had resided in the city since 1965. They also selected 98 controls from a population of residents having the same general age structure.
Usage
1 |
Format
A data frame with 147 observations on the following 7 variables.
- LC
Whether subject has lung cancer
- FM
Sex of subject
- SS
Socioeconomic status, determined by occupation of the household's principal wage earner
- BK
Indicator for birdkeeping (caged birds in the home for more that 6 consecutive months from 5 to 14 years before diagnosis (cases) or examination (control))
- AG
Age of subject (in years)
- YR
Years of smoking prior to diagnosis or examination
- CD
Average rate of smoking (in cigarettes per day)
Source
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
References
Holst, P.A., Kromhout, D. and Brand, R. (1988). For Debate: Pet Birds as an Independent Risk Factor for Lung Cancer, British Medical Journal 297: 13–21.
Examples
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 | str(case2002)
attach(case2002)
## EXPLORATION AND MODEL BUILDING
myCode <- ifelse(BK=="Bird" & LC=="LungCancer","Bird & Cancer",
ifelse(BK=="Bird" & LC=="NoCancer","Bird & No Cancer",
ifelse(BK=="NoBird" & LC=="LungCancer","No Bird & Cancer", "No Bird & No Cancer")))
table(myCode)
if(require(car)){ # Use the car library
scatterplotMatrix(cbind(AG,YR,CD), groups=myCode, diagonal="none",reg.line=FALSE,
pch=c(15,21,15,21), col=c("dark green","dark green","purple","purple"),
var.labels=c("Age","Years Smoked","Cigarettes per Day"), cex=1.5)
}
# Reorder the levels so that the model is for log odds of cancer
LC <- factor(LC, levels=c("NoCancer","LungCancer"))
myGlm <- glm(LC ~ FM + SS + AG + YR + CD + BK, family=binomial)
if(require(car)){ # Use the car library
crPlots(myGlm) }
# It appears that there's an effect of Years of Smoking and of Bird Keeping
# after accounting for other variables; no obvious effects of other variables
# Logistic regression model building using backward elimination (witholding BK)
myGlm1 <- glm(LC ~ FM + SS + AG + YR + CD, family=binomial)
summary(myGlm1)
myGlm2 <- update(myGlm1, ~ . - SS)
summary(myGlm2)
myGlm3 <- update(myGlm2, ~ . - CD)
summary(myGlm3)
myGlm4 <- update(myGlm3, ~ . - FM)
summary(myGlm4) # Everything left has a small p-value (retain the intercept)
## INFERENCE AND INTERPRETATION
myGlm5 <- update(myGlm4, ~ . + BK) # Now add bird keeping
summary(myGlm5)
myGlm6 <- update(myGlm5, ~ . + BK:YR + AG:YR) # Try interaction terms
anova(myGlm6,myGlm5) # Drop-in-deviance = 1.61 on 2 d.f.
1 - pchisq(1.61,2) # p-value = .45: no evidence of interaction
anova(myGlm4,myGlm5) # Test for bird keeping effect
(1 - pchisq(12.612,1))/2 # 1-sided p-value: 0.0001916391
BK <- factor(BK, levels=c("NoBird", "Bird")) # Make "no bird" the ref level
myGlm5b <- glm(LC ~ AG + YR + BK, family=binomial)
beta <- myGlm5b$coef # Extract estimated coefficients
exp(beta[4]) # 3.961248
exp(confint(myGlm5b,4)) # 1.836764 8.900840
# Interpretation: The odds of lung cancer for people who kept birds were
# estimated to be 4 times the odds of lung cancer for people of similar age, sex,
# smoking history, and socio-economic status who didn't keep birds
# (95% confidence interval for this adjusted odds ratio: 1.8 times to 8.9 times).
# See bestglm library for an alternative variable selection technique.
detach(case2002)
|
Example output
'data.frame': 147 obs. of 7 variables:
$ LC: Factor w/ 2 levels "LungCancer","NoCancer": 1 1 1 1 1 1 1 1 1 1 ...
$ FM: Factor w/ 2 levels "Female","Male": 2 2 2 2 2 2 2 2 2 2 ...
$ SS: Factor w/ 2 levels "High","Low": 2 2 1 2 2 1 1 2 2 1 ...
$ BK: Factor w/ 2 levels "Bird","NoBird": 1 1 2 1 1 2 1 2 1 2 ...
$ AG: int 37 41 43 46 49 51 52 53 56 56 ...
$ YR: int 19 22 19 24 31 24 31 33 33 26 ...
$ CD: int 12 15 15 15 20 15 20 20 10 25 ...
myCode
Bird & Cancer Bird & No Cancer No Bird & Cancer No Bird & No Cancer
33 34 16 64
Loading required package: car
Call:
glm(formula = LC ~ FM + SS + AG + YR + CD, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.3910 -0.9718 -0.5519 1.1733 2.5020
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.37895 1.67206 0.227 0.82070
FMMale -0.74923 0.50501 -1.484 0.13792
SSLow 0.07303 0.43893 0.166 0.86785
AG -0.05799 0.03432 -1.690 0.09112 .
YR 0.07955 0.02636 3.018 0.00255 **
CD 0.01978 0.02422 0.817 0.41421
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 165.87 on 141 degrees of freedom
AIC: 177.87
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ FM + AG + YR + CD, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.4134 -0.9744 -0.5430 1.1749 2.5123
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.46101 1.59688 0.289 0.77282
FMMale -0.76832 0.49178 -1.562 0.11821
AG -0.05858 0.03415 -1.715 0.08628 .
YR 0.08027 0.02603 3.083 0.00205 **
CD 0.01959 0.02420 0.810 0.41820
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 165.90 on 142 degrees of freedom
AIC: 175.9
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ FM + AG + YR, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.2597 -0.9794 -0.5462 1.1718 2.4894
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.82886 1.52662 0.543 0.587172
FMMale -0.73638 0.48914 -1.505 0.132210
AG -0.06363 0.03359 -1.894 0.058195 .
YR 0.08776 0.02452 3.579 0.000344 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 166.55 on 143 degrees of freedom
AIC: 174.55
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ AG + YR, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.2933 -0.9869 -0.5682 1.2448 2.5943
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.67653 1.49597 0.452 0.651100
AG -0.06568 0.03291 -1.996 0.045976 *
YR 0.07815 0.02321 3.368 0.000758 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 168.83 on 144 degrees of freedom
AIC: 174.83
Number of Fisher Scoring iterations: 5
Call:
glm(formula = LC ~ AG + YR + BK, family = binomial)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.5466 -0.8649 -0.4911 0.9763 2.2584
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.34296 1.58002 0.217 0.828159
AG -0.04610 0.03430 -1.344 0.178952
YR 0.07485 0.02296 3.261 0.001111 **
BKNoBird -1.37656 0.40073 -3.435 0.000592 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 187.14 on 146 degrees of freedom
Residual deviance: 156.22 on 143 degrees of freedom
AIC: 164.22
Number of Fisher Scoring iterations: 5
Analysis of Deviance Table
Model 1: LC ~ AG + YR + BK + YR:BK + AG:YR
Model 2: LC ~ AG + YR + BK
Resid. Df Resid. Dev Df Deviance
1 141 154.60
2 143 156.22 -2 -1.6163
[1] 0.4470879
Analysis of Deviance Table
Model 1: LC ~ AG + YR
Model 2: LC ~ AG + YR + BK
Resid. Df Resid. Dev Df Deviance
1 144 168.83
2 143 156.22 1 12.612
[1] 0.0001916391
BKBird
3.961248
Waiting for profiling to be done...
2.5 % 97.5 %
1.836764 8.900840
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.