Passenger data from the RMS Titanic disaster of April 15, 1912, containing survival status and demographic information for 1,309 passengers. This widely used dataset illustrates logistic regression with both continuous and categorical predictors, and serves as a compelling teaching example of how passenger class, sex, and age influenced survival probability.
Format
A tibble with 1,309 rows and 6 columns:
- pclass
Passenger class (ticket class). Type: numeric. Values: 1 = 1st class (upper), 2 = 2nd class (middle), 3 = 3rd class (lower). Serves as a proxy for socioeconomic status.
- survived
Survival indicator. Type: numeric. Binary (0/1) where 1 = survived, 0 = did not survive. Overall survival rate is approximately 38 percent.
- age
Passenger age in years. Type: numeric. Range: (0.17, 80.00). 263 missing values. Fractional ages (e.g., 0.17) represent infants.
- fare
Passenger fare in British pounds. Type: numeric. Range: (0.00, 512.33). 1 missing value. Varies substantially by class.
- embarked
Port of embarkation. Type: character. Values: S = Southampton, C = Cherbourg, Q = Queenstown (Cobh, Ireland). Two cases have missing embarkation port.
- sex
Passenger sex indicator. Type: numeric. Binary (0/1) where 1 = female, 0 = male.
Source
British Board of Trade (1990). Report on the Loss of the
"Titanic" (S.S.). Compiled from various historical passenger records.
Original data file: titanic.dta
Details
This dataset is used in Chapters 9-11 (Logistic Regression) to illustrate simple and multiple logistic regression modeling. The well-known "women and children first" evacuation policy makes this dataset effective for demonstrating how sex, class, and age predict binary outcomes. Key analyses include: simple logistic regression of survival on sex, multiple logistic regression with class and age, odds ratio interpretation, predicted probabilities, handling missing data in age, and model comparison using likelihood ratio tests and AIC.
Examples
data(titanic)
head(titanic)
#> # A tibble: 6 × 6
#> pclass survived age fare embarked sex
#> <int> <int> <dbl> <dbl> <chr> <int>
#> 1 1 1 29 211. S 1
#> 2 1 1 0.917 152. S 0
#> 3 1 0 2 152. S 1
#> 4 1 0 30 152. S 0
#> 5 1 0 25 152. S 1
#> 6 1 1 48 26.5 S 0
# Logistic regression: survival predicted by sex and passenger class
glm(survived ~ sex + factor(pclass), family = binomial, data = titanic)
#>
#> Call: glm(formula = survived ~ sex + factor(pclass), family = binomial,
#> data = titanic)
#>
#> Coefficients:
#> (Intercept) sex factor(pclass)2 factor(pclass)3
#> -0.4059 2.5150 -0.8808 -1.7231
#>
#> Degrees of Freedom: 1308 Total (i.e. Null); 1305 Residual
#> Null Deviance: 1741
#> Residual Deviance: 1257 AIC: 1265