4 (20 pts.) Recall the accident data, where the response is 1 if there was an ac

Question

4 (20 pts.) Recall the accident data, where the response is 1 if there was an accident last year and 0 otherwise, vision is 0 if there are no vision problems and 1 otherwise, drivers ed is 1 if yes and 0 otherswise, and age is given in years. Define AGEGROUP by the data statement: DATA LOGISTIC; INFILE '/homes/bnk/driver'; INPUT ACCIDENT AGE VISION DRIVE_ED; IF AGE 65 THEN AGEGROUP=1; ELSE AGEGROUP=0; DATALINES; Running logistic regression using "DATA LOGISTIC", the AIC se- lects the model log(p/(1-p)) =-1.3334 + 2.1611 x AGEGROUP + 1.6258 x VISION. Compute the odds for an accident for a driver with poor vision and belonging to the "young and old" age group, and also for "middle age" drivers with good vision. Then get the odds ratio. Draw your conclusions based on the odds ratio.

Explanation / Answer

The equation is

log(odds) = -1.3334 + 2.1611(Age grp) + 1.6258(vision)

for young and old, age grp = 1 and middle age = 0.

for poor vision, vision = 1 and 0 otherwise

Thus, odds for young and old grp is log(odds) = -1.3334 + 2.1611 + 1.6258 ===> odds = e^(2.4535) = 11.63

Odds for middle age and good vision is log(odds) = -1.3334 ====> odds = e^{-1.3334} = 0.26.

The odds ratio is 44.73 (11.63/0.26). Thus, young and old group with poor vision are 44 times more likely to get into an accident than middle age grp with good vision.

Related Questions

Navigate

• Browse (All)
• Browse 4
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.