Question I

A study of patients with cancer of the prostate was undertaken by Dr Don Young at the Ohio State University Comprehensive Cancer Center (see Hosmer & Lemeshor, 2000). A multiple logistic regression model was performed using data from a sample of 380 patients to determine whether variables at the baseline exam can be used to predict whether the tumor has penetrated the prostatic capsule to predict the probability of tumor penetration of prostatic capsule (ProstateData.sav).

1. Use the multiple logistic regression model to summarize the relationship between the probability of patient tumor penetration of prostatic capsule and the various potential predictors. (5 points)
2. Which ones of the variables are important predictors of the probability of patient tumor penetration of prostatic capsule? Explain why. (5 points)
3. Does the model fit well the data at hand? Explain why. (5 points)
4. Given a patient with: AGE = 45, WHITE, DPROS=No, DCAPS = No, PSA = 16, VOL = 16, and GLEASON = 0, estimate the probability of patient tumor penetration of prostatic capsule (Note: In the regression model, for categorical variables, the reference category is given by the last category in ascending order; e.g. RACE:2 = BLACK is the reference). (10 points)

Price: $11.44

Solution: The downloadable solution consists of 6 pages, 544 words and 4 charts.
Deliverable: Word Document