Home  |   Schedule   |   Classes    |   Problem Sets   |   e-Quizzes    |   Contact Us

 Biostatistics 140.624
 Statistical Methods in Public Health IV
  Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health

 

Use the TAB key to move from field to field.  Double-check your entries BEFORE you click the "Submit" button.  Once the exam has been submitted, there is no way to retrieve it.  Please do not submit multiple times.  If you do not receive a confirming e-mail within
1 hour, check your SPAM folder. If the confirmation has still not arrived, send an email to mvnatta1@jhu.edu or jtonasci@jhu.edu. Don't forget to double-check your entries BEFORE you click the "Submit" button.

Your name (Last, First)

 

Student ID

 

Re-enter student ID

 

Your SIS e-mail

 

Re-enter e-Mail

 

 
Exam Answer Form
q00

 I have complied with the Academic
 Ethics Statement (type 0 or 1)

q01a

 What were the mean bilirubin values
 (round to two decimal places)

q01b

 What were the median bilirubin values
 (round to two decimal places)

q02

 Which variable had the greater number of outliers
 (type 1 or 2)

q03

 What form of model between mean log
 bilirubin and age is suggested by the
 logbil vs ageyr ksm plot (type 1, 2, or 3)

q04

 What is Pearson's correlation coefficient
 between age and log bilirubin (round
 to two decimal places)

q05

 Within what approximate 95% limits
 can log bilirubin be forecasted (round
 to two decimal places)

   
q06

 What is the estimated change in mean
 log bilirubin per 10 years of age (round
 to two decimal places)

q07

 In the MLR model for log bilirubin
 with ageyr, histologic stage, and sex as
 predictors, did the studentized residuals
 appear normally distributed on the q-q plot
 (type 0 or 1)

q08

 The residuals for the MLR model are
 (type 1 or 2)

q09

 The Shapiro-Wilk test for normality of
 residuals shows (type 1 or 2)

q10

 The case number with the highest leverage
 in the MLR model fit was

   
q11

 The largest studentized residual in the
 MLR was (round to two decimal places)

q12

 How many cases had extreme influence
 (dfits) on MLR fit

q13

 Did the case with the most influence also
 have the most extreme residual
 (type 0 or 1)

q14

 Compared to the original MLR, did R2
 increase when the influential cases were
 excluded (type 0 or 1)

q15

 The variance inflation factors from the
 MLR show that the predictor with the
 highest multiple correlation with the other
 predictors was

   
q16

 Forward and backward variable
 selection procedures yielded
 (type 1 or 2)

q17

 Downweighting influential points
 using resistant regression (rreg)
 gave a P-value for ageyr of
 (round to two decimal places)

q18

 What were the 95% confidence limits on
 ageyr using the bias corrected method
 (round to two decimal places)

 to 
q19

 In the MLR model with age centered
 at the mean age (part o), what is the
 interpretation of the constant term (short answer)

q20

 Slope > age 60
 (round to two decimal places)

   
q21

 Slope ≤ age 60
 (round to two decimal places)

q22

 Difference in slopes, (a)-(b)
 (round to two decimal places)

q23

 What is the P-value for H0:Sex does
 not modify the effect of age on mean
 log bilirubin, controlling for disease stage
 (round to two decimal places)

q24

 In the MLR model Part 15, what is
 the estimated age slope for males
 (round to two decimal places)

q25

 In the MLR model Part 15, what is
 the estimated age slope for females
 (round to two decimal places)

   
q26

 What are the odds of stage 4 disease
 for a female of average age and
 average log bilirubin
 (round to two decimal places)

q27

 By how much is the odds of
 stage 4 disease multiplied for a patient
 whose bilirubin value is 2 log units higher
 (round to two decimal places)

q28

 What is the probability of stage 4
 disease for a female of average age
 and average log bilirubin
 (round to two decimal places)

q29

 What is the relative risk of stage 4
 disease comparing otherwise
 similar females vs males
 (round to two decimal places)

q30

 Does the binary logistic regression
 model of stage 4 disease fit the data
 (type 0 or 1)

q31

 What type of goodness of fit test
 was used (type 1, 2, or 3)

q32

 Does the quadratic term for centered
 log bilirubin significantly (p < 0.05)
 improve the model fit (type 0 or 1)

q33

 Was the goodness of fit improved
 by adding the quadratic term for
 centered log bilirubin (type 0 or 1)

q34

 Does the odds of stage 4 disease for
 females vs males vary significantly
 by age (type 0 or 1)

q35

 What are the odds of stage 4 disease
 for a female of average age and average
 log bilirubin (round to two decimal places)

q36

 What are the relative cumulative
 odds of a higher stage
 (round to two decimal places)

q37

 Is the proportional odds assumption
 significantly rejected by the Brant
 test (type 0 or 1)

q38

 What are the conditional odds of
 stage 2 vs stage 1 disease comparing
 a female vs a male who both have
 average age and average log bilirubin
 (round to two decimal places)

q39

 What is the estimate derived from the
 multinomial model of the conditional
 odds of stage 3 vs stage 4 disease
 comparing a female with a male of
 average age and average log bilirubin
 (round to two decimal places)

q40

 Does the multinomial model fit the data
 (type 0 or 1)

q41

 Which variable has the most siginificant
 relationship to disease stage (type 1, 2, or 3)

q42

 What is the probability under the null
 hypothesis that none of the predictors
 effect the odds of a seizure free 2 week
 period (round to two decimal places)

q43

 What is the upper 95% confidence
 limit for the odds of a seizure free 2 week
 period, comparing active treatment to
 placebo, adjusted for the baseline
 covariates (round to two decimals)

q44

 According to the Hosmer-Lemeshow
 goodness of fit test, does the
 longitudinal binary logistic model
 fit the data (type 0 or 1)

q45

 What method was used to correct for the
 within patient correlation in repeated
 2-week seizure free periods (type 1 or 2)

   
 

  Home   |    Schedule    |    Classes    |    Problem Sets    |    e-Quizzes    |    Contact Us

 Last edited: 07 May, 2019


 © 2019, Department of Biostatistics ,
               Johns Hopkins Bloomberg School of Public Health
               All Rights Reserved