1.      A researcher would like to compare the observed newborn weights (DBWT), measured in pounds, for the following three groups: (1) newborns whose mothers developed gestational diabetes but did not have pre-pregnancy hypertension; (2) newborns whose mothers developed gestational diabetes and had pre-pregnancy hypertension; and (3) newborns whose mothers did not have/develop any of the following – gestational diabetes, gestational hypertension, pre-pregnancy diabetes, and pre-pregnancy hypertension.  Report the means and standard deviations of the newborn weights for these three groups in a single SAS table, and test whether the means of the first two groups, (1) and (2) above, are different from each other (you can assume normality for the purpose of this test).  Finally interpret the results of this test.  [40 points]

SAS code (20 points):

Data newborn_wt;

Set

Keep DBWT RF_GDIAB RF_PHYPE;

Run;

Data htn_diab;

Set newborn_wt;

SAS table (2.5 points):

SAS test output (2.5 points):

Tests Interpretation (15 points):

2.      Gestational diabetes (RF_GDIAB) has long been associated with a higher probability of pre-term labor (birth before the 37th week, for which GESTREC3 can be used).  Discuss the observed proportions for this example. Interpret the test of whether mothers with gestational diabetes are more likely to have a higher chance of delivery before the 37th week compared with those not having gestational diabetes. Also interpret the odds ratio provided by SAS, as well as the relative risk of a pre-term birth.  [25 points]

SAS code (4 points):

SAS table & output (4 points):

Observed Proportions Interpretation For the Context of This Question (5 points):

Tests Interpretation (4 points): 

Odds Ratio Interpretation (4 points): 

Relative Risk Interpretation (4 points):

3.      The following code performs the following three distinct tasks: (1) reads the relevant data; (2) creates new variables; and (3) performs a test for whether, among the subgroup of countries who provide economic incentives (EI), which is a multi-category variable (1=Yes), birth rate in log-terms (BRL) is larger than death rate in log-terms (DRL) across a sample of multiple countries (rows of the dataset).  Both BRL and DRL are assumed to be normally distributed.  Based on this information, fill the blanks with the missing SAS commands and variable/dataset names, and fill the circles with any of the following missing symbols [{*},{.},{+},{ – },{>},{<},{;}].  [20 points – i.e., 0.5 points per item]

a.      __________       ______________ ;

b.      Set ‘S:/ASRU/LES.sas7bdat’ ⃝

c.      Keep ________        ________        ________;

d.      __________ ;

e.      __________      ______________ ;

f.       Set myDB ⃝

g.      EINEW  = ⃝ Economic Incentive=No or N/A ⃝;

h.      __________ = . ;

i.        __________= . ;

j.        If BR~= .    ___________      __________    ⃝   __________(BR);

k.      If DR~= .    ___________      __________     ⃝  __________(DR);

l.        If    ___________   ⃝   1 then  ________     =   ‘Economic Incentive=Yes ⃝;

m.    ___________ ⃝

n.      PROC ___________  _________ = newData  _________ ⃝  L ;

o.      __________        _________       ⃝  ___________ ;

p.      _________       _________  ⃝ ‘Economic Incentive=Yes’;

q.      ___________ ;

4.      A researcher hypothesizes that the proportion of WIC-using mothers (using WIC) who are smokers before pregnancy (using CIG_0) is less than 20%. Write the hypotheses, and perform the appropriate test in SAS using only those reporting a pre-pregnancy smoking status. Provide an interpretation of the results.  [15 points]

Hypotheses (4 points):

SAS Code (5 points):

SAS test output (2 points):

Interpretation (4 points):