Page 239 - Applied Statistics Using SPSS, STATISTICA, MATLAB and R
P. 239

220      5 Non-Parametric Tests of Hypotheses


               variables describing the previous knowledge of Boole’s Algebra and binary arithmetic
               are independent.

           5.15 Redo Example 5.14 for the variable AB.

           5.16 The  FHR   dataset contains measurements of foetal heart rate baseline performed by
               three human  experts and an  automatic system.  Is there evidence at  the 5%  level of
               significance that there is no difference among the four measurement methods? Is there
               evidence, at 5% level, of no agreement among the human experts?

           5.17 The Culture   dataset contains budget percentages spent on promoting sport activities
               in samples of Portuguese boroughs randomly drawn from three regions. Based on the
               sample evidence is it possible  to conclude that there  are no significant differences
               among those three regions on how the respective boroughs assign budget percentages
               to sport activities? Also perform the budget percentage comparison for pairs of regions.

           5.18 Consider the flow rate data measured at Cávado and Toco Dams included in the Flow
               Rate   dataset. Assume that the December samples are valid random samples for that
               period of the year and, furthermore, assume that one wishes to compare the flow rate
               distributions at the two samples.
               a)  Can the comparison be performed using a parametric test?
               b)  Show that the conclusions of the sign test and of the Wilcoxon signed ranks test
                   are contradictory at 5% level of significance.
               c)  Estimate the power of the Wilcoxon signed ranks test.
               d)  Repeat the previous analyses for the January samples.

           5.19 Using the McNemar Change test compare the pre and post-functional class of patients
               having undergone heart valve implant using the data sample of the Heart V alve
               dataset.

           5.20 Determine which variables are important in the discrimination of carcinoma from other
               tissue types using the  Breast Tissue   dataset,  as well as  in the discrimination
               among all tissue types.

           5.21 Consider the bacterial counts in the spleen contained in the Cells’ dataset and check
               the following statements:
               a)  In general, the CD4 marker is more efficacious than the CD8 marker in the
                   discrimination of the knock-out vs. the control group.
               b)  However, in the first two weeks the CD8 marker is by far the most efficacious in
                   the discrimination of the knock-out vs. the control group.
               c)  Two months after the infection the biochemical markers CD4 and CD8 are unable
                   to discriminate the knock-out from the control group.

           5.22 Based on the sample data included in the Clays’ dataset, compare the holocenic with
               pliocenic clays according to the content of chemical oxides and show that the main
               difference is in  terms of alumina, Al 2 O 3 . Estimate what  is the  needed difference  in
               alumina that will correspond to an approximate power of 90%.
   234   235   236   237   238   239   240   241   242   243   244