Authors :
Taiwo, Abraham Soji; Adejumo Adebowale Olusola; Sulaimon, Ibikunle Idris
Volume/Issue :
Volume 6 - 2021, Issue 7 - July
Google Scholar :
http://bitly.ws/9nMw
Scribd :
https://bit.ly/3CD7i4i
Abstract :
When performing a test of conditional
independence or Homogeneity of associations, one of the
tests that come to mind is the Cochran or Mantel
Haenszel test, furthermore, the Breslow day test. The
attributes of these tests are notable and it is being
recommended that log linear models can be utilized to
test similar theories. This research paper takes a look at
the two tests in the framework of loglinear modeling.
Expanding on the knowledge of both tests by delineating
how theories of homogeneous associations and
conditional independence can be applied along with
extra speculations and introducing a speculation of the
ideas of homogeneity of associations and conditional
independence in higher request connections in the
context of loglinear modeling.
Methods were illustrated using real life data on HIV
/AIDS. The data which was classified into a
225
partial table consisted of the morbidity pattern and
mortality for both genders from 2010-2014. It was
revealed that the odds ratio of gender by survival status
for each of the year was heterogeneous. It was also
discovered that the survival status for HIV/AIDS
depended on gender conditionally while controlling for
the effect of the years. Finally, we observed that the
conditional odds ratio for male deaths is 1.612 based on
the available data
Keywords :
Conditional Independence, Homogeneity of Associations, Loglinear Modeling, Cochran or Mantel Haenszel Test, Breslow Day Test
When performing a test of conditional
independence or Homogeneity of associations, one of the
tests that come to mind is the Cochran or Mantel
Haenszel test, furthermore, the Breslow day test. The
attributes of these tests are notable and it is being
recommended that log linear models can be utilized to
test similar theories. This research paper takes a look at
the two tests in the framework of loglinear modeling.
Expanding on the knowledge of both tests by delineating
how theories of homogeneous associations and
conditional independence can be applied along with
extra speculations and introducing a speculation of the
ideas of homogeneity of associations and conditional
independence in higher request connections in the
context of loglinear modeling.
Methods were illustrated using real life data on HIV
/AIDS. The data which was classified into a
225
partial table consisted of the morbidity pattern and
mortality for both genders from 2010-2014. It was
revealed that the odds ratio of gender by survival status
for each of the year was heterogeneous. It was also
discovered that the survival status for HIV/AIDS
depended on gender conditionally while controlling for
the effect of the years. Finally, we observed that the
conditional odds ratio for male deaths is 1.612 based on
the available data
Keywords :
Conditional Independence, Homogeneity of Associations, Loglinear Modeling, Cochran or Mantel Haenszel Test, Breslow Day Test
