Homework 4
Use the
SOCR Neuroimaging dataset of visceral pain, irritable bowel
syndrome, ulcerative colitis, and Crohn's disease.
- Problem 1: Is there evidence of hippocampal volume
asymmetry in the IBS cohort? Formulate an appropriate research hypothesis,
identify a suitable statistical technique, apply the corresponding
software tool, and report your findings.
- Problem 2: Assume the gender distributions of the
participants in this study accurately represents the 4 cohorts
(Normals, IBS, Ulcerative
Colitis and Crohn's disease). Let's say we wish to compare the
proportions of males in the Normals and IBS groups. Use the
Methods of Moments to compute estimates (p1 and p2) of
the probabilities that a randomly selected male is from the
asymptomatic normal control group (p1) and a randomly
selected male is from the IBS control group (p2).
Once you find the point estimates (p1 and p2),
perform a
statistical test to assess the differences between the
proportions of males in the Normals and IBS groups.
Note that you need to invert the conditioning: p'1=P(male|NC), and
p'2=P(male|IBS).
What is the
confidence interval of the difference p'1 - p'2?
- Problem 3: Use the
SOCR Fligner-Killeen test to investigate if the variances of the
white matter volume (WMV) between the 3 cohorts (Normals, IBS, Ulcerative Colitis) are
significantly different. Why is this important? Repeat this
analysis looking for differences in the variability of the gray
matter volume (GMV) in the controls group between genders. Elaborate on your fingings.
- Problem 4: Complete the following table, where patients indicates all study
participant except the asymptomatic controls.
| Groups |
Patients |
Controls |
Total |
| Female |
|
|
|
| Male |
|
|
|
| Total |
|
|
|
Compute a
point estimate and a 95% CI for the Odds Ratio of controls relative to gender.
Next, apply the
χ2 association test to investigate for association between participants'
gender and diagnosis.
