Normalizing metrics
We previously saw that there was not a significant difference between the brain volumes of elderly individuals with and without Alzheimer's Disease.
But could a correlated measure, such as "skull volume" be masking the differences?
For this exercise, calculate a new test statistic for the comparison of brain volume between groups, after adjusting for the subject's skull size.
Using results.statistic and results.pvalue as your guide, answer the question: Is there strong evidence that Alzheimer's Disease is marked by smaller brain size, relative to skull size?
Diese Übung ist Teil des Kurses
<Kurs>Biomedical Image Analysis in Python</Kurs>Interaktive praktische Übung
Versuche dich an dieser Übung, indem du diesen Beispielcode vervollständigst.
# Import independent two-sample t-test
____
# Divide `df.brain_vol` by `df.skull_vol`
df['adj_brain_vol'] = ____
# Select brain measures by Alzheimers group
brain_alz = df.loc[df.alzheimers == ____, 'adj_brain_vol']
brain_typ = df.loc[____, 'adj_brain_vol']
# Evaluate null hypothesis
results = ____