Tip:
Highlight text to annotate it
X
This question compared the effect of calcipotral cream versus ultra based
cream on the reduction of skin lesions. We had p values of .017 for the
calcipotral cream versus p value of .056 for the ultrabase base.
The answer to the first question is D, the difference between the creams may not
be significantly different even though the p value for one is less than .05 and
for the other is greater than .05. It appears that there was some reduction
on the placebo side. That would just missed statistical
significance because we know that the p value is slightly greater than .05.
it's unlikely that the between group difference will be significant here.
The next question asked us the unit of observation for the study.
Because we were studying the side of a face, that is the unit of observation and
that's the answer to this question. This study does involve correlated
observations because we are studying two sides of the same person's face and these
obviously will be correlated. This question has the authors running a
lot of tests, so they have 15 different levels of black tea, coffee, decaf
coffee. They have two diseases [UNKNOWN].
They also have two types of participants, pre menopausal versus postmenopausal.
In such a large number of test that are run, we would expect to find a few p
values that are less than .05 just by chance.
The fact that we have a couple of values that are less than .05 does not reflect
actual correlation in this case. Because this study examined individual
women, the unit of observation here is a person.
The observations in this study are not correlated.
This is because this study examined independent women.
This question gives us the increased risk for various diseases comparing two
groups, exposed versus unexposed. we can see that the, the risk ratios in
case B and D are very close to one, which tells us that these are not really very
clinically significant. And thus the answer to this question is
A, C and E where we actually do see a significant departure from 1.
For the risk ratios to be statistically significant at a .05 level.
we should have the confidence intervals such that one does not cross them.
Which means that one should either be lower than the lower bound or one should
be greater than the upper bound. the only two cases for which this is
true, is A and B. Thus these are the only risk ratios that
are statistically significant at a .05 level.