A pediatric hospital is considering changes to the timing of
doctor rounds and they are worried that these changes will
negatively affect the sleep patterns of babies in their care.
They decide to implement the changes for a sample of 6 babies and
compare their total sleep to known population parameters.
Mean of HO pop.
µ = 16 hours of sleep
Standard deviation of pop.
? = 2.7
Mean of H1 pop.
µ = 14.9 hours of sleep
Sample size
N = 6 infants
Standard Error
?M = 1.1
Calculate statistical power for “the standard”; a two-tailed
test (? = .05).
Since the research hypothesis is that those in the sample will
sleep fewer hours, on average, than those in the population, one
could make the case for a one-tailed test. Recalculate statistical
power for a one-tailed test (? = .05).
How did switching from a two-tailed to a one-tailed test impact
power? Is this generally an acceptable strategy? Why or why
not?
Recalculate statistical power for a one-tailed test (? =
.10)
How did increasing alpha from .05 to .10 affect power? Is this
generally an acceptable strategy? Why or why not?