2. Confidence Intervals for the
difference between two population
parameters
In this section, we will use samples from two
populations to create confidence intervals for
the difference between population parameters.
All the examples of this section will involve
independent random samples.
3. Confidence Intervals for the
difference between two population
parameters
There are several types of confidence intervals
for the difference between two population
parameters
Confidence Intervals for 1 – 2 (1 and 2
known)
Confidence Intervals for 1 – 2 (1 and 2 Are
Unknown)
Confidence Intervals for 1 – 2 (1 = 2)
Confidence Intervals for p1 – p2
12. Example Page 381
In his book Secrets of Sleep, Professor Borbely describes research on dreams
in the Sleep Laboratory at the University of Zurich Medical School. During
normal sleep, there is a phase known as REM (rapid eye movement). For most
people, REM sleep occurs about every 90 minutes or so, and it is thought that
dreams occur just before or during the REM phase. Using electronic equipment
in the Sleep Laboratory, it is possible to detect the REM phase in a sleeping
person. If a person is wakened immediately after the REM phase, he or she
usually can describe a dream that has just taken place. Based on a study of
over 650 people in the Zurich Sleep Laboratory, it was found that about one-
third of all dream reports contain feelings of fear, anxiety, or aggression. There
is a conjecture that if a person is in a good mood when going to sleep, the
proportion of “bad” dreams (fear, anxiety, aggression) might be reduced.
Suppose that two groups of subjects were randomly chosen for a sleep study.
In group I, before going to sleep, the subjects spent 1 hour watching a comedy
movie. In this group, there were a total of n1 = 175 dreams recorded, of
which r1 = 49 were dreams with feelings of anxiety, fear, or aggression.
In group II, the subjects did not watch a movie but simply went to sleep. In
this group, there were a total of n2 = 180 dreams recorded, of which r2 = 63
were dreams with feelings of anxiety, fear, or aggression.
13. Example Page 381
a) Check Requirements Why could groups I and II be considered
independent binomial distributions? Why do we have a “large-
sample” situation?
Solution:
Since the two groups were chosen randomly, it is reasonable to
assume that neither group’s responses would be related to the
other’s.
In both groups, each recorded dream could be thought of as a
trial, with success being a dream with feelings of fear, anxiety, or
aggression.
15. Example Page 381
Interpretation What is the meaning of the
confidence interval constructed in part (b)?
Solution:
We are 95% sure that the interval between –
16.6% and 2.6% is one that contains the
percentage difference of “bad” dreams for group
I and group II.