312 Chapter 8:Hypothesis Testing
8.4Testing the Equality of Means of Two Normal Populations
A common situation faced by a practicing engineer is one in which she must decide whether
two different approaches lead to the same solution. Often such a situation can be modeled
as a test of the hypothesis that two normal populations have the same mean value.
8.4.1 Case of Known Variances
Suppose thatX 1 ,...,XnandY 1 ,...,Ymare independent samples from normal populations
having unknown meansμxandμybut known variancesσx^2 andσy^2. Let us consider the
problem of testing the hypothesis
H 0 :μx=μy
versus the alternative
H 1 :μx=μy
SinceXis an estimate ofμxandYofμy, it follows thatX−Ycan be used to estimate
μx−μy. Hence, because the null hypothesis can be written asH 0 :μx−μy=0, it seems
reasonable to reject it whenX−Yis far from zero. That is, the form of the test should
be to
reject H 0 if |X−Y|>c
accept H 0 if |X−Y|≤c
(8.4.1)
for some suitably chosen valuec.
To determine that value ofcthat would result in the test in Equations 8.4.1 having
a significance levelα, we need determine the distribution ofX−Y whenH 0 is true.
However, as was shown in Section 7.3.2,
X−Y∼N
(
μx−μy,
σx^2
n
+
σy^2
m
)
which implies that
X−Y−(μx−μy)
√
σx^2
n
+
σy^2
m
∼N(0, 1) (8.4.2)
Hence, whenH 0 is true (and soμx−μy=0), it follows that
(X−Y)
/√
σx^2
n
+
σy^2
m