8.2. Population Mean: Small Sample Case
t-distribution
(“student” t distribution):
This distribution was
invented by W. S. Gossett (published in the name “Student”).
The t-distribution is a family of similar probability distributions,
with a specific t distribution
depending on a parameter known the
degree of freedom (d.f.). Denote
to be the
random variable having t-distribution with n degree of freedom.
Note:
. The t-distribution is symmetric about 0.
: a t value with an area of
in the upper tail
of the t-distribution with degree of
freedom equal to n.
.
Important Result:
When
the population has a normal probability distribution and
is
unknown, then
,
where
,
, and
are random
variables with associated possible values
.
Derivation of
confidence
interval:
Suppose
the population has a normal probability distribution. Since

thus

confidence interval based on t-distribution:
As
the population has a normal distribution and
is unknown,
![]()
is a
confidence
interval of
.
Example 2:
Consider the following
random sample of 4 observations, 25, 47, 32, and 56. Suppose
the population is normally
distributed. Please construct a 95% confidence interval for
.
[solution:]
![]()
In
addition,
and
. Thus,

JavaStatSoft:
Confidence
interval:
Statistics
-> Estimation -> One Sample -> Mean -> T Interval