190 Chapter 5: Special Random Variables
0.4
0.3
0.2
0.1
− 3 − 2 −10 1 2 3
t-density with 5 degrees of freedom
Standard normal density
FIGURE 5.15 Comparing standard normal density with the density of T 5.
The mean and variance ofTncan be shown to equal
E[Tn]=0, n> 1
Var(Tn)=
n
n− 2
, n> 2
Thus the variance ofTndecreases to 1 — the variance of a standard normal random
variable — asnincreases to∞. Forα,0<α<1, lettα,nbe such that
P{Tn≥tα,n}=α
It follows from the symmetry about zero of thet-density function that−Tnhas the same
distribution asTn, and so
α=P{−Tn≥tα,n}
=P{Tn≤−tα,n}
= 1 −P{Tn>−tα,n}
Therefore,
P{Tn≥−tα,n}= 1 −α
leading to the conclusion that
−tα,n=t 1 −α,n
which is illustrated in Figure 5.16.
The values oftα,nfor a variety of values ofnandαhave been tabulated in Table A3
in the Appendix. In addition, Programs 5.8.2a and 5.8.2b on the text disk compute the
t-distribution function and the valuestα,n, respectively.