pivot statistic

We define [Graphics:../Images/index_gr_1432.gif], and the corresponding z-value [Graphics:../Images/index_gr_1433.gif] We denote this z-value as z8[v]=zc[v,{0,0,0,0},0,1].

[Graphics:../Images/index_gr_1434.gif]
[Graphics:../Images/index_gr_1435.gif]
[Graphics:../Images/index_gr_1436.gif]
[Graphics:../Images/index_gr_1437.gif]
[Graphics:../Images/index_gr_1438.gif]
[Graphics:../Images/index_gr_1439.gif]

We slightly alter z8[v] and denoted zq[v]=[Graphics:../Images/index_gr_1440.gif] below.

[Graphics:../Images/index_gr_1441.gif]

Here we define a function to collect [Graphics:../Images/index_gr_1442.gif]'s for later use

[Graphics:../Images/index_gr_1443.gif]

check if this is correct.

[Graphics:../Images/index_gr_1444.gif]
[Graphics:../Images/index_gr_1445.gif]

Now we continue to calculate the distribution function of zq[v]

[Graphics:../Images/index_gr_1446.gif]
[Graphics:../Images/index_gr_1447.gif]
[Graphics:../Images/index_gr_1448.gif]
[Graphics:../Images/index_gr_1449.gif]

The distribution function of zq is obtained here. [Graphics:../Images/index_gr_1450.gif]=Φ{zfzq[w,{q0,q1,q2,q3},v0,tau]}.

[Graphics:../Images/index_gr_1451.gif]
[Graphics:../Images/index_gr_1452.gif]

For tau=1, zfzq becomes

[Graphics:../Images/index_gr_1453.gif]
[Graphics:../Images/index_gr_1454.gif]

When v0=0, tau=1,  zfzq becomes

[Graphics:../Images/index_gr_1455.gif]
[Graphics:../Images/index_gr_1456.gif]

In particular, the distribution function of z8 under v0=0 is Pr{Z8<=w;0,1)=Φ(w).

[Graphics:../Images/index_gr_1457.gif]
[Graphics:../Images/index_gr_1458.gif]


Converted by Mathematica      July 21, 2003