Y [ln

po = -0.322232431088 P2 = -0.342242088547 P4 = -0.453642210148 • 10 91 = 0.588581570495, 93 = 0.103537752850,

P3 = -0.0204231210245, 9o = 0.0993484626060, 92 = 0.531103462366, 94 = 0.38560700634 • 10

If 0.5 < ft < 1 then ) = -z(1 - ft). This approximation produces, for example, the values z(1 - 0.025) « 1.959964 and z(1 - 0.05) w 1.644854. For 10-20 < ft < 1 -10-20, Eq. (3.54) yields an approximation that is accurate to seven decimal places (Odeh and Evans 1974). The percentage point of the standard normal distribution can be used to calculate approximate percentage points of other distributions such as Student's t and chi-squared (see following paragraphs). See the following for more details on the Gaussian distribution: Johnson et al. (1994: Chapter 13 therein) and Patel and Read (1996).

Student's t distribution with v degrees of freedom has following PDF:

f (x) = r((v + 1)/2) (1 + /v)-(v+1)/2 , v = 1, 2, (3.57) (nv)1/2 r(v/2) V !

Approximations have to be used for calculating the percentage point, tv (0). For the Monte Carlo simulation experiments in this book, the following formula (Abramowitz and Stegun 1965: p. 949 therein) is employed:

0 0