invnor(P [,upper:T or lower:F]), P REAL with elements of P between 0 and 1, df > 0 |

invnor(P) computes the quantiles (probability points, critical values) of the normal distribution corresponding to each element of P. P must be a REAL vector or array with elements between 0 and 1. The result has the same size and shape as P. invnor(P,upper:T) and invnor(P,lower:F) compute upper tail quantiles of the standard normal distribution. The result is mathematically equivalent to invnor(1 - P) but may be more accurate for small P. A critical value for a two-tail Z-test with significance level alpha or for a 1-alpha confidence interval may be computed as invnor(alpha/2, upper:T) or invnor(1-alpha/2). Critical values for a one-tail Z-test with significance level alpha are computed as invnor(alpha) (lower tail test) and invnor(alpha,upper:T) or invnor(1-alpha) (upper tail test). invnor() is the inverse of cumnor() in the sense that invnor(cumnor(z)) should be the same as z within rounding error and cumnor(invnor(P)) should be the same as P within rounding error. See also cumnor(), rnorm()

Gary Oehlert 2003-01-15