qtukey(p, k, df)
Inf is accepted, though if
k>2, then
df will be set to 1000.
Missing values (
NA) are not accepted.
k
independent standard normal variates based on
df degrees of freedom.
This is the random variable as originally defined by Tukey;
for example,
qtukey(1-alpha, 2, df)
equals
sqrt(2)*qt(1-alpha/2, df).
For many choices of the arguments,
the function does a table lookup.
For large values of
k,
df, and/or unusual values of
p,
the function obtains the critical point by numerical integration
and a bisection method.
Hochberg, Y. and Tamhane, A. C.(1987).
Multiple Comparison Procedures.
Wiley, New York.
Hsu, Jason C. (1996).
Multiple Comparisons: Theory and Methods.
Chapman and Hall, London.
qtukey(.95, 2, Inf)/sqrt(2) qtukey(.99, 50, 100)