use Math::CDF qw(qnorm);

for my $p (0.33, 0.67) {
    print qnorm($p) * 2.5 + 44, "\n";
}