When reading an old paper from 1921* I find formulas like:
$\rho + \frac{\rho(1- \rho^2)}{2\overline{n - 1}} \big( 1+ \frac{9 - 14\rho^2}{6\overline{n-1}} \big)$
which is said to be the median of the estimated sample correlation, $r$.
What is the meaning of the horizontal bar above (n - 1)?
My best guess is that it is an old alternative notation for paranthesis but I am not able to find any real justification for this assumption. The paper also include expressions such as:
$2 \overline{s-1}r = \overline{s-2}+s \tanh(z - \phi)$ where the need of paranthesis around $s-1$ can be questioned.
If there where complex numbers involved I guess it could also be the complex conjugate but that does not really make sense here.
*R A Fisher "On the probable error of a coefficient of correlation deduced from a small samlpe", Metron 1921,
The horizontal bar is called a vinculum and is an alternative to brackets. See Wikipedia and Jeff Miller.