After discussing with Sarah and based on need for SOE, I decided that the Standard Error calculation should be SE = sqrt(S^2) rather than SE = sqrt(S^2 / N). Issue is really a sloppy definition of Standard Error verses Standard Deviation. However, everything I've found on line seems to confirm that this is the calculation that is used.
After discussing with Sarah and based on need for SOE, I decided that the Standard Error calculation should be
SE = sqrt(S^2)
rather thanSE = sqrt(S^2 / N)
. Issue is really a sloppy definition of Standard Error verses Standard Deviation. However, everything I've found on line seems to confirm that this is the calculation that is used.