> I thought the appropriate calculation for SEM was to divide
> the standard deviation (across subjects) by the square root
> of (the number of subjects minus 1), or is that not true? Am
> I mixing up sample stdev and population stdev?
There is some subtle difference here. The sample error of the mean (SEM) is the standard deviation of the sampling distribution of the mean, which measures the accuracy of the sample mean as the estimate for the population mean. According to central limit theorem, it can be calculated as the standard deviation of population distribution divided by the square root of the sample size.
The standard deviation of population distribution is usually unkown, and is estimated by the standard deviation of the sample, which is the square root of (the sum of the squared deviations around the sample mean divided by the sample size minus one).
Therefore the n-1 part occurs in the calculation of the standard deviation of the sample, not directly in the SEM.
Gang