How does Prism do linear and nonlinear regression with data entered as mean, SD and N (or mean, SEM and N)?

Prism lets you enter XY data as mean, SD (or SEM) and N.

Prism 5 and later

• If you analyze these data with ordinary nonlinear regression, Prism gives exactly the same results as if you had entered raw data, even with weighted regression. This document explains the mathematical details about how Prism does it.
• If you choose robust nonlinear regression, Prism fits only the means and ignores the n and SD or SEM values you entered.  
• If you choose outlier detection, this uses only the information from the means (ignoring n and SD or SEM). But once any outliers are identified and removed (if you chose that), the fitting proceeds as usual and does account for n and SD or SEM. 
• If you analyze the data with linear regression, Prism 6 and earlier only analyzed the means, and ignored the n and SD  or SEM values.  If you use nonlinear regression, you can choose the straight line equation, and Prism will account for the SD (or SEM) and N values you entered. Prism 7 and later accounts for the n and SD or SEM, so the linear regression results (like nonlinear regression results) are identical for analyses of raw data vs. data entered as mean, n and SD or SEM. 
• If you want Prism to only fit the means with nonlinear regression, this is a choice on the Weights tab. 

Prism 4

If you entered data as mean, SD (or SEM) and N, Prism 4  (by default) fits the means and weights by the sample size (N). This is one of the two options on the weighting tab (the other option is to fit means only, ignoring N).  Prism 4 only accounts for differences in N, but not SD (or SEM). The best fit values of the parameters, and thus the appearance of the curve, is the same with Prism 4 and 5. But Prism 5 does a smarter job with standard errors, confidence intervals, and comparisons of models.