If you enter data onto XY or two-way tables with replicate Y values in subcolumns, and want to view a table of mean and SD (or SEM) values, click Analyze and choose to do a built-in analysis. Then choose Row means/totals, and choose one total/mean for each data set.
Note that you rarely need this analysis. Prism will automatically create graphs with the mean and SD (or SEM). You don't have to choose any analyses to make these graphs. Prism computes the error bars automatically. Use settings on the Format Graph dialog (double-click on any symbol to see it) to plot individual points or to choose SD, SEM, 95%CI or range error bars. The only purpose of this analysis is if you want to see the SD or SEM values.
This choice is used rarely, but it helps you consolidate a larger table into a single data set.
If the data were entered onto a single subcolumn for each data set, then there is no ambiguity.
But what if you entered data on a table with subcolumns? In this case, the calculated total/mean values are based on the mean value of each data set for corresponding row. For example, let's calculate "Row means with SD" for three datasets with 3, 2 and 3 replicates in each. Here are the data for the first row.
Data set A: (2, 3, 4)
Data set B: (4, 6)
Data set C: (7, 8, 9)
Prism will first compute the mean values for each data set, which are 3, 5 and 8. It then computes the grand mean of those three values (and their standard deviation) so the results are Mean = 5.333, SD = 2.517, N = 3.
If Prism simply looked at those data as eight independent values, the mean would be 5.375, but since the values in different data sets are unlikely to be independent, Prism does not do this calculation.
•When Prism computes the grand mean, it does not account for the fact that data sets A and C are in triplicate, while data set B is only in duplicate.
•If you entered your data as mean, n and SD or SEM, these calculations use only the mean value you entered and ignore n and SD or SEM.