Please enable JavaScript to view this site.

This guide is for an old version of Prism. Browse the latest version or update Prism

The results are presented on multiple pages.

Analysis summary page

This page presents information on the specific test performed (according to the options selected in the parameters dialog. In addition to listing the two columns from the input data table that were analyzed, information including the name of the specific test, assumptions of variance, method for calculating and reporting multiple comparisons, and the number of tests performed/omitted is given.

t tests page

This page shows the t test for each row. The P value for each comparison is shown, along with the multiplicity adjusted P value, which is higher. Additionally, columns for transformations of the P value are included if selected on the Options tab of the parameters dialog. These transformations include -log10(P value) used in volcano plots and -log2(P value) also known as "S values" or "Surprise values".  

If you chose the statistical significance approach, the first column is  labeled "Below threshold?" and provides a "Yes" or "No" to indicate if the comparison resulted in a P value smaller than the specified alpha level, after any adjustments you requested for multiple comparisons. In prior versions of Prism, this column either contained a single asterisk (to indicate that a comparison was declared statistically significant) or not. However, it never showed several asterisks regardless of the corresponding P value. To avoid confusion, this notation was replaced with a simple "Yes" or "No".

If you chose the FDR approach, then the first column is labeled "Discovery?" and contains a simple "Yes" or "No" to indicate if these rows were deemed to be "discoveries".

P small enough and Discoveries page

This page only shows the rows with results which meet the criterion of being "Below threshold" or a "Discovery" based on your choices in the dialog. The rows are sorted by P value, with the smallest P value on top.

What is n?

In most cases, you have complete data on each row and the number of t tests equals the number of rows, so there is no ambiguity.

In some cases, it may not be possible to compute the t test for some rows (because the SD of the replicates is 0.0, or because there is only one value for each group). Prism only counts the rows for which a t test could be computed. If a t test could not be computed for some rows, Prism points this out in a floating note.

 

 

© 1995-2019 GraphPad Software, LLC. All rights reserved.