When fitting a family of curves with nonlinear regression, I want to constrain a parameter to a different constant value for each data set.

The Constrain tab of Prism's nonlinear regression dialog lets you define a parameter to be a constant value, but one value for all data sets. What if you want to set different constant values for different data sets?  Here are three ways to work around that limitation:

• Run the curve fit several times, once for each data set. With each analysis, choose which data set to include (in the "Analyze Data", set "Data to analyze").
• Enter the constant value into the column titles of your data table, and then assign the parameter to be a column constant. 
• Write a user defined equation, and use special notation to assign a different value for each data set. You'll need to hard code the constants within the user-defined equation. For example:
  <A>Bottom=4.5
  <B>Bottom=34.5
  <C>Bottom=45.6
  Y=Bottom + span*(1-exp(-1*K*X))