[phth]

Recently I have been having the problem of trying to assess the pH of a phosphate buffer in a mixed aqueous system after adding an acidic species to the buffer. I get a regression that is precise, but it is not accurate.  e.g. the pH should change 0.2 pH and I calculate changes of about 1-1.5 pH. With other buffers like HEPES I get values that are accurate. I made the regression by diluting a concentrate phosphate buffer to the correct cosolvent and 1/10[buffer], then plotting the regression of the conc buffer versus the measured values. My question: is there some way to correct the calculation?