Our goal in this section is to implement established methods and fit a robust calibration model that reliably returns sample concentration estimates under controlled conditions. At this time we will concentrate on immunometric tests; that is, tests with a positive relationship between concentration and response. Most of the same concepts apply to competitive assays, but the interpretation of some parameters changes with the inverse relationship.

Methods

Methods for calibration curve fitting may differ from ordinary least squares linear regression in three ways:

Try the analyses described in this chapter with the preliminary measurements that accumulate as you develop your system. Do not, however, put too much stock in the results yet. Use this practice to get comfortable with the software and the methods. You will gather clues about the variance structure, curve model and potential problems, which can be tested once you have a large number of replicates from a stable system.