Understanding the Calibration Curve

What is wrong with this calibration curve?

What is wrong with this curve?

The linear fit "looks good" and the correlation coefficient is "three nines." This must be a good curve, right? Unfortunately, if you reanalyzed the standards as samples, you would find that five of the eight standards were in error by more than five percent. Three of the standards are in error by more than 15% from their known values! Clearly, the usual procedures for evaluating calibration curves are not always accurate.

There is more to understanding calibration curves than just running standards and measuring the correlation coefficient! The calibration curve is one of the most important steps in the generation of analytical data. We will show you how to do it the right way (and it isn't difficult)!

When properly designed, the calibration system provides accurate results with a minimum of effort. However, a poorly constructed system can waste valuable instrument and analyst time, and result in poor quality data.

This seminar will provide a complete set of instructions on how to set up and evaluate any analytical calibration system, with an emphasis on providing the highest quality results with a minimum of effort. Examples are drawn from GC, LC, GC/MS, and AA data sets. Spreadsheet programming examples are included in the course notes.

This course provides valuable and practical information that is not currently available from any other single reference source.

Students will learn:

  • The fundamentals of linear regression
  • Different options for preparing data
  • What calibration options are available
  • Why the correlation coefficient is not a good measure of calibration quality
  • Simple procedures and tests for evaluating calibration data


  1. Linear Regression Basics
  2. Data Transformations
    • External Standard Method
    • Internal Standard Method
    • Isotope Dilution
  3. Selecting Calibration Levels
    • How many levels are needed and how should they be spaced?
  4. Calibration Options
    • Single Point
    • Response Factor/Average Response Factor
    • Linear Through Zero
    • Linear With Intercept
    • Method of Standard Additions
    • 2nd Order Polynomial
  5. Evaluation of Data
    • Correlation Coefficient
    • Analysis of Residuals
    • Zero Intercept Test
  6. Evaluating Real World Data Sets

Course Length: 1/2 day

A condensed version of this topic is available as an on-demand webinar.
Contact us for information.