Determination of Limits for a Linear Regression or Calibration Curve

Article

Swallow, William H.; Trout, J. Richard   (1983, ASQC)   North Carolina State University; Rutgers University

Journal of Quality Technology    Vol. 15    No. 3
QICID: 5458    July 1983    pp. 118-125
List $10.00
Member $5.00

FOR A LIMITED TIME, ACCESS TO THIS CONTENT IS FREE!
You will need to be signed in.
New to ASQ? Register here.

Article Abstract

Trace analysis has become increasingly more important because of its connection with recent environmental, chemical, and biological concerns; therefore, measurement techniques often must be evaluated particularly for their ability to measure low concentrations with acceptable accuracy and precision. A methodology is presented here for determining objectively the lower (or upper) limit associated with a linear regression; that is, the point below (or above) which a regression model fails. Methods are also given for determining, provided the data include multiple observations at some x values, whether problems observed beyond the limit are due to increased variability or to breakdown of the linear relationship. The methodology is applied to calibration curve data from our chlorine measurement techniques, to estimate lower limits and to show that the principal problem found below the lower limit is dramatically increased variability.

Keywords

Statistics,Accuracy and precision,Calibration,Heteroscedasticity,Linear regression,Nonlinear models,Precision


Browse QIC Articles Chronologically:     Previous Article     Next Article

New Search

Featured advertisers





ASQ is a global community of people passionate about quality, who use the tools, their ideas and expertise to make our world work better. ASQ: The Global Voice of Quality.