Detailed Notes on Residuals and Calibration Methods
Residuals
Definition of Residuals
- A residual is the difference between an observed data point and the value predicted by a line of best fit.
- Residuals are crucial for calculating the standard error of the slope and intercept derived from the line of best fit.
- They are also used to determine the standard error of concentration as calculated from a calibration graph.
Calculating Residuals
- Collect Data: Gather response data (y) against analyte concentration data (x).
- Fit Line: Determine the line of best fit parameters:
- Slope (m): 4.9286
- y-intercept (c): 2
- Coefficient of determination (R²): 0.9923
- Compute the Fitted Line: The line of best fit can be expressed as:
y_{calc} = 4.9286x + 2 - Calculate Residuals: For each data point, compute the residual by:
Residual = y - y_{calc}
Using Residuals
- When a straight line is fitted, the residuals should be randomly scattered around 0, indicating a good fit.
- It is important to note that for calibration graphs, the magnitude of errors concerning y is rarely constant. This means that the error in y may increase as the analyte concentration (x) increases.
- Analyzing residuals can help identify non-linear relationships in the data and qualitatively detect outliers from a residual plot.
- Residuals are foundational in calculating the standard error of the predictor values, denoted as s_{y/x} and essential for determining the error of the slope and intercept.
Standard Error
- The standard error for the y estimate is denoted by:
s_{y/x} - The variable n represents the number of points on the calibration graph.
Error in Slope and Intercept
- The errors associated with the slope and intercept are typically presented in examinations as follows:
- s = standard error
- n_S = number of sample repeat measurements
- n_C = number of points on the calibration graph
- m = slope of the line of best fit from the calibration graph
- The subscripts used refer to each individual data point (i) or the unknown data point (0), while the over bar indicates the mean value for x and y.
Standard Additions
Overview
- An alternative calibration technique to standard series measurement is termed Standard Additions.
- In this method, the sample is spiked with a standard solution of the analyte being measured.
- This approach is particularly useful in overcoming matrix effects, which occur when components in the sample either suppress or elevate the instrumental signal generated by the analyte.