3.2.5. Preferences
Various parameters can be set in the worksheet
Max deviation for bad average
A percentage is entered in this field. After pressing the "Calc average" button, an average is calculated from all calculated results. If the standard deviation of these measurement points divided by the average is >15%, the result will be coloured red. All averages with a CV% ≤15% turn green.
Extrapolate (0=no, 1=yes, 2=between blank and highest dose)
The calculation range of a calibration line is between the lowest and highest dose / concentration.
The calibration line can be extrapolated above the highest dose and below the lowest dose but often results obtained from these extrapolated responses are not validated.
In special cases option 1 can be chosen to use results from extrapolation. The area between the blank and the lowest dose / concentration is below the determination limit (Lower Limit of Detection (LLOD)) of the test. In cases when results below the LLOD must be calculated option 2 can be chosen.
Matrix? (0 = no, 1 = yes)
Option 1 displays the results in a matrix on the "Logit" sheet. This result matrix with results is always shown in the "Datasheet".
**Linear between blank and lowest dose (0 = No, 1 = yes, 2 = spline) **
The measurement range between the lowest dose / concentration and the blank is linear in some cases.
By default, between blank and lowest measurement point results are not calculated, but they are calculated if option 2 is chosen for extrapolation.
In this the range between blank and lowest dose the calibration line is curved due to the logit regression.
If it is proven that the measurement area between blank and lowest measurement point can be better calculated as a straight line, option 1 can be switched on.
The calculated results, with responses between blank and lowest dose, are underlined with the results. This option has been added for a test in which many greatly reduced samples are measured and it has been proven that the area in this measurement area behaves as a straight line.
Keep in mind this option was added for special cases.
With option 2 set to 'spline', straight line segments connect the average responses for each measuring point, and the concentration is computed along these segments. Subsequently, the logit regression is deactivated.
Significant digits The number of significant digits in a result is a representation of the reliability of the result. The result 10.1 suggests a reliability of 0.5 - 1% in the result. The result 10, on the other hand, suggests a reliability of 5 - 10%. It can be 9.51 but also 10.44. The same applies to 101 and 100, and 1010 and 1000.
Using three significant digits is advisable for ELISAs and RIAs when test reproducibility is approximately 10%. Numbers are rounded to values such as 1.23, 9.98, 10.1, or 19.9, allowing for minor multiplication or division without significant information loss. With two significant digits, the results would be 1.2, 10, 10 and 20. This is fine when reporting the results to a customer.
> or <result displayed (0 = no, 1 = yes). If a result falls outside the measuring range, this is shown as> the upper limit or and < characters can be undesirable if the results are to be further calculated. With the 0 option, the> and <signs are not shown.