Q & A: Whole Number Standard Deviations

Posted by Sten Westgard, MS

This question comes to us from a CLS student in Texas:

"I was hoping that someone might be able to answer a question that is causing me and some of my
classmates some confusion. There is some confusion when you are plotting your QC chart and all your values are a whole number. Would you keep you SD as a whole number or make it to one decimal place? And if you did make it to one decimal place, would you be making your SD more accurate that your original value?"

Answer after the jump...

Dr. Westgard replies:

There can be difficulties due to data rounding.  In general, it is good practice to maintain one extra significant figure for QC purposes - for the test results, the mean, and the SD.  The problem is that some instrument systems round all test results and that may compromise the QC data and its interpretation.  For example, if BUN is rounding to a whole number and the SD of the method is really 0.6 mg/dL, 2 SD limits would be ± 1.2, but would actually get set as ± 1, which will cause the false rejection rate to be higher and an even bigger problem that expected.  If the SD were rounded to 1, then 2SD limits would be ± 2 units, which actually correspond to ± 3.3 SD.  If rounded to 1, 3 SD limits would be set as ± 3 units, which really corresponds to 5 SD - so wide there would be little error detection available.  So the best practice is to  maintain one more significant figure all the way through until the test result is finally rounded for the report.

We welcome your This email address is being protected from spambots. You need JavaScript enabled to view it..