5 ms·
As an undergrad at Georgia Tech you should know that if one of my professors caught me writing some significant digit nonsense like 1.0272303 ± 1.031768 Bq/kg I
by carlob 14y ago
As an undergrad at Georgia Tech you should know that if one of my professors caught me writing some significant digit nonsense like 1.0272303 ± 1.031768 Bq/kg I would not have gotten a degree.
1.0272303 ± 1.031768 Bq/kg or rather 1.0 ± 1.0 Bq/kg means that your experimental error is so large that your result is compatible with both 0 and 2 Bq/Kg. What's the point of adding all those significant digits if you're telling us that your result is imprecise, or rather that your experiment can't effectively tell if there is any activity from Cs-60.
(edit: I wasn't an undergrad at GT)
- drharris 14y agoI thought this exact thing. When I was at GT, this would have resulted in a grade of 50 +/- 50. Once you get to that error, you need to redesign the experiment.
- blueintegral 14y agoThat data came from the HPGe, which really does have that precision. I know it seems like the accuracy is bad since, like you said, it could be either 0 or 2 Bq/kg. However, a Bq/kg is REALLY tiny and that accuracy (which is proportional to the length of the experiment) is good enough for this analysis. So the precision is fixed based on the HPGe and the accuracy increases with the length of time the HPGe took measurements.
- carlob 14y agoMy point stands: just write 1.0±1.0. If 1 Bq/Kg is small then giving a precision of 0.1 nBq/Kg is just plain insane.
- blueintegral 14y agoYou're right, I'll correct it.
- carlob 14y agoThat's great! I see you're not a physics major, so I apologize if I came out as aggressive. In general you want to keep 1 or 2 digits in the s.d. and align the precision of the mean accordingly.