L1 Craig: INTRODUCTION TO THE STATISTICAL ANALYSIS OF EXPERIMENTAL DATA Flashcards
what should uncertainty be estimated and quoted as?
error bar
what does uncertainty determine?
precision of the measured result
If you now the true value what can you do ?
measure ether accuracy
what is absolute uncertainty?
actual magnitude of the uncertainty
what is relative uncertainty ?
expressed as a fraction or percentage of the measured value e.g. Dx/x or 100 x Dx/x %
should the uncertainty be quoted to the same level of precision as the measurement itself ?
yes
what does it mean if no uncertainty is quoted?
last figure in the measurement is the best that can be determined
so uncertainty is half of this
what are the 2 categories of uncertainties ?
random and systematic
what’s random uncertainty?
produces a scatter of values due to random effects such as poor resolution, electrical noise, human failings; for single measurement it must be estimated; for repeat measurements it may be calculated.
what is systematic uncertainty ?
arises through incorrect calibration or mistake in experimental design often resulting in an off-set value (bias); once identified it is often possible to correct for such an error.
why do we repeat measurements?
- Identify possible outliers–
- Calculate a mean
- Estimate the uncertainty
what are outliers?
values that appear way out of line and may be due to some one-off, gross error.
When only random errors are present the best-value is…..
average or mean
In practice often only a few (n < 12) repeat measurements are available. In this case the______ is a better estimate?
range method
what is standard error?
property of a set of measurements
