a****n
发帖数: 3082 | 1 I am doing small angle scattering on biomolecules
Someone says I need to run as long time as I can on buffer since the error
bar is larger for buffer than the sample
Somebody else says I need to run longer time on Sample than buffer.
Any idea on that and please explain.
Thanks | d*b
发帖数: 21830 | 2 This is a typical counting experiment problem.
Suppose you do N times of counting experiemnts and found n signal events.
When N is larger (typically > 1000 or so) the distribution follows Gaussian
distribution. When N is smaller, it is a Pisson distribution. You can
calculate statistical error based on those distribution and your counting
size very quickly.
Now, problem is, in physics analysis, we do not need n over N, or something
like dn/dx. Because every experiment has different setups. The a
【在 a****n 的大作中提到】
: I am doing small angle scattering on biomolecules
: Someone says I need to run as long time as I can on buffer since the error
: bar is larger for buffer than the sample
: Somebody else says I need to run longer time on Sample than buffer.
: Any idea on that and please explain.
: Thanks
| a****n
发帖数: 3082 | 3 i don't have a solid background on statistical physics. thank you.
Gaussian
something
need
cares
【在 d*b 的大作中提到】
: This is a typical counting experiment problem.
: Suppose you do N times of counting experiemnts and found n signal events.
: When N is larger (typically > 1000 or so) the distribution follows Gaussian
: distribution. When N is smaller, it is a Pisson distribution. You can
: calculate statistical error based on those distribution and your counting
: size very quickly.
: Now, problem is, in physics analysis, we do not need n over N, or something
: like dn/dx. Because every experiment has different setups. The a
|
|
