This could as well go to the statistics, but I am looking at it from particle physics point of view...(adsbygoogle = window.adsbygoogle || []).push({});

Why adding systematic uncertainties worsen the expected upper limits to the signal strength?

I am trying to find where the flaw enters in the following logic:

0. The model most analyses use is the following likelihood:

[itex]L( N_{obs} | b(\theta ) + \mu s(\theta ) ) = P(N_{obs} |b(\theta ) + \mu s(\theta ) ) U(\mu) \Pi_i Gaus(\theta_i | 0,1)[/itex]

Where [itex]N_{obs}[/itex] is the observed events, [itex]b/s[/itex] are the background/signal expected events, [itex]\theta_i[/itex] are the different nuisance parameters and [itex]\mu[/itex] is called the signal strength. In a Bayesian approach, one has to also to feed in a prior distribution for the signal strength parameter, which is the [itex]U(\mu)[/itex]- let's consider itUniform. [itex]P(x|n)[/itex] is the poisson probability to get x observed given the expectation of n, and [itex]Gaus[/itex] is a way to represent the variation of the nuisance parameters (given you have symmetric errors).

1. In order for one to get the expected limits, they would set [itex]N_{obs}=N_{exp}=b[/itex].

2. Once they do it, they can start varying the background+signal uncertainties, [itex]\theta_{stat}[/itex] (+[itex]\theta_{sys}[/itex]) [these uncertainties don't affect the signal and background in the same way]

3. On the varied result, they would try to figure out what is the [itex]\mu[/itex] so that the [itex]b'+\mu s' = N_{obs}[/itex]

4. Doing that several times, you get a distribution for [itex]\mu[/itex] (after you marginalize over the uncertainties) which is called the posterior pdf....

5. From μ-distribution get the 95-quantile point.

Now for some reason, adding nuisance parameters (such as [itex]\theta_{sys}[/itex] on top of the statistical), moves the [itex]\mu_{.95}[/itex] higher.

Is that because the uncertainties are not the same for bkg/signal?

Intuitively I can see how eg by subtracting the background from the observed result with higher uncertainties is going to give a more unclear picture of how much signal you allow in the game... but I don't see where this fits in the above logic.

**Physics Forums - The Fusion of Science and Community**

Join Physics Forums Today!

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

# A Help with the statistics of Upper Limits?

Have something to add?

Draft saved
Draft deleted

Loading...

Similar Threads - Help statistics Upper | Date |
---|---|

A I need a spherically symmetric spin-dependent NN potential | Mar 10, 2018 |

I Why do particle physicists use Gaussian error estimates? | Jan 29, 2018 |

I Best way to get an overview of research happening at CERN | Sep 24, 2017 |

A Walecka mean field theory, Help to find its M* | May 1, 2017 |

**Physics Forums - The Fusion of Science and Community**