TU Wien:Statistik und Wahrscheinlichkeitstheorie UE (Levajkovic)/Übungen 2022W/HW09.1
- Distribution of the -value
In the context of a left-sided statistical test, let be the test statistic which under has a continuous and strictly monotone distribution function . Then the -value writes as
Note that this is the evaluation of the data summarized in the statistic . THe question arises of 'how the random p-value' is distributed, i.e., if the random test statistic is evaluated. In other words, what the distribution of under ?
Note that this distribution is the same for right-sided tests, and also for two-sided tests if is symmetric around zero, i.e., for all .
{{Beispiel|1= Angabetext }}
oder
{{Beispiel| Angabetext }}
zu (im Falle einer korrekten, unverifizierten Lösung "solved". Auch möglich "unsolved", "wrong", "verified_by_tutor". Alle möglichen Werte sind hier: Vorlage:Beispiel dokumentiert.)
{{Beispiel|status=solved|1= Angabetext }}
Lösung von Tutorin[Bearbeiten | Quelltext bearbeiten]
Left-sided test:[Bearbeiten | Quelltext bearbeiten]
We test against . Let be the test statistic, whose distribution is given via the cdf (continuous, invertible, differentiable). Let be the value of the -statistics computed from the data set. Then, the -value in this case is
For different samples we obtain different values . Thus, the distribution of the -value (as random variable) is uniform, i.e., , because of we find (using that is invertible)
Also, for the probability is as well as for the probability is . Alltogether, these lead to the cdf of the uniform distribution on (0,1).
For completeness:
right-sided:
two-sided:
Dieses Beispiel hat noch keinen Lösungsvorschlag. Um einen zu erstellen, kopiere folgende Zeilen, bearbeite die Seite und aktualisiere den status=unsolved
Mögliche status=...
Werte stehen hier: Vorlage:Beispiel
== Lösungsvorschlag von ~~~ == --~~~~
Siehe auch Hilfe:Formeln und Hilfe:Beispielseiten.