Lemma 4.7, p. 23 — f_q(k) ≤ ¼(k + C(q)) with a constant C(q) depending only on q
ProvedCorrClusterExp.IBR.lemma_4_7cluster-randomizationexperimental-designp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1worst-case-variance
Let . There is a constant , depending only on , such that for every integer
where is the optimal value of (1) for clusters all of size one. Thus grows linearly in with slope ; this is the rate used in the asymptotic analysis of §4.2.
Differences from the source:
- reading: " is a constant that depends only on " → chosen after and before ; reason: the literal meaning of a constant depending only on .
- encoding: for clusters → ; reason: an instance needs at least one cluster.
Preamble
import Mathlib import Definitions.Def_CorrClusterExp_IBR_Setting
Formal statement
namespace CorrClusterExp.IBR
theorem lemma_4_7 (q : ℝ) (hq0 : 0 < q) (hq : q ≤ 1 / 2) :
∃ C : ℝ, ∀ k : ℕ, 1 ≤ k → fq q k ≤ 1 / 4 * ((k : ℝ) + C) := by sorry
end CorrClusterExp.IBR
Source
Candogan, Chen, Niazadeh, Correlated Cluster-Based Randomized Experiments: Robust Variance Minimization, SSRN 3852100 (version of March 22, 2023), p. 23, Lemma 4.7
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.