[Paper Review] Consistency of effective field theory analyses of the BOSS power spectrum

Théo Simon, Pierre Zhang+2 more|arXiv (Cornell University)|Aug 11, 2022

Cosmology and Gravitation Theories5 citations

TL;DR

This paper investigates inconsistencies in cosmological constraints from the BOSS power spectrum using effective field theory (EFTofLSS), finding that differences between two major likelihood pipelines—PyBird and CLASS-PT—mainly arise from distinct priors on EFT parameters ('West-coast' vs. 'East-coast'). Despite non-Gaussian posteriors causing shifts up to ~2σ in cosmological parameters due to marginalization, results remain consistent at ~1σ when priors are broadened, and cosmological constraints from BOSS are not in tension with Planck under ΛCDM when combined with CMB data.

ABSTRACT

We assess the robustness of $Λ$CDM results from the full-shape analysis of BOSS power spectrum using the one-loop prediction of the Effective Field Theory of Large-Scale Structure (EFTofLSS). The public likelihoods PyBird and CLASS-PT lead to results in agreement only at the $1σ$ level, despite the fact that they are derived from the same BOSS dataset and theory model. We perform a thorough comparison of the various analyses choices made between the two pipelines, and identify that the differences come from the choice of prior on the EFT parameters, dubbed "West-coast" (WC) and "East-coast" (EC) prior, respectively associated to PyBird and CLASS-PT. In particular, because posteriors are non-Gaussian, projection effects from the marginalization over the EFT parameters shift the posterior mean of the cosmological parameters with respect to the best-fit up to $1σ$ in the WC prior and up to $2σ$ in the EC prior. We quantify that best-fit cosmological parameters extracted from BOSS given the two prior choices are consistent at $\sim 1σ$. The consistency improves to $\sim 0.5σ$ when doubling the prior widths. While this reveals that current EFT analyses are subject to prior effects, we show that cosmological results obtained in combination with CMB, or from forthcoming large-volume data, are less sensitive to those effects. In addition, we investigate differences between BOSS measurements. We find broad agreements across all pre-reconstructed measurements considered ($<0.6σ$), but the two available BOSS post-reconstructed measurements in Fourier space, once combined with the EFT full-shape analysis, lead to discrepant Hubble parameter $H_0$ at $\sim 0.9σ$. Given the various effects we discuss, we argue that the clustering amplitude $σ_8$ measured with BOSS is not in statistical tension with that inferred from Planck under $Λ$CDM.

Motivation & Objective

• To assess the robustness of ΛCDM constraints derived from BOSS full-shape power spectrum analysis using one-loop EFTofLSS.
• To identify the source of discrepancies between cosmological results from two leading likelihood pipelines—PyBird and CLASS-PT—despite using the same data and theory model.
• To evaluate the impact of EFT parameter priors, measurement choices, and analysis settings on cosmological posteriors, especially in the presence of non-Gaussianity.
• To determine whether the clustering amplitude σ₈ from BOSS is in statistical tension with Planck under ΛCDM.

Proposed method

• A systematic comparison of analysis choices between PyBird and CLASS-PT likelihood pipelines, including EFT parameter priors, scale cuts, number of multipoles, and reconstructed vs. pre-reconstructed measurements.
• Direct comparison of cosmological posteriors using identical datasets but different priors and analysis settings to isolate the effect of prior choice on parameter shifts.
• Quantification of posterior shifts due to marginalization over EFT parameters, particularly the non-Gaussian projection effects that shift posterior means relative to maximum a posteriori estimates.
• Evaluation of consistency across multiple BOSS power-spectrum measurements, including pre- and post-reconstructed data in Fourier space.
• Assessment of the impact of fixing or varying nuisance parameters like ω_b and n_s on cosmological constraints.
• Comparison of results with independent analyses (e.g., Chen et al. 2022, Brieden et al. 2022) to contextualize the findings within the broader literature.

Experimental results

Research questions

• RQ1 What drives the ~1σ discrepancy in cosmological constraints between PyBird and CLASS-PT likelihoods despite using the same BOSS dataset and EFTofLSS model?
• RQ2 How do different priors on EFT parameters ('West-coast' vs. 'East-coast') affect the posterior distribution of cosmological parameters, especially under non-Gaussian posteriors?
• RQ3 To what extent do analysis choices such as scale cuts, number of multipoles, and reconstruction status influence the final cosmological constraints?
• RQ4 Are the H₀ constraints from post-reconstructed BOSS power spectrum measurements in Fourier space consistent with each other when combined with EFTofLSS?
• RQ5 Is the σ₈ value inferred from BOSS in tension with Planck under the ΛCDM model when accounting for EFT parameter prior effects?

Key findings

• The primary source of discrepancy between PyBird and CLASS-PT results is the choice of prior on EFT parameters, with the 'East-coast' prior causing larger shifts in cosmological posteriors due to non-Gaussian marginalization effects.
• Posterior mean shifts in cosmological parameters relative to maximum a posteriori estimates reach up to ~1σ for the 'West-coast' prior and up to ~2σ for the 'East-coast' prior due to projection effects from non-Gaussian posteriors.
• Maximum a posteriori cosmological parameters from the two pipelines are consistent at ≤1σ, improving to ≤0.4σ when both priors are doubled in width.
• The two post-reconstructed BOSS power spectrum measurements in Fourier space yield discrepant H₀ constraints at ~0.9σ when analyzed with EFTofLSS.
• The clustering amplitude σ₈ measured with BOSS is not in statistical tension with Planck under ΛCDM when EFT prior effects are properly accounted for.
• Cosmological constraints from BOSS combined with CMB data are less sensitive to EFT prior effects, suggesting robustness in future large-volume surveys.