The Axis of Evil: revisited

1
The Axis of Evil revisited
Kate Land

• astro-ph/0611518, 0502237, 0507289

2
THE ISSUES

• AXIS OF EVIL
• Planarity
• Alignments
• Low-power
• Ecliptic issues

• OTHER
• Asymmetries
• Cold spot

3
PLANARITY AOE
Intriguingly, both the quadrupole and the
octopole are seen to have power suppressed along
a particular spatial axis, which lines up between
the two, roughly towards (l, b) (-110, 60) in
Virgo.
Tegmark al.PRD(68)123523
4
A posteriori planarity
Quadrupole l 2
5
A posteriori planarity
Octopole l 3
6
Planarity is quite normal -
BUT THE AXIS IS NOT!
7
Axis of Evil
Independent of Gaussianity
Hard to test
8
Maximum Likelihood

• Avoid instabilities
• Allow model comparison (Bayesian V Freq)

9
Models

• Null

10
Maximum Likelihood
11
Significance
Compare to simulations
Information criteria
planarity AIC 2, BIC 2.8 m-pref
AIC 2.8, BIC 0
Jeffreys scale gt 2.5 strong
12
Bayesian evidence
13
What can we conclude?

• Evidence for AOE depends on model and method!
• Similarly for nslt1 (Liddle astro-ph/0701113)
• Important consequences
• Combined with other anomalies
• SYSTEMATICS Satellite, data pipeline,
• ASTROPHYSICAL EFFECT Foregrounds, local lensing,
  
• COSMOLOGICAL FEATURE Non-FLRW models, non-trivial
  topology,

14
The Axis of Evil II - Summary
Correlations seen for l 2,3,4,5
But how do we test the significance? Propose a
model
Frequentist
Bayesian
Planarity
m-preference
15
Priors
16
Axis of Evil (260,60) Dipole (264,48) Virgo
(260,70)
Max asym axis (57,10) Ecliptic pole (96,30) SG
pole (47,6)
17
Information criteria
18
??