deletions | additions       diff --git a/spinning followup.tex b/spinning followup.tex  index bc9439a..b9f6c15 100644  --- a/spinning followup.tex  +++ b/spinning followup.tex 
...
Though the main purpose of this study is to quantify parameter estimates while accounting for spin, we will include two non-spinning analyses for comparison. The first is the analysis using the ``TaylorF2'' frequency-domain model, the results of which were presented in \citet{Singer_2014}. This model is very fast, cheap to generate, and is always used for medium-latency follow-up before more expensive analyses are done. The second non-spinning analysis uses the time-domain ``SpinTaylorT4'' model, with spins fixed to 0. This is not typically used in a follow-up scenario, but is included to demonstrate that any differences between the spinning and ``TaylorF2'' analyses are due to the inclusion of spin, and not systematic differences between waveform models.  When analyzing a GW trigger, the Bayesian analyses use `uninformative' priors for the source's properties. For some parameters, such as the location and orientation of the binary, uniformative priors (i.e., uniform in volume and isotropically oriented) are well-motivated. For other parameters (e.g., component masses, spins), current observations and theory do not motivate any particular choice in prior. For this study we use a prior distribution uniform in component masses, as in \citet{2013arXiv1304.0670L}, and spins with magnitudes uniformly distributed between 0 and 1 and isotropically oriented. We make the choice to not rule out compact objects with high spin because, since choosing any particular upper bound for spin magnatude would require choices regarding the unknown EOS of neutron stars. In Sec. \ref{subsec:prior_constraints} we look at more constraining spin priors, and particularly how such choices can affect mass estimates. \subsection{Mass Estimates}  For the sake of sampling efficiency, it is common to reparameterize the model to reduce the degeneracy between parameters, particularly those specifing the binary's masses. GW detectors are most sensitive to a combination of component masses referred to as the chirp mass, $\mathcal{M}_\mathrm{c} = (m_1 m_2)^{3/5} (m_1 + m_2)^{-1/5}$. For this study, we use the assymetric mass ratio $q = m_2/m_1$, where $0 < q < 1$, as the second mass parameter. Detectors are much less sensitive to the mass ratio, and strong degeneracies with spin make constraints on $q$ even worse. It is primarily the uncertainty in $q$ that governs the uncertainty in component masses $m_1$ and $m_2$.  Figure \ref{fig:mass_std} shows the distribution of mass standard deviations for the 250 simulated signals from the three analyses previously described. The consistency between the non-spinning analyses shows that the drastic increases in the uncertainty of mass parameters from the spinning analysis is due purely to degeneracies with spin, and not systematic difference between waveform families.