Additional-Review-(marginalization-over-joint-population+EoS-uncertainty).md

@@ -30,6 +30,9 @@ for each component mass, meaning that the total contribution to the weights shou
 
 The new version of the code relies on the user to specify the draw probability, so this is rendered moot. Within the tests described below, this was implemented correctly within [this script](https://git.ligo.org/reed.essick/mmax-model-selection/-/blob/gw-distributions/test/o3-nsbh/src/convert-event-csv#L48). **Update link once merge request is accepted**
 
+- [ ] Reviewer signoff on implementation in example
+- [ ] Reviewer signoff that in new version of the code, implementation of Jacobians should be done through inputs, not on this code base.
+
 **bad population reweighing**
 
 The previous version of the code implemented population priors by reweighing single-event PE samples. This was done within [this code block](https://git.ligo.org/reed.essick/mmax-model-selection/-/blob/c8b0d297d11b0a350c23ee44832d24b899729d8f/bin/mmax-model-selection#L227-270). Importantly, we see that the total weight for each PE sample was computed as the sum of population priors (implicit from [this block](https://git.ligo.org/reed.essick/mmax-model-selection/-/blob/c8b0d297d11b0a350c23ee44832d24b899729d8f/bin/mmax-model-selection#L263-270)). That sum was then normalized by the default PE prior [here](https://git.ligo.org/reed.essick/mmax-model-selection/-/blob/c8b0d297d11b0a350c23ee44832d24b899729d8f/bin/mmax-model-selection#L279). This means the overall weight assigned to each PE sample was
@@ -69,6 +72,10 @@ The correct sum is implemented within the updated code [here](https://git.ligo.o
 This difference can cause discrepencies between the old and the new code when there is a nontrivial marginalization over the population (i.e., when there is more than a single population sample `\Lambda_p`).
 These can be much larger than the statistical uncertainty from the finite number of Monte Carlo samples (see [below](#comparison-to-previous-results)), but have never been large enough to change the scientific conclusions drawn. Furthermore, [ApJ 915 L5 (2021)](https://iopscience.iop.org/article/10.3847/2041-8213/ac082e) only reported results assuming a fixed population (flat in source-frame component masses) and therefore this issue did not affect the published results.
 
+- [ ] Reviewer signoff on desired expression (within technical note)
+- [ ] Reviewer signoff on implementation within the code
+
+
 ## technical note
 
 A technical note deriving the Monte Carlo sums used to estimate the various probabilities and ratios of probabilities, along with their associated uncertainties, is included alongside the source code. It is available here:
@@ -91,6 +98,9 @@ In brief, I find
   * excellent agreement when the mass, spin, redshift distributions are exactly known (i.e., no marginalization over population parameters)
   * statistically significant (but scientifically inconsequental) differences when we additionally marginalize over uncertainty in the population. As described [above](#identified-bugs-and-associated-fixes), this is to be expected because of bugs in the previous version of the code.
 
+- [ ] Reviewer signoff on the above statements
+
+
 **GW200105 C01_PHMcombined_highspin**
 
 |EoS distrib|population|previous results|new results|difference (sigmas)|