Models of gravitational waves: Do we need to modify our understanding? So, now we got the first observation of gravitational waves, and that's indeed an amazing revolution in the field of astronomy and cosmology. So one can start discussing whether the previous model(s) were completely to the point or not, isn't it ? I mean, for decades gravitational waves were kind of a myth in science, but now one can confront models and experiments, as in any experimental science. 
So I was wondering whether the theory of gravitation needs to be adapted to some unexpected behaviours of the gravitational waves already detected, if they are alternative models describing the observations, and so on. In short, where are we in the scientific debate about modelling gravitational waves ? 
I guess it's well too soon for discussing this now, but still I can ask the question ;-) 
 A: I think the answer is clearly no, we don't.  The detections have agreed extremely well with numerical models based on GR so far.  So this is another test that GR has passed with flying colours.
It may well be that we will learn new things -- and with luck surprising things -- about the statistics of objects and this might influence cosmological models.   Indeed the initial detection was of black holes with masses which were thought to be uncommon I think.  However this will take time as there need to be many detections.
(It's clear from the comments below that 'myth' in the question is not meant in the sense it now often has of something presumed not to be true).
A: The best review of tests of general relativity, and including the results  that can be obtained from gravitational waves is from a Living Reviews article at http://relativity.livingreviews.org/Articles/lrr-2014-4/title.html
The other good place to look is at the LIGO publication from the first and second detections. The first one in particularly had a whole paper or section on tests of GR. See https://dcc.ligo.org/public/0122/P1500213/031/paper.pdf
All results so far have been as predicted by General Relativity. 
The tests and parameters of the theory depicted in the first reference above are the 10 parametrized post Newtonian expansion of any metric theory of gravity. Those 10 are the only ones determining any metric theory. It includes GR, any linear-tensor theory such as Brans-Dicke and many others related, and for instance vector-tensor theories. The latter have been proposed as one way to predict the galaxy rotation curves without any need for dark matter. That first reference also shows the limits on those 10 parameters form many of the relativistic effects measured before gravitational waves were detected, most are pretty tight but a couple in the 1 percent uncertainty wrt GR. It also talks about some of the gravitational wave (GW) tests possible - for instance a metric theory could conceivably predict 6 different polarizations for GWs, but GR says only two are possible, Scalar-tensor theories adds another possible scalar polarization, and others add 3 others. The GW detectors for now can only detect two. There will be other LIGO like detectors where those other possibilities can be ruled out.
The LIGO papers and many others before the LIGO results also identified other test, for instance the possible detection of gravitational wave multipole patterns that are NOT consistent with the known only-possible solutions for a black hole (BH) that are consistent with the 'No Hair' theorem (that BHs can only have mass, angular momentum and charge, nothing else. The observations from LIGO have been consistent with that. With additional LIGO like detectors aligned differently, longer baselines, and later space-based interferometers, this will be able to be mapped much more accurately, and allow us to determine if BHs became as GR dictates. With longer baselines and space based 1 million km baseline Interferometers  we will be able to see gravitational waves from much larger structures such as supermassive BHs, and even from cosmological effects. Those will test the cosmology theories, as well as GR.
With GW astronomy we wil be able to see where no electromagnetic observations are possible, such as inside neutron stars and earlier in the cosmological expansion, before the recombination time from where electromagnetic waves can not escape (and thus we can not see).
It s going to test more exactly GR, as well as get more astrophysical and cosmological data. 
