On Tue, 18 Dec 12, Steve Willner wrote:
Exactly so. Substantial data sets of supernova distance moduli are
published, so at least a first test shouldn't be hard. The DMs are
luminosity distances, so one has to use some theory to convert to
whatever distance 1/z is supposed to represent.

A difficulty is that SN DMs are presented with the assumption that
redshift represents time dilation of the SNe. That this yields that
the most distant SNe have sub-par luminosities seems to ring no alarm
bells amongst the researchers. One of the models that I'm juggling
treats time dilation as the square root of the redshift, which would
restore the expected luminosities of the farthest SNe. If raw SNe
data were published instead of the redshift-processed stuff, that
would be a boon to testing alternative models.

Measuring angular size is easy.
In principle, at least. Not always so in practice.

Indeed, just today there's a new preprint
http://arxiv.org/abs/1212.3869 , "Evolution of the Sizes of Galaxies
over 7z12 Revealed by HUDF", in which the highest-z galaxies look
unresolved. They jump through hoops to resolve them, but it reminds
me of the "elliptical hosts" of many hi-z quasars which also look
unresolved. And in that paper they don't publish the raw angular
sizes, it's all presented as kPc after processing using the FRW model.
Oh, for just one graph of raw angular size vs z !!

Eric