Non-Null results of recent MMX type experiments
I think the most likely explanation for the Null result of the MMX experiment was that the result wasn't null. I have reposted below links from very knowledgeable
people that the measured aether speed is not the result of reading
error bars incorrectly and that very recently perfomed expermiments
(2005) are producing mixed results:

I have done some recent research about recent attempts to detect the
aether with MMX type experiments. The most recent attempts use rotating
cyrogenic cavities. I have listed links to the relevant papers, some of
which show significant non-NULL results. Only one of the experiments
show genuine NULL results. In addition, there has been some analysis of
these non-NULL results indicating that they are consistent with Dayton
Miller's research which showed the aether did exist and the speed of
the aether they detect is consistent with the motion detected in the
cosmic background radiation from the NASA COBE mission.

This paper does genuinely produce results which are consistent with
zero ( The range -1.9 +-5.2 x 10^-15 does include zero). However, their
previous experiments and similar experiments are not consistent with

This is a previous experiment done by the same scientists as the
previous citation. This experiment returned non-NULL results as 2.6 +-
1.7 X 10^-15. I personally contacted Holger Mulller via email and asked
him about this result since  the results were not consistent with zero.
He replied that he thought that the experiment returned a no-anistropy
result. He explained that if you divide the main number 2.6 by the
range value 1.7, you get only only 1.5 (one sigma error) which in
mathematical terms means that the result has a 23% chance of falling
outside of the range and could possibly include zero. A result would
need to have a 5 sigma error (like 2.6 +- .5) to claim the detection of
an effect. An experiment I cite below (0504109) does have the required
5 sigma error.

This paper indicates that an Feb 2005 experimental result is consistent
with Dayton Miller's results. It also goes on to predict new results.
This shows that although the results appear to be small, they are
consistent from what previous MMX experiments have produced.

This is the original paper cited as the Feb 2005 experiment. The
results are buried in Table I on page 3. The C2 value which represents
the correlation between frequency shift and the rotation of the
experiment is 11 +- 2 X 10-16. This has the required 5 sigma range to
be considered valid. This range is totally inconsistent with a zero
value. The experimenters do acknowlege and attempt to discredit this
result by attributing it to thermal causes, but do not elaborate.
Considering the rotation period was 10 minutes and that the experiment
is cooled to close to zero kelvin, I can hardly see how an external
labratory thermal gradient could cause a difference.

This paper (2005) indicates that the speed detected by MMX like
experiments match up with the motion detected by the cosmic background
radiation. This confirms the CMB represents the preferred reference
frame from which absolute motion should be detected. The experiment
references 2 similar cyrogenic rotator experiments.

This is another paper indicating that the old MMX experiments indicate
speeds which match up with the motion of the cosmic background
radiation as measured by the NASA COBE satellite.

See Table 2 - while the experimenters don't say they found anistropy -
their graph clearly shows a repeating variation in the data occuring
over a 24 hour period corresponding to the rotation rate of the
experiment. This is what one would expect if the aether exists. Their
reduction of the limit to 10^13 might sound impressive, but is larger
than the expected value that one would predict on Dayton Millers work.

In conclusion, it appears we have one recent result which points to
anisotropy and all the rest that don't. However, there is reason to
believe that the values coming out of these experiments are consistent
with the aether drift rates predicted by Dayton Miller and by the
observation of the cosmic background radiation. I would say that more
and better experiments need to be conducted to gain a consensus on what
is really happening. However, the question of anisotropy (aether drift)
is far from being answered.

From what I can see, there is an extreme bias towards no-anisotropy (no
aether) - to the point where experimenters even don't believe their own
results. My challenge to anyone interested in this topic is to look a
the new references I have cited and suspend your bias long enough to
see if there is any merit in the claims rather than just grabbing hold
of the non-NULL results and ignoring the rest.

If you think about it, anisotropy exists, it could open up a whole new
avenue of physics, so it shouldn't be dismissed so lightly. Some people
do currenty theorize that such ansiotropy may exist as the result of
string theory, etc. so even mainstream science continues to investigate
the question.