sol invictus said:
Question: do you know what an "error bar" is?
I assume it means the same thing in physics as it does in other fields.
One of the things I tried to explain - unsuccessfully it seems - is that one cannot rely upon
the assumption that an error bar (or uncertainty) in the abstract of a paper is estimated by any particular method ... you actually need to read the paper to learn what the estimated uncertainty is and how it was calculated.
Is it one standard deviation? As such, it would be based on the uncertainty of all the measurements made in calculating some value.
In the case of WMAP team papers, a very great deal of work has gone into estimating the uncertainties, and I think you'll find the +/- values are 68% confidence limits. While this
seems to the same as one sigma, I think you'll find there are some differences. Further, at least some papers also quote 95% and 99% CLs (IIRC, I could be wrong), and if you do some simple arithmetic on them, you'll see they cannot be 2 and 3 sigma (respectively).
Normally I'd offer to go through a paper or two on this, to help you (and any other reader) understand what has actually been done, but in this case I am not prepared to do so ... the effort would be far greater than the time I have available (the analyses are very extensive, and are presented in a number of quite heavy papers); perhaps another JREF Forum member might like to step in?
So, was the 15 to 20 billion years so often thrown around not so long ago based on one or two standard deviations? In any case, 13.7 billion years is certainly outside of that range.
I see this point has already been addressed ...