The image shows the data for the angular power spectrum of the cosmic microwave background as obtained by different experiments as well as a 'best fit' to the data.
It is obvious that only for the Archeops experiment are the error bars significantly smaller than the data itself, i.e. the other data should be considered as statistically insignificant (it is a well known fact of statistics that for overlapping error bars the difference between the corresponding data points have to be considered as statistically insignificant (see
http://www.graphpad.com/articles/errorbars.htm) and most of the error bars of all experiments apart from Archeops do indeed overlap).
The Archeops data, like the WMAP results, on the other hand, can be considered to be crucially affected by systematic errors (see the
WMAP- page).
