In terms of star colors, that is exactly what we should, and actually do see. But that's because our ability to see red stars falls off rapidly with distance. The red stars are invisible in the older galaxies, we can see only the blue stars, because they are hotter & brighter, and their light therefore carries over larger distance. So if you look only at the star colors, you will get a very skewed picture of the evolution of galaxies, because most stars are too red to see. Blue stars do not mean young galaxies, they mean young stars. Star formation is episodic, so that a very old galaxy could easily undergo a burst of star formation, pump up its blue star population, and make you think it is young when it is really very old.

It is better to look at the global star formation rate as a function of time (redshift). That's why I linked to the Hopkins & Beacom paper where you can see the history you expect to see. At high redshift the star formation rate is low, simply because things haven't got started yet, except in a few places. But the star formation rate goes up dramatically as star formation really gets going, peaking about redshift 3 (about 11.5 billion years ago). Since then the global star formation rate has fallen off by about a factor of 10.

Aside from the selection effects involving star color, there is also the small numbers problem. You cannot hope to learn anything at all about star formation in the universe by looking only at 2 or 3 galaxies, as there are always exceptions. You need to look at thousands of galaxies, more if you can. That's how the star formation history studies work. So, even though we know that the global star formation rate is low at high redshifts, that does not exclude the local star formation rate from being quite high in some places (i.e, HUDF-JD2, if it is really as far as redshift 6.5; Chary, et al., 2007 argue that it is actually much closer, about redshift 2; this also serves to show the difficulty of determing the redshift for distant galaxies).

Consider the galaxies featured in the OP story. The "blue" galaxy at 7 million light years is NGC 300. But is it "blue", or do we only see blue stars slectively? And remember that the GALEX instrument looks only in UV, so it would not efficiently see red stars even if they were nearby. See Butler, Martinez-Delgado & Brandner, 2004. They conclude that "The main disk stellar population is predominantly old, consisting of red giant branch (RGB) and asymptotic giant branch (AGB) stars ..." and "Taken at face value, this finding would agree with the Davidge report of suppressed star formation there during the past 10⁹ yr ...". We expect blue star to be in the minority in any galaxy, and so we see here. The point is that the stellar population of the galaxy is peculiar to that galaxy and its own specific star formation history (not all galaxies have the same distribution of gas from which stars form).

The "red" galaxy at 62 million light years is the giant elliptical galaxy NGC 1316, also known as the bright radio source Fornax A. But this galaxy is a merger remnant (Horellou, et al., 2001). The merger event(s) trigger short term bursts in star formation activity. What we see now is a galaxy that was very blue, due to active star formation, about 100,000,000 years ago, but star formation activity has dropped off since then. So you look at the "red" color, as far as GALEX is concerned, and think perhaps that the galaxy is far older than it really is (about 3 billion years since the major mergers that formed it; Kaneda, Onaka & Sakon, 2007). The appearance of this galaxy, as with NGC 300, is dominated by its own peculiar history, not by cosmic evolution.

__________________
Don't try this at home - We're what you call "professionals" - MythBusters.