Hmm, so what is the overall take-home message from these two quite different studies? The amount of evidence available is far too small to draw strong, generalizable conclusions, but one can combine the results with one’s personal understanding of the world of software development and come out with at least better-founded prejudices than one had before. When I do this, the result reads somewhat like this:
With respect to programming languages, at least for small programs that perform some kind of text handling (or some such general activity), it appears safe to say that scripting languages are more productive than conventional, statically typed languages.
With respect to the efficiency of such programs, it is usually much more important to avoid the wrong programmer than to avoid the wrong language. There are many ways (and languages to use) to get it right.
If very high performance or very low memory usage are needed, C and C++ are still superior if (but perhaps only if) the programmers are sufficiently capable.
There appear to be language-specific cultures that can sometimes lead programmers using different languages to take different approaches to design.
Regarding web development platforms in particular, for straightforward web-based systems in the hands of capable teams, it appears that the language is the particular framework used and the level of mastery the team has with that framework are much more important than the language.
In particular, when capable professionals are using it, PHP is a lot better than its reputation.
There appear to be cultures particular to web platforms that can sometimes lead teams working on different platforms in different directions.
Your own view may vary, and your mileage will vary as well. A lot more studies like these are needed before we really understand where the individual strengths of individual languages and platforms lie or what typical characteristics emerge from their use. Furthermore, both of these questions are moving targets.