The Mind-Reading Salmon: The True Meaning of Statistical Significance: Scientific American.
Should we worry that so many of the research results published in Construction Management and Economics are justified on the basis of statistical significance?
Management and Economics in Construction 
P-values are better than 100 % speculation.
The 0.05 is rather too low than too high.
No research result is the final answer, and as long as the researcher tells what he or she has done and what the conclusions are based on, it is ok.
Instead, my personal opinion is that the researchers not considering the effect of randomness etc. whatsoever, are the ones whose works need not be published.
P-values might be “better than…” but what are they “better at…”?
Seemingly contrasting quantitative research with speculation (qualitative research?) you write that “…as long as the researcher tells what he or she has done and what the conclusions are based on, it is ok.” In my view, this only leads to incrementalism and bland research dictated by procedural correctness!
Take a look at the ill-conceived questionnaires posted at cnbr-l by students and see what I mean.
In a field like CM, where we continuously struggle to define the scientific boundaries, anatomy and mechanisms, we should perhaps focus more on the substantive (i.e. speculative) rather than procedural matters.
I, for one, welcome the new editorial policy of CM&E!