How KENNETH UDUT approaches Science News Headlines (step by step
   process): Here is my process:
   a) See news headline.
   b) Read article.
   c) Find data-points within news article. Numbers are usually
   good.
   d) Ask question: Can I draw conclusion based on limited data?
   Yes/No?

   If Yes:
   e) Draw conclusion, publish opinion.

   If No:
   e) Find authors of study.
   f) Study background of authors of study.
   g) Look for possible bias. [their horse-shit in particular]

   IF BIAS FOUND:
   h) Stop. Draw conclusion. Publish opinion.

   IF BIAS appears minimal, or the horseshit factor is
   statistically insignificant:
   h) Continue.

   i) Find original study. Priority order:

   j) Full article is preferred. Look at images first, then look at
   bold-faced type throughout text., then conclusion, then
   introduction.

   If that is enough, stop. Draw conclusion, publish opinion.
   If not, read inbetween text.

   or

   j) Abstract. Treat Abstract as identical to news article.

   or

   j) worst case: count how many articles REFERENCE said study, and
   determine the possible influence of said article upon other
   articles that reference it. This requires excessive inferring on
   my part, and I consider it less reliable, but sometimes that is
   all that I have to work with, as not everything is on the
   Internet yet.

   I live in Google Scholar sometimes.

   I may be a (knee) jerk at times, but not without reason.