Academia Stack Exchange is a question and answer site for academics and those enrolled in higher education. It only takes a minute to sign up.
Sign up to join this community
I encounter this term "principled approach" in some papers of computer science. Since I' m not a native speaker, I don' t quite understand what this means. And I didn' t find any results online.
I' m not sure if this site is appropriate for such questions. Please let me know if I posted at the wrong place.
A 'principled approach', at least the way that I've been exposed to this term, implies due care and diligence with regards to the rigor and discipline used in the materials context. A paper that describes a principled approach would be one that is presenting a procedure for the execution or evaluation of a given subject matter. For example, if I picked up a paper titled 'A principled approach to algorithm selection and implementation' then I would expect the contents of that paper to clearly enumerate a system of algorithm analysis, with exhaustive supporting documentation.
Conversely, a paper which uses a principled approach would be one that follows such a detailed and rigorous methodology that the data collected from its research may be considered to be functionally with out bias and with a low probability of corruption or inaccuracy.
(I am in mathematics, but similar language is used roughly similarly, I believe.) As a place-holder answer: a "principled" approach in science is at least opposite to a quick-and-dirty, or ad _hoc_, or "kludge-y" approach, the latter three synonymous expressions meaning that the priority is getting some result out, perhaps even finding some rationalization for the conclusion one wants. Obviously a non-principled approach more lends itself to corrupted (but also quick, desired, easy) results.
The "principled" approach "takes the high road", does not bias conclusions, does not rationalize-away weaknesses or flaws in methodology or information.
That is, one could hope that a "principled" approach involves no conflict of interest for the parties involved, and could be trusted. At its worst, "unprincipled" approaches (which no one would ever admit to, except perhaps as a mildly perverse claim to fresh unorthodoxy) produce completely untrustworthy outcomes, because those outcomes are chosen in advance, and whatever results are obtained are "interpreted" to support the original premise.
A hilarious example I witnessed was a computer science M.S. (details elided to protect privacy), on which I was an "outside examiner", in which "the goal" was to prove that two bunches of events were correlated, thus proving that the people who were promoting the one as "cause" of the other were right, and people should invest in their product. (Nevermind that correlation is not causality.) The guy failed to find any correlation in any of the first twenty or so statistical tests he applied... but he kept at it, until he found a statistical test that did seem to assert a slight correlation.
Of course, what he had really proven was that there was apparently no correlation... but, taking an "unprincipled" approach, claimed the opposite of what his own evidence showed, etc.
