taxiq
100+ Head-Fier
- Joined
- Dec 27, 2015
- Posts
- 140
- Likes
- 41
The theory of statistical significance requires a sample size with substantial a number of units for there to be statistical significances. My recollection is 15 units in the sample are needed for 90% confidence the results are reasonably representative of universal population, i.e. the production run. Obviously this degree of rigor is never met by reviewers who are basing their findings on a single units.I seriously believe there needs to be some type of standardized protocol where a headphone being reviewed is tested to be within certain standard deviation to the "average" or "typical" unit in production. This needs to be tested and stated at the beginning of the review to be a legitimate one. Now that I think of it, isn't this sort of common sense?
Come on! Just after the LCD4 vs Utopia debacle, another one?? We are better than this. How am I to ever trust any review, I wonder... (no offense and nothing personal)
We live in an imperfect world and are making due with what we have, flawed as it is.