Weather Banter - View Single Post

Martin Brown · August 25th 15, 11:23 AM posted to uk.sci.weather

On 25/08/2015 10:40, wrote:
On Monday, 24 August 2015 22:02:01 UTC+1, wrote:

We got a bad review recently about the cleaning in our cottage. But I know
for a fact that the place was spotless, the customer had clearly had bad
weather and was just having a pop. Pointless and unhelpful. Most of our
reviews are either excellent or very good overall. Yes some make useful
comments, but single outstanding good or bad reviews are often safely
ignored :-)

Tripadvisor would be a lot more useful for measuring quality if you could trim out
those who have review fewer than, say, 5 times, or indeed select your
lower bound
of reviewers' review count. I actually went to the lengths of
suggesting this to them!

Not sure that culling the low review counts ones helps all that much.
The dodgier ones are those mass produced positive reviews from the same
IP addresses under different pseudonyms.

Often one single review will be either a knee-jerk moan, sabotage, or
someone trying to big up their own place. I'd rather isolate it to reviews
of people who actually use TripAdvisor regularly to write stuff and
should (hopefully) be a little bit more level-headed.

Always worth topping and tailing the distribution and looking to see why
if something has exactly 3* average made up of half 5* and half 1*.

My favourite in 2012 was a particularly nice Estwing geologists hammer
that due to a moronic survivalists review was languishing with a 2*
rating. It wasn't very good as a ice climbing tool apparently!

The main reference work for JPEG has a similar bimodal distribution of
scores with experts rating it 5* and beginners 1*.

--
Regards,
Martin Brown