Since most of my reviews are 1/72 scale, I always assume that 90% of the hits are by mistake, and as soon as the scale is discovered they leave and therefore hardly anyone actually reads them!
I think it is true to say that there are both good and bad inbox and built reviews. Many in box reviews are super informative about what is in the box and accuracy issues, but to write those can take a lot of research, and it's not that unusual for me to buy reference books for the purpose of writing that type of review, because I like to sound like I know what I'm on about even if I don't. Oddly enough in some ways built reviews can be easier to write because you only have to describe what you did and you have the photos and the instructions to guide you (plus a few notes). If I can't find any decent reference material then the build review is the best route. If you have no references and don't build it, you can end up with a pretty bland review (like my recent Ju87/JS-2 effort... not that great.)
The best reviews have both I guess, although bear in mind the number of releases that are just minor variants of kits we've already seen built, so the accuracy and what's in the box is probably more relevant (like my several Dragon Panther reviews - I'm not building everyone of them... ).
In terms of getting things wrong and then being pilloried for it, I always like to pepper everything with lots of words like think, probably, possibly, appears to be... Makes the text more verbose, but I am happy to admit when I'm not 100% certain and for almost all of us in most cases we are relying on research done by others.
Vettejack / John, don't be discouraged by that thread. I thought it was quite funny that the comment that kicked it off, quite unnecessarily impolitely, appeared, as I understand it, itself to be incorrect. Don't attach significance to it though. This is about plastic kits. They're toys. It's about fun.