How can a "review" be accurate without being firsthand? Otherwise it will be "true" he-said-they-said..... may be true but not "review"....
In some cases the specs alone count for much and they can be reviewed even if you've never touched the item. If the upcoming 5th gen iPad were to list at $1500, 12.5" screen, 640x480 graphics, no wifi, no cellular, 16GB memory I think you could pretty easily start ragging on it without ever having used it.
I can see both sides here.
Example 1.
The EOS 7d. A cemera folk shouldn't buy because it has noisy images and the af is lousy.
How many times do we hear that?
Am I too much of a fanboy when I point out that my experience of the 7d (& all d4/18mp EOS cameras I've owned) need some work on the raws, and that the af is good when you take time to understand it, set it up for your needs and use the right mode for the right job. Actual experience.
I bought the 7d based on a usp (only 25fps dslr on market at launch) and the reviews at the time (detailed from AP Magazine and Dpreview, hands on)
Some folk who slate it do so because they (i guess) didn't post-process correctly, want jpegs out the cam, didn't spend time getting to know the af.
Are they wrong to recount thrir experiences? Are they wrong to recount their pals experience even though they haven't used it? Are they wrong to tell you what they read on a forum?
Just where is the line?
Of course folk chip in with 'my 5d3 is much better, my 1dx is much better' well, duh! Doesn't answer the question of whether the 7d is a good buy for somebody with 7d money.
Are these guys trolls? Of course not.
Example 2.
The m. Thank god. I'll repeat that the fw1 was a pig out the box in wide flexi-zone. For me and my way of working it's actually great. Even better with fw2. Were the reviewers right? Given the target market of powershot upgraders, perhaps. But it wasn't the whole story. Are the folk who slated without ever touching one right? No. But in this case I'm glad they did.
Example 3.
Fcpx. Never installed it. Never used it. Never will.
Knew from spec sheet that without rs232 device control in suite, that without omfie export or laga y support it was a lame duck for me. Or rather my clients who still proffer tape for me to work with.
As i clunked along on 32bit fcp7 I envied the reported speed gains in rendering, the background transcoding.
But consistently every review said I had to relearn my editing technique.
Killer for me.
And adobe obliged with track based editing with 64bit rendering, no transcoding, legacy fcp shortcuts
I knew from the specs that I would need to replace my mac pro to run fcpx, and I knew there were better products for me. Or did I?
Aren't i and everybody else guilty of affirmation bias? Blind to flaws because we want to love (what for most of us) are toys, trinkets. In terms of value for money, do hobbyists really get good value from bright expensive glass? Probably not. But it gives us status, kudos amongst our peers. The white lens club.
And of course the images are so much better. Or whats the point?