On 29/9/07 09:58, in article ,
"Martin" wrote:

On Sat, 29 Sep 2007 08:09:45 +0100, Malcolm
wrote:


In article , echinosum
writes

Sacha;750615 Wrote:
This is of interest to all those who buy seeds, both veg. and
flowers:
http://tinyurl.com/yq6fd4

Apparently, several suppliers have been found to have dead seeds in
the
packets and some of them in very high numbers, too.
I read a report of this in a newspaper which said they had tested "400
seeds" from each supplier. I think this means they tested about 8
packets of seeds from each supplier. I think that is far too small a
sample in order to decide which are the rogues and which are the
reliable suppliers.

But that's to suggest, as Martin does, that Which should be carrying out
a statistically significant test taking large samples from each supplier
and of each plant.

Believe it or not in the days when Which? did reliable testing it is exactly
what they did for this sort of thing. This was the reason why the subscription
fees were high.

But, as I understand it, this isn't, and never has
been, Which's approach. That kind of testing is for the seed producers
to carry out. Which behave as the ordinary shopper (in this case,
gardener) does. They acquired a few packets from each source and tested
them. Actually, 400 seeds is quite a lot for some plants. The commoner
veg like lettuces and carrots often have as many seeds as that in a
packet, but many have much less than that, down to a few tens or
sometimes only 10-20. The Organic Gardening delphiniums, for example,
have 65 seeds for one variety and 38 for another, so Which have tested 8
or 10 packets.

When Which test washing machines, they don't test hundreds of them, but
take a single example just as would the average shopper.

because mass produced objects like washing machines have been quality
controlled
by the manufacturer. They should all have the same level of quality.
What you buy is the same as what left the factory. In the case of seeds what
you
buy isn;t necessarily in the same condition it was when it left the seed
producer

I take a
computer magazine which does comparative tests. They test a single
example of equipment from each manufacturer. It's up to the latter to
make sure that their products are reliable enough to allow for this kind
of minimal sampling.

The fact that who were the rogues in relation to "flower seeds" had
very little correlation with who were the rogues in relation to
"vegetable seeds" suggests to me that the results had a substantial
random factor arising from it being a poorly designed experiment of
little statistical significance.

I don't see how it can possibly be called an "experiment". It wasn't,
and nor did it need any kind of statistical significance applying to it.
It was a simple test of a product (seeds) on behalf of the consumer,
doing exactly what the consumer does, buying some seeds and sowing them.
And, of course, it is up to the consumer (and the seed selling
companies) how much notice they take of it.

In the case of seeds the results are meaningless. The history of the seeds is
also important.
This is one of the reasons I no longer subscribe to Which?

This is a proper independent test of a camera
http://www.dpreview.com/reviews/nikoncp885/ compare this to the test of the
same
camera done by Which? and wonder why Which? bothered.

Compare the conclusion with Which?'s conclusion and wonder if they both tested
the same camera http://www.dpreview.com/reviews/nikoncp885/page18.asp

Note "Red / Cyan can appear over saturated, can 'dominate' an image"

The result of this is that almost no photos of flowers produce realistic
results. Campanilla appeared pink instead a deep blue.

But but but - seeds either come up or they don't. It isn't a matter of
subjectivity in terms of personal preference, how you see blues or reds etc.
They live and sprout or die and don't.
--
Sacha
http://www.hillhousenursery.co.uk
South Devon
(remove weeds from address)
'We do not inherit the earth from our ancestors, we borrow it from our
children.'