I want to know how do I best determine if my thermometer(s) display
correct temperature. I have about 6, all of wich are digital, and some
are wireless, with an operating range of about -50f to 160f. Should
one expect different testing requirements for steel and plastic
sensors? In my case, testing in boiling water is out of question.

I've heard about giving it an 'icebath', but the opinions were
divided; some sources claim icewater holds exactly 32f, others about
33-34f. So which is it?
Also, setting your thermometer in a glass of water stored in
refrigirator. The temperature drops constantly, then stops for a
while, then drops again as ice encircle the sensor. Would such a way,
then, be accurate and/or better than the 'icebath', and what
temperature should I expect here? Any other ways?

Provided you are a normal healthy human, you already have a
temperature source that is accurate within about 36.1 - 37.8 degrees
Celcius - your body temperature. Put the probes in pairs under your
tongue for two minutes and record each thermometer's result.

Cross-check each thermometer against two or three others.

Owain

Hi owain-ga, thanks for the reply. It could be an option, but I am
interested in more correct results. 1.7 c seem too high a variation to
me. I am more interested in determing wich of my thermometers fall
within at least 0.5 C gap.
Also, I think an outdoor thermometer is not that suitable taking body
temperature. But i'll try it, anyway! Still, I'd like to know wether
it's possible to point an exact temperature in ice water (opinions
were devided wether it should be 0.0 or even 1.0 C). In my experiment
(ice covered glass of water taken out of the refrigerator), 4 of the
thermometers displayed following results:

1. -0.5

2. +0.6

3. +0.3

4. -0.7

So I wonder wich one is the most accurate! Can it be possible the ice
water hold minus temperatures (-0.1, -0.2)? After all, the freezing
process cannot take place untill temperature fall below 0 C. Any
opinions?

It is possible for ice water to be minus Celcius. If water is impure,
its freezing point varies.

Owain

The Febuary 1999 issue of Scientific American has an article in the
Amature Scientist feature describing how to make a triple point
calibration device. It's been a while since I read the article so I
can't say for sure how difficult it is to make but I'd check my local
library to see if you can find that issue. I'm pretty sure that
Scientific American also allows people to download old articles for a
small fee if you can't find a copy of the article or any other
reference to it online.

From memory, the basic idea is that the freezing/melting and
boiling/condensation points of water depend on the pressure, this
device puts the water in a point where all three states exist at the
same time.

I don't know if this device is what you need, it might be outside the
range of the thermometers you're trying to test.

The thermocouple devices used in your wireless/digital thermometers
have two different metals in electrical contact.  The bimetallic
junction generates a voltage based on the temperature of the two
metals.  Classically a duplicate junction is placed in an ice-water
bath and the voltage is generated based on this difference.  Your
digital themometers have an electronic ice-point in the thermometer
which maybe correctly set or not.  This may account for the variation
you see in the thermometers.  Also depending on the resolution of your
thermometer electronics it may not be accurate to less than 0.5
degrees C in any case.  In the laboratory, it is uncommon to have an
accurate measurement of less than 0.1 C in any case without speciali
equipment.

Retry your ice bath test with a styrofoam of ice filled with water. 
Let stand for 15 minutes to come to decent equilibrium.  If the
thermometer is waterproof put it into the ice bath.  If not place it
in a platic bag and submerge into the water.  Allow 10-15 minutes
before reading temperatures.  If you have a variation at that time -
that is due to the manufacture of the temperature.  If you have
further questions about the thermometer you should consult the
manufacturer about the specifics of the thermometer operation.
Depending on the commonly dissolved materials in the water and ice, 
the ice bath may be  -0.1 tenth of a degree from 0 degrees Celcius. 
But for your measurements, it is safe to assume that the bath is at 0
degrees.

PS - This is simple - where's the answer guy?

Wow, arbitrary and drimagine, many thanks for your help! I haven't
been able to locate the Scientist article, but then again, drimagine's
clarification was more than enough for me. Both the thorough
explanation on the basics of digital theromometers, as well as a
step-by-step guidance for attaining correct testing results were
completely satisfactory. This is the answer I was looking for! Now,
when I know the exact temperature of an icebath (under sea-level
pressure with pure water in a clean container), I can easily tell wich
of the thermometers are closer to the norm. I repeated the testing and
got very much the same results I've posted. In this case, nr. 3 & 1
were the most reliable.

Drimagine-ga, had you been a researcher, I would rate your answer 5/5.
Unfortunately as we see, the researchers aren't always doing their
job. It is here that other users can help eachother out. Thanks again
everyone for your involvement!

Antarctidus-ga