The OP was trying to determine why his readings in his unexplained
circuit were variable particularly when the resistor was hot. He
didn't say what was the function of the circuit in which this resistor
was fitted, but since it was a 2 ohm 100W unit one would assume the
circuit was for some practical purpose. If you are building a circuit
with this sort of component it implies you have some intended use or
purpose for it and you wouldn't just pluck an old 2 ohm 100W resistor
out of somewhere and use it unless you were anticipating some hefty
current. It sounds to me as if the circuit wasn't behaving as expected
and he determined that the 100W resistor might be the cause of the
problem.
The OP just wanted to know why the variance in his readings when
measuring the cold Vs hot resistance and a number of practical
suggestions as to how he could do that have been made. I simply
suggested an easy method which didn't require anything other than a
DMM which would give him a good idea whether he was on the right track
or not and you responded in a negative manner which basically
pooh-poohed the idea.
I have no argument against using the DC voltage/current measurement
principle and it can be extremely accurate. However, in order to
achieve good accuracy when measuring a 2 ohm 100W resistor when hot it
is necessary to have a stable, low impedance source of current capable
of at least 5-6A in order to heat the 100W resistor sufficiently (ie.
need to dissipate at least 50W in order to get it hot enough). You
also need a meter for measuring current and another for measuring
voltage, each with good resolution and low error tolerances in order
to produce meaningful results when dealing with measurements on low
resistance circuits. These items may or not be in the possession of
the OP (he hasn't said so).
However, if he has access to a good quality DMM such as a Gossen
MetraHit 25S (now superseded, but it's what I use)
http://www.gossenmetrawatt.com/resources/zz_tam/hit22-26m/db_gb.pdf or
similar - as long as it has a ZERO ohms cancellation function, you can
get very good results by direct reading.
You conveniently snipped my suggested possible cause for variance in
his initial readings and made no comment on it. If the said resistor
was a slider type (OP doesn't say exactly what type he is using) the
wire might have been damaged by over-tightening the slider or some
other abuse, thus leading to errors especially when hot.
A little more explanation from the OP might be helpful.