Test Methods for Halogen Lamp Emissions

It sounds like a valid approach to me. But you need to consider the aging effects of the other components of your circuit - the sensor, particularly, and the ADC. I would do your experiment at least twice - during the first month all your components will be aging - "burning in". By the time you start your second test, everything should have aged past the point where it will be changing with time.

In fact, I would repeat the experiment many times - your results will be more valid if they come from several trials an several bulbs.

Also - if you don't know already - don't handle your bulbs with your fingers. The oils will cause them to burn out quicker. Carefully wash each bulb with alcohol before the test.

And, needless to say - be sure your whole setup is enclosed so external light doesn't skew your results.

Good luck - it sounds like a fun experiment.

I'd be inclined to try to eliminate the effects of aging, temperature drift etc. by setting up 2 x (lamp+lens+sensor), with the sensor outputs measured by the same ADC, using an input multiplexer (e.g. a changeover relay). Drive and measure one of the lamps as you described in your OP, and use the second lamp/sensor as a reference, turning it on once per hour (or maybe 6 hours), allowing, say, 20sec warmup stabilization time (or repeated measurement until reading is stable), then save the reference reading and turn the reference lamp off. Run the reference lamp continuously for a couple of days before starting your test run.

It's then reasonable to assume that the reference lamp output reading is constant over the month, and you can correct the results from the test lamp accordingly.

Repeating the whole experiment, as bhankiii suggested, is also a good idea; it depends how long you're prepared to wait for your results.

Does your test reflect real life conditions? Halogens tend to dim if under-run but will regain brightness when run at maximum output.