Thanks so much for submitting this question! I am glad to see that you've inspired your student to ask interesting questions like this.

The results of your student's experiments are interesting. The fact that the red shirt is the coolest makes sense since a red shirt will reflect red light, limiting light absorption and consequently, heating. With that said, the black should be the hottest as black pigment absorbs all light so if you haven't already reproduced the result multiple times, I would recommend attempting to do so. If you already have proved reproducibility, I believe that part of the issue could be in the experimental design.

One potential pitfall could be in the manner in which the temperature of the shirt is assessed: if an IR thermometer is being used, the shirts could have different emissivities, leading to inaccurate temperature readings. For a further explanation on the role of emissivity in IR thermometer errors, please read the following link . I would recommend the use of a temperature probe that is either in direct contact with the shirt or with a thermal insulator that is touching the shirt so that heat can be amassed there. Leaving the bulb on pointed at a shirt that is resting on a plastic pane with a temperature probe attached to it would be a relatively reliable setup.

Another possible issue could be caused by degradation of the light-bulb's color coating, which would engender different incident light fluxes for different shirt tests. Testing with a coating-less lightbulb might give you some intuition there.

I am glad you reached out about this interesting conundrum and I wish you the best with resolving it.