jmki - Your measurements and calculations all appear to be within range and accurate as to power dissipated by the 50KΩ load. The solar panel is really not working to produce power, however, and leads to a flawed experiment. It needs an appropriate working load from which to take readings and make your calculations. You are essentially just making 'no load' open circuit voltage measurements. I don't know how much current your solar panel is capable of producing, but it seems to have been manufactured to produce an operating voltage of 12 VDC (you measured over 16 VDC with your highest intensity light source). I would try various loads using, say, 68Ω resistors. Take three of them and take measurements as a single resistor, 68Ω, then add an additional resistor, in series, one at a time, to give you several different loads of 68Ω, 136Ω, and 204Ω. This should put your panel into an active working zone, producing real power, so that you can make your measurements and calculate power. Use at least a 2 watt rated resistor for the first, single resistor test. The rest can be 1 watt. The experiment will also demonstrate the fact that your source and load impedence selection, or matching, will affect optimum power transfer. For example, each different bulb will produce a different power output with different loads. Good luck, report your results if you can. As an aside, you expected a linear relationship between illumination and voltage/power. There are few purely linear outcomes in the real world, so expect some non-linear readings in your experiment.