I changed the lights in the entry way “chandelier” yesterday, with the help of a couple of my kids. I climbed the ladder, they took the burned out bulbs and handed me the new ones. It was a team effort.
It made me think of our previous house, where we had a similar setup with a chandelier in the entry way that had 7 or 9 or so light bulbs and the only way to access them is with a ladder. But with that one, all the lights were working when we moved in, and then, over the next few years, we saw the bulbs burn out, one by one, until there was only one of those 40W bulbs still working. This was the bulb at the far end of the bell curve of light bulb longevity. As an engineer, I appreciate a good bell curve.
My family whined and said it was too dim and we should replace the bulbs. But I didn’t want to replace all but one bulb and then have to go up there and replace that one before the others need replacing again, and then… well, you get the idea. Plus, this was a great experiment. I convinced them (my interpretation) of the need to find how far out that bell curve went, seeing that we had a 99.999 percentile bulb up there. We waited. We waited 2 or 3 years past the penultimate light bulb failing for that last bulb to finally burn out. We were exploring the very, very far end of the bell curve, and it was great. A little dim. But great.
I should have documented it.
Engineering Daze 
Jan 03, 2013 @ 13:38:38
How do you know if it was the bulb and not the inconsistency with the circuit of the lighting. Considering the builder in the neighborhood, they probably used the cheapest fixtures.
Jan 03, 2013 @ 17:38:11
Thank you for your comment of concern. I could review the quality control and quality assurance programs for the factories of all lighting production companies. I think you would see that even the low end of the quality system companies have sufficient consistency in their product line as to not merit such a great discrepancy in light bulb time life. Therefore, the majority of the effect is due to normal distribution, therefore, the bell curve. Again, thanks for your questions, reader.