# Thread: Which distribution to use?

1. ## Which distribution to use?

A light bulb is warranted to last for 3,000 hours.
a) What probability distribution would you recommend to model the time to failure?
b) If the true mean of time to failure is 2,750, what is the probability that the bulb will last at least 3,000 hours?

I'm thinking its poisson distribution that I have to use but I did the formula and got a wack answer. Any help would be appreciated.

2. Originally Posted by Marko_02
A light bulb is warranted to last for 3,000 hours.
a) What probability distribution would you recommend to model the time to failure?
b) If the true mean of time to failure is 2,750, what is the probability that the bulb will last at least 3,000 hours?

I'm thinking its poisson distribution that I have to use but I did the formula and got a wack answer. Any help would be appreciated.
For starters, the distribution must be continuous. The Poisson distribution is not.

I suggest you review your class notes or textbook for a suitable continuous distribution.

3. Thanks. I have tried using both exponential and normal distribution as they are both continuous but I am still hitting snags in my calculations. Not sure where I am going wrong.

4. Originally Posted by Marko_02
Thanks. I have tried using both exponential and normal distribution as they are both continuous but I am still hitting snags in my calculations. Not sure where I am going wrong.

5. Hard to do all this statistical data is being put into excel

6. Originally Posted by Marko_02
Hard to do all this statistical data is being put into excel
Weibull Distribution

7. For the purposes I need I am relitavely sure it is not Wiebell. I am thinking along the lines of exponential distribution but I feel as if I dont have enough data to work with. Any help please?

8. Originally Posted by Marko_02
A light bulb is warranted to last for 3,000 hours.
a) What probability distribution would you recommend to model the time to failure?
b) If the true mean of time to failure is 2,750, what is the probability that the bulb will last at least 3,000 hours?

I'm thinking its poisson distribution that I have to use but I did the formula and got a wack answer. Any help would be appreciated.
Under the assumption that the probability of failure in a unit interval of time is independently of life so far we use the exponential distribution for remaining time to failure.

This ignores "infant mortality" and "wear out"

CB