[R] - Test to say if I can accept or not my data in a right-skewed distribution

#1
Hi, I'm new in Statistics world and I have a problem I don't know how to deal with.
Sorry in advance for my English, I'll try to be as clear as possible.

I have a dataset of 9000 machines that need to have a specific repair. For each observation I have the repair time used.
I know from the maintenance manual that the theoretical repair time should be 5 hours for this kind of procedure but I have many repair times different from this value, some are very bigger and other are smaller than 5. Repair time has a right-skewed distribution.
I thought to have a statistical method (a test?) that can tell me if a repair time is acceptable or not (I do not know if it makes sense saying if "it is statistically significant").
I do not know if it is necessary consider a "tolerance range", if yes I could say 2-6 hours.

Is there anyone who can tell me if it is possible and how? I use R software.

Thank you so much for your time! :)
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
I typically like to initially visualize my data with a histogram. @Miner will tell us about relevant stats.

Welcome to the forum and your English was very clear!
 

obh

Well-Known Member
#3
How many people/groups do this fix?
internal or external to your organization?

You may find from your fix data the 99% percentile (or 99), meaning in 99% of the fixes the time is less than this value. but in this case, on average 50 times (0.01*5000), the fixed time will be larger.

You may test the average, to see it is not significantly larger than the 5 hours?
But there is also the question of the 5 hours, is it really correct number based on experience...?