Determining a confidence interval for data.

[NewtonianAlch]

Homework Statement

I have 20 data points of the surface roughness of oil pipes, the mean, and the std of the sample.

I need to get a 95% confidence interval of the mean surface roughness of these pipes. I did a boxplot, and found that it's negatively skewed, and therefore not normally distributed:

Would using the t-table be OK? I'm usually used to doing this for data when it is normally distributed, I'm not too sure what to do when it's not normal.

This is what I did:

x ± t-val(s/√n)

Where the t-val is t19 and I looked up the t-table for the value. s = std and n = 20 and x is of course the mean.

 

[hedipaldi]

when the sample size is greater than 30 you can use the t-confidence interval even if the population is not normal.

 

[phiiota]

several of my professors have said n=30 is a number that's been around for a while and is used as a rule of thumb, but you don't have to follow it super strictly. How skewed is the data? If it's not horribly skewed, if it's even vaguely bell-shaped, I'd imagine n=20 is acceptable. But please correct me if I'm wrong, I'm still new at this.

 

[hedipaldi]

I suppose n=20 is reasonable for using t-intervals