Import the following text data file into JMP
(Instructions for importing a text file into JMP)
The JMP data table should look as follows:
Assume that these data are hours to failure obtained in a life test of a piece of electronic equipment. The piece of electronic equipment has a part from one or another of two suppliers: Supplier (1000, 2) and Supplier (5000, 2).
First, we do an exponential plots of the hours to failure ignoring supplier identity:
As we can see there is sharp decrease in the slope which is the hazard rate. It looks like the hazard rate was constant up to 2000 hours and then decreased.
Decreasing hazard rates are frequently encountered when two populations are "mixed."
We now analyze the data taking into account supplier identity. In particular, we do two separate Exponential and Weibull plots:
We see that two Weibull distributions fit the data quiet well. In fact the data for supplier (1000, 2) was generated from a Weibull (1000, 2) and the data for supplier (5000, 2) was generated from a Weibull (5000, 2).