8.3 Reliability Data
Analysis
From a reliability assessment viewpoint, failure data are used
to:
(1) 

Determine the underlying probability distribution of
time to failure and estimate its parameters (if not already
known)

(2) 

Determine a point estimate of a specific
reliability parameter, e.g., MTBF

(3) 

Determine a confidence interval that is
believed to contain the true value of the
parameter 
Two methods are used to analyze failure data:
(1) 

Graphical methods

(2) 

Statistical
analysis 
In many practical cases, graphical methods are simple to apply
and produce adequate results for estimating the underlying distribution. They
are virtually always a useful preliminary to more detailed statistical
analysis. The two methods will be discussed in more detail in the following
subsections.