Type of Censoring
Use the right-censoring options if you test each unit until a preset amount of time
elapses (time-censored) or until a preset proportion of failures occur
(failure-censored).
Use interval censoring when you plan to inspect units for failures at pre-set
intervals. You can space these intervals equally in time or in log time, or you can
set intervals so that the expected number of failures in each interval is the same.
For more information on right-censored data, go to Data censoring.
The options allow entry of different values for groups. For example, if the test will
observe some units for 10,000 hours and other units for 20,000 hours.
- No censoring
- Select to plan a test that runs until all units failure.
- A single right censoring time for all units
- Select to test each unit for a preset amount of time, then enter one or more
censoring times. Minitab displays a test plan for each censoring value that
you enter.
- Multiple right censoring times, one for each group of units
- If your data are multiply censored (the test units run for different amounts
of time), the analysis requires at least one column in the worksheet.
- From Allocation of units to groups, select whether to allocate the units to equal size groups or to specify
the percentages of units in each group.
- Equal percent per group
- In the worksheet, create at least one column with the censoring
times for each group. In Censoring times in column, enter one or more columns. If you enter multiple columns,
then Minitab displays a test plan for each column. For more
information, go to Data censoring and click "When
should I use time censoring for test plans?".
In this worksheet, Plan
A contains the censoring times
for 3 groups. Plan
B contains the censoring times
for 2 groups
| C1 |
C2 |
| Plan
A |
Plan B |
| 2500 |
5000 |
| 5000 |
7500 |
| 7500 |
|
- Specified percent per group
- In the worksheet, create at least two columns. One column has
the censoring times for each group. The other column specifies
the percent of the units to allocate to the censoring time in
the same row. Specify percents of units to allocate so that the
sum of the column is 100%.
- In Censoring times in column, enter one or more columns. If you enter multiple columns,
then Minitab displays a test plan for each column. In Percents of units in column, enter the columns of percents so that the order of the two
sets of columns is the same.
In this worksheet, Plan
A contains the censoring times
for 3 groups. Percent A contains the percents of units
to assign to each group in the test plan.
Plan B
contains the censoring times for 2 groups.
Percent B
contains the percents to assign to each group. To create
test plans for these 4 columns, enter "Plan
A" "Plan B" in Censoring times in column. Enter "Percent A" "Percent B"
in Percents of units in column.
| C1 |
C2 |
C3 |
C4 |
| Plan
A |
Percent A |
Plan B |
Percent B |
| 2500 |
50 |
5000 |
50 |
| 5000 |
40 |
7500 |
50 |
| 7500 |
10 |
|
|
- A single right censoring percent of failures for all units
- Select to test each unit until a preset percent of failures occur. In
Censoring percent of failures, enter one or more percents of failures. Minitab displays a test plan for
each percent that you enter.
- Multiple right censoring percents of failures, one for each group of units
- To plan that test units are in groups that run until different percents of
failure, the analysis requires at least one column in the worksheet. Each
row contains a percent of failures between 0 and 100 for a group of units.
- From Allocation of units to groups, select whether to allocate the units to equal size groups or to specify
the percentages of units in each group.
- Equal percent per group
- In the worksheet, create at least one column with the percent of
units to test until failure in each group. In Censoring percents of failures in column, enter one or more columns. If you enter multiple columns,
then Minitab displays a test plan for each column. For more
information, go to Data censoring and click "When
should I use failure censoring for test plans?".
In this worksheet, Plan
A contains the percent of units
to test until failure for 3 groups.
Plan B
contains the percents for 2 groups
| C1 |
C2 |
| Plan
A |
Plan B |
| 75 |
90 |
| 50 |
50 |
| 10 |
|
- Specified percent per group
- In the worksheet, create at least two columns. One column has
the percent of units to test until failure for each group. The
other column specifies the percent of the units to allocate to
the percent of failures in the same row. Specify percents of
units to allocate so that the sum of a column is 100%.
- In Censoring percents of failures in column, enter one or more columns. If you enter multiple columns,
then Minitab displays a test plan for each column. In Percents of units in column, enter the columns of percents so that the order of the two
sets of columns is the same.
In this worksheet, Plan
A contains the percent of units
to test until failure for 3 groups. Percent A contains
the percents of units to assign to each group in the
test plan. Plan
B contains the percents of
failures for 2 groups. Percent
B contains the percents to
assign to each group. To create test plans for these 4
columns, enter "Plan A" and
"Plan B" in Censoring percents of failures in column. Enter "Percent A" and
"Percent B" in Percents of units in column.
| C1 |
C2 |
C3 |
C4 |
| Plan
A |
Percent A |
Plan B |
Percent B |
| 75 |
50 |
90 |
50 |
| 50 |
40 |
50 |
50 |
| 10 |
10 |
|
|
- Interval censoring
- In Number of inspections, enter how often to check the units for failures during the test. Then,
select how to plan the inspection times.
- Equally spaced
- Select to equally space the inspection times. This option is
usually convenient from an operational standpoint because the
inspection schedule is easy to remember.
- In Last inspection time, enter the time of the last inspection. The last inspection
usually coincides with the end of the test.
- Equal probability
- Select to make the expected proportion of failures the same in
each interval. This option is usually convenient to limit the
amount of work done on failed units at each inspection
time.
- In Total percent of failures, enter the expected percent of failures for the entire test.
For example, if the test runs until all the units fail, enter
100.
- Equally spaced in natural log times
- Select to equally space the inspections on a natural log scale.
Equal space on the natural log scale usually produces
inspections that are closer together at the beginning of the
test, then become more spread out as the test proceeds. This
characteristic often pairs well with failures from a lognormal
distribution or from a Weibull distribution with a decreasing
hazard function. (A Weibull distribution with a shape parameter
that is less than 1 has a decreasing hazard function.) For such
distributions, multiple early inspections offer increased
resolution for early failures, but the inspections still cover
failure times that vary by mutiplicative factors.
- In First inspection time, enter the time of the first inspection. In Last inspection time, enter the time of the last inspection. The times to enter
are in the natural units rather than the natural log units.
