Complete the following steps if your data indicate the exact time that each item failed or was retired from service.
- In Variables/Start variables, enter the column or columns that contain the exact failure times or retirement times. Enter one column per sample. Minitab assumes that all the data within each column are from identical processes and provides a pooled growth curve estimate for the data within each column.
- If you have frequency data for each variable, in Freq. columns (optional), enter the column (or columns) of positive integers that indicate the number of items for each failure time. Enter one column per sample.
- If the failure data in a column is from multiple systems, in System ID (optional), enter a column or columns to identify each system. Enter one column per sample. Minitab pools the data to provide a single growth curve analysis, and tests for equal shapes or scales or mean time between failures across the systems.
- If all the samples are stacked in one column, but belong to separate groups, select By variable and enter a column of grouping indicators. Minitab provides a separate growth curve analysis for each group, and tests for equal shape or scale or mean time between failures across the groups.
In the following worksheet, the Time column contains exact failure/retirement times. The System column (optional) identifies the system for each failure time. Because the data are both time-truncated and failure-truncated, the Retirement column indicates whether the data in each row is a failure time or a retirement time. By default, the lower value in the column (0) indicates a retirement time and the higher value (1) indicates a failure time. For example, system 1 failed at 1 hour, failed at 5 hours, and was retired at 9 hours.
| C1 |
C2 |
C3 |
| Time |
System |
Retirement |
| 1 |
1 |
1 |
| 5 |
1 |
1 |
| 9 |
1 |
0 |
| 4 |
2 |
1 |
| 7 |
2 |
1 |
| 10 |
2 |
0 |
| 8 |
3 |
1 |
| 9 |
3 |
1 |
| 11 |
3 |
0 |
Note
With exact data, multiple failures are not possible at any one instant for a given system. Therefore, the optional Frequency column is usually not used. However, if a rounding error occurred during data collection, you can use a frequency column to indicate that more than one failure is associated with a specific time value, due to rounding.
