Change Analysis


Change analysis is a Root Cause Analysis technique that uses the precise specification of a single deviation (problem or adverse event) so that changes and/or differences (potential causal factors) can be found by comparison to closely related un-deviated situations.

Pros and Cons


  • Conceptually simple, easy to grasp.
  • Works well in combination with other methods.
  • Results translate naturally into corrective action recommendations.
  • Can be used to find causes that are obscure, or that defy discovery using other methods.


  • Requires some basis for comparison.
  • Resource intensive, requires exhaustive characterization of deviation.
  • Applicable only to a single, specific deviation.
  • Provides only direct causes for a deviation.
  • Results may not be conclusive; testing usually required.


Change: A discrete difference between an occurrence exhibiting the deviation, and a similar occurrence that did not exhibit the deviation.

Deviation: A situation in which actual results or actual performance differed from what was expected.


As suggested by the name of the technique, change analysis is based on the concept that change (or difference) can lead to deviations in performance. This presupposes that a suitable basis for comparison exists. What is then required is to fully specify both the deviated and undeviated conditions, and then compare the two so that changes or differences can be identified. Any change identified in this process thus becomes a candidate cause of the overall deviation.

What is a suitable basis for comparison? There are basically three types of situations that can be used. First, if the deviation occurred during performance of some task or operation that has been performed before, then this past experience can be the basis. Second, if there is some other task or operation that is similar to the deviated situation, then that can be used. Finally, a detailed model or simulation of the task (including controlled event reconstruction) can be used, if feasible.

Once a suitable basis for comparison is identified, then the deviation can be specified. Various schemes exist for performing this specification. Perhaps the most useful scheme (attributed to Kepner and Tregoe) involves four dimensions (WHAT, WHERE, WHEN, and EXTENT) and two aspects (IS and IS NOT). Regardless of the scheme used, the end result should be a list of characteristics that fully describe the deviated condition.

Given the full specification of the deviated condition, it becomes possible to perform a detailed comparison with the selected undeviated condition. Each difference between the deviated and undeviated situations is marked for further investigation. In essence, each individual difference (or some combination of differences) is a potential cause of the overall deviation.

After the potential causes are found, each is reviewed to determine if it could reasonably lead to the deviation, and under what circumstances. The most likely causes are those that require the fewest additional conditions or assumptions. In this way, a large list of potential causes can be whittled down to a short list of likely causes. Finally, given the likely causes, the actual or true cause(s) must be identified. Generally speaking, the only way to verify which likely cause is the true cause is by testing.

The purpose of change analysis is thus to discover likely causes of a deviation through comparison with a non-deviated condition, and then to verify true causes by testing. True causes found using change analysis are usually direct causes of a single deviation; change analysis will not usually yield root causes. However, change analysis may at times be the only method that can find important, direct causes that are obscure or hidden. Success in change analysis depends ultimately on the precision used to specify a deviation, and in verification of true cause through testing.



Change is introduced in all factors of life continuously. Some sources of change are planned, as in deliberate actions taken to achieve a purpose. Other sources of change are unplanned, as in natural, random variation, or as in factors introduced unintentionally due to outside influences or as the result of error. Whatever the source, change is often a source of disruption in the normal, expected, or usual flow of events. When change is not accounted for or compensated, it can lead to deviations.

As discussed above, change analysis depends on the recognition of changes or differences that could have led to a specific deviation. Sometimes, however, multiple changes may have occurred over time that combine to cause the deviation. Therefore, it is important for the investigator to consider combinations of changes or differences as potential causes, in addition to individual changes or differences.


Change analysis is heavily dependent on comparison with similar situations. However, there are varying degrees of similarity, depending on how close the undeviated condition is to the deviation under investigation. The best case scenario for change analysis is when you have previous operational history for the exact same task or operation. In this case, changes or differences that could have contributed to the deviation are easily identifiable.

The problem with trying to compare situations that are less similar is that other, inherent differences in underlying conditions may mask differences that were responsible for the deviation. Since each difference identified in the change analysis procedure is considered a potential cause, the list of potential causes may include some of these inherent differences -- which may or may not bear any causal relation to the specific deviation under investigation.

It therefore is critical that an appropriate basis for comparison be selected when performing change analysis. Furthermore, inherent differences between the actual deviated condition and the situation chosen for comparison must be fully identified and handled with extreme care. Finally, when verifying true cause by testing, the test condition must be made as identical to the actual deviated condition as possible.

Web Resources

KT_Equip_Root_Cause.pdf, Presentation illustrating Kepner-Tregoe Analysis, by Stephen Davis to HPRCT, 2000.

Revision History

  • 00.2 25-Sep-2014 Minor content/formatting changes + link updates/replacements.
  • 00.1 12-Jul-2005 Minor rev - added link to US Coast Guard summary page.
  • 00.0 24-Jun-2005 Initial revision issued for public consumption/comment.

by Bill Wilson
Loading Quotes...

Site Map
Bill Wilson © 2004-2020

Last updated: November 17, 2014 at 20:08 pm

Leave a Reply

Your email address will not be published. Required fields are marked *