Layman’s Abstract for Quality and Reliability Engineering International article: Criteria for the prioritization of hypotheses in root cause analysis

Each week, we publish layman’s abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
 
The article featured today is from Quality and Reliability Engineering International with the full article now available to read here.
 
Barsalou, MCriteria for the prioritization of hypotheses in root cause analysisQual Reliab Eng Int20221– 11https://doi.org/10.1002/qre.3224
 

Explanatory hypotheses are formed and evaluated in root cause analysis. However, prior to investigation, the hypotheses must be prioritized. Often, methods such as nominal group technique, multi-voting, and simple voting are used to decide which to investigate first. This research was performed to provide concrete criteria for the prioritization of hypotheses using three levels of prioritization based on the strength of the available evidence.  

Quality professionals were asked to rate various scenarios as confirmed, strong, moderate, and weak evidence towards supporting a hypothesis. There were thirteen scenarios presented and only two did not have statistically significant results. 

Based on the research, criterial for three levels of prioritization have been established as well as criterial for declaring an identified cause to be the root cause. Weak evidence consists of one weak study or experiment with confounding or ambiguity present and a single data point or observation. Moderate evidence costs of a single data point or observation verified through experimentation, a representative sample from the same batch, and related previous experience. Strong evidence consists of one study or experiment under simulated conditions, a representative sample from the same batch statistically analyzed, multiple observations, expert opinion, and multiple studies or experiments under simulated conditions.  

Both representative sample from a different batch and representative sample from a different batch statistically analyzed were not statistically significant. A representative sample for the same batch is rated moderate; therefore, a recommendation is to consider a representative sample from a different batch to be weak, which is one level lower than a representative sample from the same batch. A representative sample from the same batch statistically analyzed is strong evidence, so a representative sample from a different batch statistically analyzed should be considered one level lower, which would mean moderate evidence.  

The criteria are for the prioritization of hypotheses to investigate and not for declaring a hypothesized root cause to be the actual root cause. Criteria for declaiming a hypothesis to be the root cause was found to be the ability to turn the problem on and off by recreating the problem and then removing the source of the problem to make the problem go away.

More Details