Complex systems often contain many people. The SHEL model can be used to organize information (and guide fact-finding) by considering that people interact with several people, each of them may interact with several more people. Accordingly, in complex systems, the SHEL model can be represented with multiple people:

In a practical sense, the number of possible SHEL components affecting each person involved in a marine casualty or pollution incident is very large. If a marine investigator were to describe each possible element of the marine transportation system in detail and then to comprehensively rule out every possible mismatch, the process of conducting fact-finding would be exhausting. Accordingly, the marine investigator should consider the four principal components in the general sense, and should inquire exhaustively into only those areas that are immediately relevant to the incident. Questions about a person's marital status, for instance, are not appropriate when a person's performance is either a) tangential to the incident; or b) not in question (i.e., they did nothing wrong). Generally, information about the other conditions (type of equipment in use, weather and sea state, vessel characteristics, names of people involved and their roles) should be collected first. The section below provides guidance on what order to acquire certain topics.

　

The temptation during any interview, and particularly during interviews of a cooperative witness, is to presume that the witness is entirely forthcoming and accurate. Marine investigators should remember that extreme care must be used in establishing a rapport with the witness. Part of building this rapport involves gradually easing into discussion of those subjects most uncomfortable for the witness. Witnesses generally feel most comfortable giving factual information (as opposed to opinion information) about matters seemingly external to themselves. Accordingly, marine investigators will begin the interview by obtaining descriptions of the SHEL components involved. Such descriptions might include:

　

Once the outlines of the SHEL model for that person have been established, the marine investigator will proceed "around the diagram" once again, this time focusing on the interactions. Such questions are slightly more uncomfortable, but build on previous discussion. For example, these questions might include:

Finally, the marine investigator will proceed into the most personally uncomfortable aspects of the interview: the factors affecting the witness themselves. Recognizing that people may act to protect their own self-image and/or to hide matters from the investigation, marine investigators should carefully corroborate information provided by a witness pertaining to themselves. Experience indicates that the majority of this information about a person is derived from interviews with other people. For example questions might include:

Once these topics have been adequately discussed, the marine investigator should then turn his or her attention to the mismatches in their interaction. These mismatches are Latent Unsafe Conditions (LUCs) that will be used in the Causal Analysis.