The reliability of an educated guess on the cause of a failure or accident

I thought to remind readers that some initial hypotheses by experts, educated guesses, on the cause of accidents or failures are very reliable – and others are less so. And there’s every gradation in between. This sometimes causes problems for client and expert alike.

A client decides to take a case or commit to settlement of an insurance claim and all’s good – additional forensic investigation supports the initial hypothesis.

But often enough the hypothesis must be modified a little or a lot – as per the scientific method – and on occasion – rarely – rejected completely. (Ref. 1) This can be difficult.

There’s also the hope – expert’s feel it as a subtle, unintended pressure – that the expert’s additional investigation will support the initial hypothesis. One colleague commented on this several years ago.

One solution is not to ask for or give an initial hypothesis. I’ve experienced that in the past, not often but on occasion – complete the forensic investigation then give us the findings.

Another might be to assign reliability to a hypothesis on a scale of 1 to 10 so a client has a feel for what s/he is getting. Or more simply: Is it probable or only possible? Maybe give the possibilities.

Still another solution, state clearly that in forensic work an initial hypothesis is an educated guess based on the available evidence. And additional investigation may change things as per the scientific method that is followed in forensic work. (Ref. 1) Evidence that an expert gets from:

  1. The client’s briefing on the problem – the accident or failure
  2. A document review (text, pictures, video)
  3. A visual site assessment – in person or virtually
  4. The expert’s past experience


All this came to mind when I saw a large barn on the verge of collapse during a drive in the New Brunswick countryside a few days ago. (Ref. 2) The ridge of the roof sagged an estimated 10 feet from where I was driving. It was something to see.

If an expert was retained to assess the cause of a collapsed structure and saw pictures of a structure like this then a very reliable initial hypothesis would jump off the page.


My thoughts about the reliability of hypotheses were reinforced when a party in Ottawa contacted me a few days ago about the cause of a leaning retaining wall. He sent me 23 pictures of the wall. (Ref. 3)

I studied them for an hour and the cause gradually presented based on:

  1. The nature of the cracks in the wall
  2. Water stains on the face of the wall, and their location
  3. What was not seen on the face of the wall
  4. Wall foundation construction as seen at one location
  5. The location of the lean with respect to a feature behind the wall, and
  6. Past experience with this type of retaining wall construction

On a scale of 1 to 10 – where 10 is like the reliability of a 10 foot sag in a barn roof – I would give the reliability of my assessment of the cause of the lean an 8.

To be really sure, I’ve asked for low level aerial photography of the site with a camera mounted on a kid’s drone – a few dollars of expense for what I’m certain will be clinching evidence.


Following are my assessments of the reliability of an educated guess, on a scale of 1 to 10, of the cause of accidents and failures I investigated in the past:

1. I’ve investigated the cause of slip, trip and fall accidents different times. If I had been asked about a slip and fall on one particular occasion, I would have given my assessment of cause a 9.

2. A trip and fall accident would have got an 8.

3. The cause of a nail gun accident would also have got an 8.

4. Sagging, sloping floors in a high rise would have got a 4 maybe a 5.

5. The cause of large cracks in the wall of a three story building would get a solid 6. I didn’t even see this wall, not even virtually. I was just told the type of wall and the size of the cracks – not hairline, which was significant – and, right away, I knew their configuration and cause.

6. The cause of the John Morris Rankine fatal motor vehicle accident would have got a 10, if I had been asked. No question. I drove the test vehicle during my re-enactment of the accident. I’m here to tell the tale which, if I could tell you more, would say a lot about the reliability of an initial hypothesis of cause in this case. Three dimensional video – video from three directions, including early drone-like video from above – was the clincher.


And so it goes, the up and down reliability of initial hypotheses, and the importance of realizing this. The expert must form an initial hypothesis, an educated guess, but, if possible, keep your distance from it till s/he is well into the forensic investigation and has a good understanding of where it’s heading.


  1. Google the scientific method and read all about it, in fairly jargon-free language
  2. Update: You could be excused for thinking that everything is falling down. Posted October 13, 2020
  3. Morry, C. J., Personal communication, October, 2020


(Eric E. Jorden, M.Sc., P.Eng. Consulting Professional Engineer, Forensic Engineer, Geotechnology Ltd., Halifax, Nova Scotia, Canada

Comments are closed.