ExposureBlog

I was reading a deposition awhile back in which the expert witness, a smart guy with a few letters behind his name, was asked if he thought something would happen.  His reply: “In all probability, it would.”  The attorney moved on to the next question.

He shouldn’t have. 

Instead, he should have asked Mr. Expert something about those words “in all probability.” Something like. . .”what exactly do you mean when you say those words?” 

I’ve read a lot of depositions and I see those words all the time, coming from people who want the jury to believe the attorney questioning him (or her) Just Doesn’t Get It and statistics (and thus, perhaps, The Almighty) is on their side.  Attorneys on the other hand, are often people who were once liberal arts majors and never, ever took a course in statistics.  Or if they did, it was probably the “statistics for liberal and fine arts majors” course where they learned that the correct term for “average” is “mean.”  Okay. Maybe that’s harsh.

But attorneys really should be on the lookout for buzz words that sound like harbingers of a Foregone Conclusion but in fact may be orphans—standing between the rails as the Statistics Train fades away into the distance.

I always tell my clients to listen for words that suggest throwaway conclusions but in fact are tied to fairly rigorous mathematical procedures.  Lazy experts (yeah, there are a few in this world) like to use them in place of real analysis.  When tagged, they fall back on their voluminous, multifaceted and most importantly, unverifiable experience.  “Based on my experience” these guys intone, “the probability for this to occur is high.”

Well, I’ve probably said something like that myself—but only after an eight-hour deposition and only after my blood sugar level was in the negative numbers.  Nowadays, I know that many probabilities *can* be quantified.  And if an expert is charging his client more than three figures an hour, he should be able to come up with a reasonable probability figure that at least vaguely resembles statistics.  And it should be correct.

If I get the chance to coach a client before deposing some of these experts, I usually recommend listening carefully for the following words: probability, chance and liklihood. All can be quantified, and if the expert is earning his or her keep, then that quantification has been made in terms of *confidence interval.*  In later posts on Safety Gig I’ll discuss how this is done.  But for now, here are the important questions attorneys should ask:

1. You said there is a high probability.  Is that probability greater than 50%?

2. Is it 90%? (remember that science generally requires the confidence interval to be in the 95 percentile.  This is the same as having a probability of 5 percent or less that the relationship is due to chance.)

3. (Pushing a calculator across the table) “Can you show the jury how you made that calculation? 

The last question is a bit unfair, because the actual calculation for probability usually (probably!) involves some pretty intensive mathematical grunt work involving such things as maximums and minimums, iterations, random numbers and such.  Fortunately, there are software packages that do this for the expert, notably Decisioneering’s *Crystal Ball* and Palisade’s *At Risk*. Both are easy to use and reasonably priced. 

Getting back to the questions.  While attorneys must understand the science of the case as well as the experts (some actually know it better than the experts!)—many hate to appear foolish, even in deposition.  Nothing—*nothing* is more fun for an expert than lecturing the opposing counsel on *how* to frame the question.  Yet, the attorney must remember that in this regard they are standing in for the jury—a bunch of decent, hardworking people who probably don’t have the time nor inclination to think much about statistics—yet, may have to use it to decide a case.

All the more reason to make that expert come up with a number—a number that can be defended, explained or attacked.  As I always tell my clients:  make ‘em *quantify* it!  If they cannot, then the proper response should really be: “you don’t know, do you?”

And the honest response under those circumstances, would be. . .

“well. . .no.”