Confusion of the Inverse??

At JOTWELL, Omri Ben-Shahar has a review of a forthcoming article in the Stanford Law Review claiming to have shown in a study that consumers are cowed by a consumer contract's fine print even if they believe they have been defrauded by the seller - i.e., have been expressed guaranteed A and learn later that (i) they aren't getting A, and (ii) the fine print says they have no legal right to A. (The reviewed piece is Meirav Furth-Matzin & Roseanna Sommers, Consumer Psychology and the Problem of Fine Print Fraud, 72 Stan. L. Rev ___ (2020)).

I've been blogging with outtakes from the not-quite-ready-for-prime time Unsure at Any Speed . Here the outtake intersects with another subject on which I have gotten involved recently: how to deal with the spread of detailed and unread consumer contract fine print, particularly given the ease by which it can appear to be made binding via internet click-throughs.

The question is not whether the conclusions Furth-Matzin and Sommers draw from their laboratory experiments are correct.  First, I don't know enough about qualitative research methods to assess their hypotheticals and questions to test subjects. Second, from what I can tell, they have given enough detail about the methodology to allow the tests to be repeated and therefore falsified. So I accept them for what they seem to say: people seem to take the fine print seriously even when they know they have gotten screwed.

My question is rather about empirical statements that underlie the study to begin with. Is it the case that widespread non-readership of fine print leaves consumers open to exploitation by unscrupulous firms? Is it true that sellers can outright lie about their products and services and then contradict the lie in the fine print?  The Stanford article takes the answer "yes" to those questions as a given, and then proceeds to assess the impact of fine print, given that there was fraud.  I cannot find, however, at least in the footnotes on the first six pages of the article anything other than a couple of anecdotes in support of the proposition that unscrupulous firms are a widespread problem.  I'm not saying they aren't; I just don't see any evidence one way or the other.

Is this an example of "confusion of the inverse," the subject of my outtake?

What I mean by "confusion of the inverse"

I cut from Unsure a detailed explanation of the "confusion of the inverse." It is, along with things like availability heuristic, the law of small numbers, hindsight bias, and confirmation bias, an example of the predictable divergences from actual probabilities to which Kahneman, Tversky, and others demonstrated humans are prone. My particular heuristic/bias peeve has to do with academic assumptions about the morality and competence of corporate oversight (Caremark doctrine for you governance nerds), exacerbated perhaps when, my having recently been been a corporate executive, a colleague blithely characterized corporate executives as "turnips" at a workshop shortly after I joined the faculty.

Here is the confusion of the inverse applied to my peeve.  Conditional probability is the quantification of the following question: given the probability that A is true (P(A)), what is the probability of B given A (P(B/A))?  The formula for deriving the answer is:

P(B/A) = [P(A/B) x P(A)]/P(B)

What we are trying to derive is the probability that we have a corrupt/incompetent board given that we have observed material corporate wrongdoing.

The probability of MW among the set of all corporations is P(A).

The probability of MW given CIB is P(A/B).

The probability of CIB is P(B).  Note that you can have a CIB even if you don't have MW, and you can have MW even if you don't have CIB.

Our formula now looks like this: P(CIB/MW) = [P(MW/CIB) x P(MW)]/P(CIB)

So...

Let's assume the following.  It turns out MW among all corporations is very rare.  Say P(MW) = .01 (one in a hundred).

The probability of material wrongdoing, however, is very high, IF you have a corrupt/incompetent board.  Say P(MW/CIB) = .95

The formula gives us the following numerator:  .95 (the probability of MW given that we have a CIB) x .10 (the probability we have MW).

But remember you can have a CIB even if you don't have MW, and you can have MW even if you don't have CIB.  So the denominator P (CIB) has to take all possibilities into account.

Hence, P(CIB) = [the probability that there is MW given CIB times the probability of MW] plus [the probability that there is MW with no CIB times the probability of no CIB].

So... P(CIB/MW) = (.95 x .01) /[(.95 x .01) + (.05 x .99)]

P(CIB/MW) = .16

So given that you observe material wrongdoing, the probability of also encountering a corrupt or incompetent board P(CIB/MW) is .16.  The confusion of the inverse is to believe P(CIB/MW) is .95.  It is not to say that you can't have corrupt or incompetent boards. It is to say instead that it is wrong to assume board members are turnips just because you observed material wrongdoing.

There are even more malignant examples of the confusion of the inverse.  When a police officer pulls over a car, what is the probability that there are drugs in the car, given that the driver is African-American?  When TSA does a search, what is the probability that the individual is a terrorist, given that he/she appears to be Middle Eastern?  When you are tested for a rare disease, what is the probability you have it, given that the test is positive?

Confusion of the inverse and contract fine print issues

As I said, I express no view on the study in the Stanford Law Review article.  I just don't see any evidence about the prevalence of out-and-out fraud. My intuition is there is probably less of it than the article seems to suggest.

That isn't to say there aren't real fairness issues with fine print. I have engaged with Rob Kar on his Harvard Law Review article with Margaret Radin, the thesis of which is to ground an attack on over-reaching boilerplate on a demarcation of the "true" agreement between the contract drafter and the consumer by way of Grice's "conversational maxims" and an actual shared meaning.  (Theirs is Pseudo-Contract and Shared Meaning Analysis; my response, just published in the Australasian Journal of Legal Philosophy (Vol. 43, pp. 90-105) is Conversation, Cooperation, or Convention? A Response to Kar and Radin.)

What I take from the Stanford Law Review study is that consumers aren't completely led down the primrose path by the fact of "fine print" - they expect there to be terms and conditions even if they don't read them.  The study seems to bear that out, even in the extreme where the consumer really does believe he/she/they got screwed. The real question is to what extent should the fine print be binding.  I agree with Omri that disclosure is not likely to be helpful - oy, more fine print disclaiming the fine print. Nor do I think trying to find the actual agreement or shared meaning is going to be fruitful.  Rather, there is a convention about what is and is not fair, and that probably ought to be reflected in regulation.