In regards to the Mary test, I initally thought it was about inductive inferences, but then when I caught wind of the last question, I began to have my doubts. But I went with my intuitions, and I passed with flying colors. :)

Heh, I just took the Mary test. I passed but I didn’t spot the error in quite the way perhaps might have been best. My reasoning was more like this: these questions are all ridiculous, there is no way to determine any of these results as being more or less likely to follow from the data presented. So I avoided the trap in 4, but not by specifically sussing out the difference between 2 and 4, but rather because all of the questions seemed to be roughly on equal ground in the sense that there was no reason for me to choose any one of them as being more or less likely than any of the others.

there is no way to determine any of these results as being more or less likely to follow from the data presented

Well for both sociological and psychological reasons, given the scenario presented, that just isn’t true.

But I’m glad you avoided the error!

Heh! I guess I meant within the context of those questions. One could do a study and find a pattern of association betweeen Mary’s attributes and likely paths, but I don’t have any of that sociological or psychological data at my fingertips. Or are you saying that those psychological and sociological reasons are right there in the test?

Or are you saying that those psychological and sociological reasons are right there in the test?

No, but the fact they are not there, doesn’t mean that they don’t exist. It just is the case that given x number of people with the characterisitcs of Mary that there are going to be systematic differences in the distribution of the various outcomes.

In other words, there are reasons for choosing x rather than y as an outcome even if you don’t know what they are! And I know from the results that a majority of people are aware of the reasons (otherwise it wouldn’t be the case that so many people - even amongst those interested in philosophy - get the thing wrong!).

Well I’m looking at those questions again and here is the interesting thing:

The text says

On the basis of what you know about Mary, judge how likely each statement is to be true in relation to the other statements (so, for example, if you think it is more likely that Mary is an executive for a housing charity than a bank clerk, you’d rate this as being more probable on the scale).

It is worth pointing out that there is no trickery going on here. The description of Mary’s life above is not in anyway misleading. Everything is at it appears!

So the questions say “probable” or “unlikely” but then stuck in the middle is “possible.” Basically the way I avoided the error was by refusing to commit to a more or less likely. Why overcommit when saying something is possible is sufficient to the task? So I guess my question is, do you need the “possible” in there? Or is a problem created if you remove it? “Possible” just means there is nothing logicaly or empirically standing in the way of it being the case. Whereas probable/improbable is asking for a different kind of valuation the kind of valuation the question asks people to make even though they can escape from that demand by just clicking possible.

I think you’re over-analysing a bit here. It’s just a 1 to 5 scale (what’s called ordinal data), where you’re ranking the probability of the various options against each other. I think that’s clear from the context, plus the statement about ranking “in relation to the other statements”.

But I’ll have a think about the point you’re making! :-)

question: what percentage of people who did not fail the test answered “possible” to all questions? Is that a report you can pull?

I can only tell you how many definitely didn’t. But actually the data wouldn’t tell you a lot at the moment because it defaults to that setup, so it’ll include the fairly large chunk of people who will have lost interest reading the instructions, and who just pressed the submit button.

I might default it to nothing selected, but then I’ll have to write in an error checking thing, which is a bit of a hassle.

I think nothing selected would be good, and would give you interesting results. But you’re a busy man :)

Jeremy,
Enjoyed the Mary test. Any more?
Read the article on microbiology. Knew some of the stuff but still very interesting. Something I found out recently about “mad cow disease,” it’s neither a viral nor bacterial disease. It’s a disease caused by malformed proteins (prions.) The human infection has a difficult name, easily forgotten. I’m not sure but that this is the only protein disease.

Faust - Yes, I think you’re right. I’ll set it up so that it defaults to nothing selected.

Ralph - Glad you enjoyed the test. There are no more at the moment, but I intend to create a huge selection of these kinds of tests, so I’ll keep people posted.

Creutzfeldt–Jakob disease - or transmissible spongiform encephalopathy - is the human version. It’s not the only prion disease, but very interesting. Actually, degenerative neurological disorders, generally speaking, are very interesting (and scary!). I was obsessed with ALS for a while: I know more about the disorder than any sane non-specialist should know!

Terrifying diseases. I can’t donate blood in the US because I was in Enlgand for more than a year between the 80s and the 90s or at least that’s what it said on the form.

Yeah. I was in the UK during that whole time, but I’ve not eaten meat for nearly thirty years, so it wasn’t really an issue for me. But there was quite a bit of worry about it for a while.