Are you a Bayesian? If so, why? If not, why not?
Sure…or else I’d keep sticking my tongue to the flagpole…
Do I get to wear a little hat?
That was my 800th post. I feel it summed up my whole being.
awwwwww, hon. Yer lil’ hat is mighty cute. Sets them Blackeyes off right nice now, it does.
I think I am. I’ll let you know after I update my priors.
Not really. But I’m not a statistician, so I don’t really have to care.
Not true, actually.
Only yes or no? Shouldn’t it be yes/ no/ sometimes? Or yes/ no/ if the data are unambiguous/ if the number of alternative states is known?
I am not a Bayesian. I renounced my Bayesian citizenship when I immigrated to this country in 1992 to escape religious oppression and the bloody Kurdo-Bayesian civil war.
That dictionary definition leaves me none the wiser, so I will repeat my initial (spoken) reaction:
Wut?
If you’ve not heard of Bayes, you’re proabably not a Bayesian.
Well, I was a Bayesian, but my membership was revoked and my Member Card confiscated when I broke the rules by using the clubhouse as campaign headquarters during my unsuccessful bid for Supreme Global Ruler.
So now I’ve got nothing. Betcha they couldn’t have kicked me out if I’d WON that election.
Yup. I’m a Bayesian. Its logical and more intuitive, once you understand the concepts. I was introduced to Bayesian statistics by a very good lecturer in my undergraduate days, who made me realise that with normal statistics, our assumptions are based on what we’d like to see. With Baysian methods, this bias is non-existant, as our priors reflect the belief we initially have in each model, and can change based on the data we are presented with, so that an initially less probable model can become more probable because of the presented data.
In short, it gives hope to the underdog! Yay!
I would tell, but I think it would take all the mystery out of our relationship.
Susan
If you pushed, I’d say yes. But I don’t feel as passionately about it as some do.
I should say I am! Nora Bayes is one of my favorite vaudeville singers!
I’m more of a non-linearist. Say, nice shoes!
While it is a more formal theory of statistical inference, one that is in continuing development (I think), it also has to do with how one how and why one comes about her beliefs. And by beliefs I don’t mean silly stuff like religion or palm reading.
Since you’re writing from the U.K., I’ll assume you’ve heard of Terry Pratchett from whom I’ll steal an example. In one of his books he mentions “autocondimentors”, i.e. people who automatically add salt to their food without tasting it first. In some circles here in the States that is considered a real no-no. If you autocondiment in front of the wrong person, you’ll be on the receiving end of an apocryphal story about the American tycoon Howard Hughes. The story goes that some poor slob was being interview for employment with Hughes and that the interview was taking place over a meal. The interviewee had the audacity to salt his food without tasting it first and Hughes immediately dismissed the gentleman as a candidate for the job. The lesson? Taste your food, I guess.
But it seems that autocondimenting is a rational strategy: if food comes to your plate insufficiently salted enough times, then it makes sense to infer that the next time it would be sensible to salt it automatically. If you find that your food tends to be too salty after autocondimenting, then you’ll be less likely to assume to salt it without tasting. That, if I understand it correctly, is a sort of Bayesian way to go about the world.
Suppose you are pulling out from a minor street to a major one, and there is a car approaching with its blinker on. Do you assume that the car will turn, in which case you can pull out without waiting for it, or do you wait for it to slow and begin its turn before you decide that its blinker is “sincere”? Do you take first impressions seriously, or do you blow them off and wait for further evidence of a person’s character? Do you assume that a salesperson is lying to you? Do you tend to believe things person X tells you and mistrust things person Y tells you? Why do you look forward to going to pub A but not pub B? It seems to me that life is full of little beliefs that can be subject to an internal and unconcsious bayesian analysis.
My very poorly written OP didn’t reflect what I was really asking (perhaps because I didn’t know myself until you post prompted me to think about it). It seems sensible that so many little beliefs are subject to bayesian analysis. But is that the only way? Do people reason out their little beliefs other ways, if so how and why? Is there a general rule, or do they use ad hoc rules for each situation: the blinker is on, the person turned it on, therefore she intends to turn; or, this salesperson is probably working on comission, so I’ll assume she’s lying; or, I like my food more salty than average, therefore it makes sense to autocondiment?
While the last post helps explain why you’ve put this in IMHO, I suspect this might go better in Great Debates: people who know enough about the terminology are likely to have an opinion that’s not quite humble.
That said, merely on the statistical issue, it depends … On most philosophical issues, I delude myself that I’m a bit of a fox and a promiscuous one at that. I’ve also read a reasonable amount about the history and philosophy of statistics without reaching many very firm conclusions, but feel able to apply insights from much of both. Pragmatically, I’ve worked in environments that are both heavily non-Baysian and heavily Baysian, and have adjusted to both. I’ve both read Fisher and paid homage at the very grave of Bayes.
Until I ever feel the need to write “bonzer’s Big Guide to Statistics”, don’t fence me in.