Statistics

Update On The On-line Probability & Statistics Course. Comments Wanted

I may wear a wig like Huygens, who waded into probability a long time ago.

It’s been a while, but I think I’ve finally figured out how to make an on-line class work. Details about this will be forthcoming, but for now I wanted to ask those who hanker after the kind of course you can imagine me offering how much they’re thinking of paying.

Write the amount (be generous: make it a big, but true, number) below in comments, email me at matt@wmbriggs.com, or use the Contact Me page.

Impossible to answer without knowing how the course would work, I know. So, very roughly: videos and write ups with relevant reading, watched and read at your pace (free for all; no support). Homework and homework support for those that want it (costs depending on level of support). And a final project for those who want the whole class experience (costs more). Again, all at your pace.

I have two initial classes in mind, but will probably do one similar to my usual summer course first. This is more introductory. See this syllabus (pdf). If you’re already a working statistician or upper-level or higher statistics student, then the pace of the course will probably be too slow for you.

This is not a mathematics course! Thousands of professors and books will teach you the math. I will teach you understanding and meaning. There is some computer work, but I have got people through it who have never even opened a spreadsheet before.

Since I’m also working on a New & Improved! version of textbook, this class will follow its development. Here’s the old one, version 1.2 (pdf).

(Besides it’s many other flaws, the old textbook had words about spreadsheets and R which are no longer necessary, given the many resources on-line.)

The other class is pure philosophy, with the practical aim of understanding the differences in classical (frequentist), Bayesian, and the logical methods of analysis. This one is more along the lines of a standard philosophy of science course but with an emphasis on uncertainty. Think of a cross between Howson and Urbach, David Stove (second half of this book), and E.T. Jaynes.

Of course, the practical class has large elements of philosophy, just as the philosophy course has elements of the practical. But, even though I reference Jaynes, the philosophy class is NOT a mathematics course.1 This class is meant for the working scientist or the statistician who has fallen into a rut.

A word on the videos, some of which you may have already seen. If it all works out, I’ll be able to fund reasonable equipment to replace the beginner’s kit I’m currently working with.

Reason I’m putting the word out now is that if folks are only willing to pay little or nothing, then I don’t think I’ll continue the scheme. There’s certainly incentive to pay nothing. On-lines statistics courses can already be had free—but they’re all standard, party-line, by-the-book offerings. If you follow this blog, you know this is not me. It’s all up to you guys about what it’s worth.

Legal something-or-other. Although I maintain a distant but cordial relationship with Cornell (I am only employed there two weeks a year), the things I do on this blog are entirely independent. The courses I offer, just like these blog posts, have NOTHING to do with that or with any university. Because of this, the only thing you’d get out of my courses is an education and not a credential.

———————————————————-

1Nothing in the world wrong with mathematics, and for a complete understanding of the field one has to have a healthy helping of mathematical knowledge. But the problem is in courses which concentrate on the proofs, the understanding is left far behind, or even eliminated. Reification enters and the equations become the meaning. This is what leads to the epidemic of over-certainty and flaming scientism we see everywhere.

Categories: Statistics

25 replies »

1. Mariner says:

I live in Brazil, which means that I make less money than most people here. I won’t give you my number because it would deflate your estimate. But I’ll probably make an effort to pay whatever it costs.

A suggestion: Give certificates anyway (for those who get the full course experience). They won’t have any legal weight, and they don’t need that. Who knows? Perhaps they’ll be worth more than your regular Coursera certificate in a few years. And people like that stuff. Look at Coursera for an example of a “certificate of achievement”.

2. Briggs says:

Mariner,

Thanks for the reminder. Yes, “Certificates of Completion” or the like are planned, as well as letters of reference. That sort of thing.

3. Matt,

This project is all about marketing.

With the proper marketing, people will pay money for nearly anything. Regardless of quality.

Without proper marketing, people won’t even take quality goods and services for free.

In marketing, the key is to figure out who your target market is. Once you’ve got your target market demographics/psychographics figured out, then you design your product/service and marketing pitch.

So before you can analyze the product/service, we’ve gotta know: Who is your target market?

Once you have potential students, you have the issue of where/how to host the sessions. There are some online training marketplaces which you could use to host/service/collect payments. Here’s one example: http://coggno.com/lms/tour

The key is still marketing. Identifying that segment that needs/wants your product/service, then letting them know about it, then getting them to act on your value proposition.

Good luck!

4. Ken says:

The course described is the type of course one takes if necessary, few out of curiosity. Thus, selections are typically based on a process of elimination: don’t like the teacher (whatever reason)–reject; don’t like the book, reject; etc. If price is competitive relative to comparable offerings, price becomes a non-factor. All of which is a windy way of saying ‘I don’t know.’

But,

What text will you use???

To a LOT of people the text can be a determining factor (at least, if the text is wrong they’ll opt out) right after rejecting instructors considered undesirable for whatever reasons.

Here’s a good reference right up your alley, and its the sort of reference most students with only a slight inclination to read would find interesting enough to read — it conveys your sentiments & outlook, but in different style (which is thus complementary & reinforcing) — and it is VERY highly rated on Amazon (i.e. if listed in the advertised curricula would be a factor for not rejecting the course offering & maybe a reason for giving the course more consideration):

“How to Lie With Statistics,” by Darrell Huff.

First published in 1954 it remains a relevant & well-written classic addressing the basics & logic over the math, and is rather short (144 pages per Amazon: http://www.amazon.com/How-Lie-Statistics-Darrell-Huff/dp/0393310728 ).

People by nature like to think they can “get away with something” even if that’s not something they can do (who doesn’t like to know how a magic trick was done?!). The book draws interest…and it’s focus is clearly on ‘how not to get duped by someone using stats to lie to you.’ People viscerally find that appealing.

5. Briggs says:

Ken,

See above for words about the text I will use.

Kent,

Thanks!

6. Bert Walker says:

\$150

7. Sheri says:

Ken: That was what my college statistics professor used many years ago! It’s still going strong–in fact, the title comes to mind often.

Briggs: Not sure. Maybe the \$150–I would have gone more toward \$100, depending on the course. Right now I am taking a course on climate change, so it would also depend on timing. Since I could work at my own pace, it should work. (Yes, I really need your unique perspective on statistics!)

8. Daniel Drumea says:

Hi,
I was thinking that \$50 per lesson, once or twice a month, is doable for me. It’s probably obviously my discretionary spending. I would suggest that you throw in the final presentation for the full package. Not very interested in the certificate but I need the mental tools for risk evaluation and mitigation in it security, and statistics applied to economics.
My offer to pitch the class on the UN Intranet in Geneva stands. Not great hopes though as the public servants know everything already or they can hire a consultant (or so I’ve heard).

9. Clayton says:

I like the idea of \$100 to \$150. At that price I would take it just to learn. BUT (there’s always a but, right?) I don’t see how you could grade homework for that price, and a project would have to be right out.

10. Briggs says:

Clayton,

You’re right, I couldn’t. Not individually. Idea is to have (password protected) homework pages with answers and explanations, a place (like any blog post, really) where students can chat and I can chime in on larger questions. But to talk to/answer emails of each person for even 15-20 minutes per homework over the entire course is a substantial amount of time.

11. Like Mariner I live in a country where money is a bit harder to come by but I would be willing to scrape up a hundred bucks and I think the course will be well worth it. Now, Dr. Briggs, all you need to do is find a thousand or so like minded people.

12. Andy says:

\$150-200 for assignments. Additional \$200 for project and presentation version.

I think there should be an upper limit to “your own pace” so as not to keep you, Briggs, on the hook for final presentations.

Any thoughts on using something like Google hangout for assignment recitations at regular intervals? Perhaps for an extra charge?

13. Dave says:

I actually read your textbook and think it’s excellent. But someone with my experience probably doesn’t need to take a class for that level.

However, I will again suggest that you build classes that get into advanced bayesian data analysis and modeling. There is no free substitute for that that I know of, so you could easily command a higher price (albeit for a smaller audience).

14. Necessity is the mother of invention, or in this case having a bias for action.

15. Ken says:

Briggs — If you include anything by D. Stove (or any philosopher a you will encourage prospects to reject the course offering. Nobody is going to reject the course because you excluded Stove, or any other philosopher.

Basically, by incorporating philosophy you’re effectively advertising your course, by whatever it’s title, is in the same genre as “physics for poets.”

CONSIDER your remark: “…the kind of course you can imagine me offering…”

NOW CONSIDER YOUR AUDIENCE: Who out there wants it? What need is to be filled?

IS this course about you, or, about arming your customers with practical pragmatic info they can apply in the real world — there’s a market for that.

There’s not much market in the liberal arts field, or pressing need anywhere for that matter, for philosopher statisticians. Who hires those, and for what??

16. Briggs says:

Ken,

I take it, then, that you’ve never had a philosophy course before? Get Howson & Urbach from the library and page through it for a better idea.

But you’re right that I need to show just why you should be interested. That is coming.

And as to your question, “who hires” philosophers of probability & statistics? Well, I’m still waiting to discover. Turns out people just don’t want to learn they are too sure of themselves.

17. Ken says:

Briggs, re: I take it, then, that youâ€™ve never had a philosophy course before?

OH yes I have–and there was zero difficulty [then & now] in seeing the mental gymnastics applied to ensure the desired conclusion was reached. This pattern is particularly apparent when one truly couldn’t possibly care less of where the analysis leads. Invariably, a philosopher will reject a conclusion if it doesn’t fit with their particular values–such conflicts are resolved by correcting the analysis to ensure a desired conclusion is “proven.” All it takes is torturing the data until it confesses.

Recent examples: Global warming alarmism fits the pattern — if one believes humanity is a blight on the Earth worthy of punishment, then data archives & models are “adjusted” accordingly. The rationale underlying an Italian courts conviction of certain scientists for failing to provide proper warning of the severity of the impending L’Aquila earthquake is another example.

18. Hi,

I would love to take the course and have homework with answers but I don’t need any intervention by you. With people like me if you do the work for one student you have done enough for a million students. You could even have a fun exit exam on-line someplace that we could take. The course could be a mooc in other words.

I would be willing to pay, but I could not go much over \$50 at this time and would like half that even more.

Also, what about selling the new version of the book on Amazon for 9.99 in electronic form? Most Bayesian statistic books are way more than that on Amazon.

And what about having your readership submit problems and solutions — help you in other words. (not me, I am still too tied to the old ways to be worth a tinkers damn)

This could develop into a large thing, and there could even be a “class option” for us teachers to us with students.

19. Walt S. says:

\$200.

20. I would be willing to pay \$300. I’ve taken on-line courses before, that cost about \$900 for a semester course. Of that, about \$150 went to the instructor, and the rest went to institutional overhead. Thus, \$300 would be good for both of us.

21. Depending up on the cost I would love to take your course, my last statistics class was in 1997.

22. Opps, meant 1987. (An edit function would be nice).

23. Francsois says:

Between 50 and 400 dollars, depending on what the product looks like.