While the other modules so far have been terrific with good levels of support and clear explanations, this module is pretty terrible for a few reasons.

1) The level of support.

Your chances of getting a response to any question are slim - which means you're pretty much on your own here. Don't understand anything? Go find the answer elsewhere.

2) The tutors.

Mine Çetinkaya-Rundel has generally been terrific so far. Speaks slowly, repeats what variours terms mean (instead of assuming we memorize them the moment we hear them) and provides good clear examples to work from.

Sadly both Merlise and David are the opposite. They whiz through the material uncomfortably reading from a telepromter often assuming we instantly grasp every possible concept. It's almost impossible to follow most of the sessions they present. Most of the time there aren't even any exercises or opportunities to check we've understood the material correctly. They would both be 100% better if they frequently reminded us of the definitions of the concepts they use.

3) The material. There is FAR too much here to be covered in a single module. This is an entire course on its own (or a much bigger module).

4) Assumptions we know things which are never taught. I've lost track the number of times a word or concept sneaks into a quiz, into a lecture, or into an R package without explaining what it means. At times it feels this material was pulled from 2 or more sources and this has created gaps in understanding.

Sorry guys, I've really enjoyed the first three modules...but this one was a bit of a disaster.

Provide better support, shrink the material, create a better lecture experience and I'll happily revise this.

Unfortunately, for me, this course did not live up to my expectations in comparison to the previous 3 courses I took as part of the Statistics with R specialisation. I gave the previous 3 a full 5 stars each.

The problems I had with this course was not that my statistics knowledge was lacking or that I found it difficult. The problems were due to the robotic delivery of the material. Specially towards the end of the course. It is very clear that the instructors have a great depth of knowledge which is incompatible with the robotic delivery structure currently in place.

For example, if you use a particular technique, even if it was introduced earlier, all it takes is a brief 2-10 second statement to re-iterate. This encourages the delivery of the material to be a hybrid of both written text the instructors are reading from, as well as a more informal aura of discussion. A guideline is: 'Can you get someone off the street to read the material you wrote to the screen?'. The more this statement is false, the more amazing your course is.

Another issue I had was that the accompanying material was immense. Am I paying a subscription to read books and passages in order to understand the material? This point is also prevalent in the forums where it was raised multiple times. These books and supplementary material would be largely not required if simple commentary was in place in the videos.

E.g. We are applying a 'BIC' prior. What does this mean? Up to this point we are used to applying priors that are distributions. This means that we are approximating the posterior probabilities of the models using the maximisation of their log-likelihoods which turns out to be easier to calculate than the posterior distributions. However, as the model space grows (>25 parameters), we may need to rely on a sampling technique, these techniques which rely on posterior probabilities to traverse the model space. If I were to say this, it would take me 10 seconds but would provide so much information to the learner.

In summary, I could have read a lot of the presented material here from a text book and found it clearer which wasn't the case in the previous 3 courses. The resources were helpful and focused on interesting points. I loved the interviews at the end with experts in the field. The statsR package is great and this is a great way to showcase its capabilities. This course is OK but I think the delivery could be improved upon.

Thanks

Starts out good in the first week and then ramps up to graduate level statistics without really a lot of notation explanation. Week 3 with the silver haired lady as the teacher was the WORST. nothing made sense when she taught.

Worst course in the specialization. Totally killed my interest in statistics and R. Warning to everyone, do not do this course if you have / want to learn statistics. Only do it if you want to re-enforce the view that statistics is not something for you.

Very large drop in quality from the previous three courses in the specialization. Unlike the previous classes, there is not a quality textbook provided. What passes for a "textbook" is essentially a written re-hashing of the lectures which provides no new examples. The lectures themselves can be hard to follow and often times skip over important calculations. There are no practice problems. The labs are also less clear and there are major leaps between what is taught and what we are expected to know to solve the problems. The lack of quality resources and poor teaching coupled with the more challenging material made this course very frustrating.

Good for reviewing Bayesian Statistic. But not for new learners.

The quality is below the previous courses in the same Specialization. The contents are rushed. The labs are impractical and sometimes confusing.

And beware of the final assignment. Since the number of students is low, the grading takes lots of days. And you might miss the enrollment window for the Capstone course.

While the rest of the courses in the "Statistics with R specialization" from Duke University were good beginner courses, intended for the rest of us, this course was a huge disappointment.

The so-called textbook is a hurriedly written set of bad lecture notes, all written in R markdown and automatically converted to HTML. On a normal 1920x1080 monitor, some of the math formulas in the later chapters are not visible, you either need a wider monitor, or must know enough HTML/CSS to use chromes "inspect" to edit the css so you can change the margins of the main content to "overflow: visible;" to see all of it.

The textbook/lecture notes and the video lectures use exactly the same content, so if you didn't understand one of them, don't bother to look at the other.

Concepts are not introduced in any logical order, and many concepts are hardly explained at all, instead they just throw a wall of jargon at you, possibly expecting you to just google it.

The amount of jargon is staggering, instead of focusing on a few basic ideas, graduate level concepts are constantly thrown around without any explanation (and remember that the basics are never covered).

The whole idea behind the course is confused, they say they will teach you Bayesian statistics without requiring calculus, yet the lectures and textbook are full of integrals. But even if you know calculus, don't expect to understand the math. The formulas introduced are not rigorous, but vague simplifications where you must already understand the content in order to understand why they can get away with the sloppiness.

In the video lectures, nobody explains any formulas or algorithms in detail. Instead the formulas (or in some cases, just the name of the formula or algorithm) are displayed in the corner while the professor continues reading the teleprompter without even stopping to breathe in.

And there is no flow to it. Reading the textbook or listening to the video lectures is like meeting a giant wall of words. No headings. No theorems or important formulas or algorithms highlighted. No dissection of things that are hard to understand. No exercises to make sure you understand the content. Instead important theorems are hidden in long examples. It was just page after page of heavy jargon without any logical structure. I found myself constantly wondering if they were still talking about the same thing, or had introduced yet another concept they didn't bother to explain.

I completed this course because I wanted to complete the specialization. But if you are curious about Bayesian statistics, and want to learn it, look elsewhere. This is not the course you are looking for.

The last two weeks are way too hard to follow and could provide more practical examples instead of focusing on mathematical theory and formulas. That would make more sense to this course when compared to the content of the previous ones in this specialization.

Like many other reviewers, I was caught out by the dramatic shift in approach and presentation style at week 3 of this class. In one of the in-class forum threads on this topic I found a perhaps sarcastic recommendation from one of the mentors to take a different class (UC Santa Cruz) which I did in the middle of this one. I learned more from that one. The in-video demonstrations do not always explain the numbers that show up on the screen, and there is much less direct connection to using R than in the previous classes in the specialization. In the final programming projects, the Bayesian magic is hidden behind packages so you don't actually work directly with the computations...just a different function call. This was the least rewarding class of the specialization, and I won't bother to continue to the capstone project because of it.

I have been doing the specialization from course 1, giving 5 stars to all of them. This course had the poorest explanations in the videos. I like the content, as it is challenging but many questions come to mind which are not thoroughly discussed in the video. For example, when you do the exercises you get topics that raises questions and there is no way to clear these doubts. Sometimes in the forums, but it shouldn't be the case.

much different than the other classes in the series. They tried to put to much into this short course and consequently its way too hard. They should drop if from the statistics specialization and produce on new and longer stand alone Bayesian class. dissappointed because I dont think I can finish this class and now I wont be able to finish the specialization.

Clearly, Professor Clyde doesn't know how to teach.

Rushed exposition of extremely complex ideas. The level of this course is not at all consistent with that of the previous courses in the series. There is really no excuse for this; it's possible to provide an intuitive understanding of the Bayesian approach that is on par with the other courses (e.g. along the lines of Stone's "Bayes Rule: A Tutorial Introduction". This course has damaged my opinion of Duke's online offerings and Coursera more generally.

Although I truly liked the first three courses in this specialization, I cannot say the same for this one. Especially, Week 3 and Week 4 are very very WEIRD. Instructors throw you new concepts and terminologies in each slide non-stop. I guess they expect you to have a doctorate in statistics beforehand and to be here for brainstorming about Bayesian statistics. Certainly, this is the WORST course in the specialization. I have finished and got the certificate but I learned absolutely nothing. I don't think it is possible to gain any knowledge in this way. I know this as I already completed the first three courses. This one was definitely a WASTE OF TIME. I am sorry for the good efforts that Mine Çetinkaya has put for the first three courses. Other instructors ruin it in this one. You can view a lot of reviews like this below. I don't understand why they don't fix this course and why they leave it as it is.

A bit of a letdown after I loved the first three courses. It's a useful and interesting course. However, what I loved about the previous courses was that they were self-contained and if sometimes challenging I always managed to do the quizzes with the knowlegde and practice from the lectures and labs. With this one, I literally had to guess my way through some of the quizzes which is both frustrating and not very educational.

Do not take this if you have never done any upper level stat courses in university. Do not take this if you do not want to learn by reading 300 pages pdfs. Do not take this if you are here for the certificate.

The experience of this course is like jumping off a cliff. The first week is fantastic. Thanks to Mine Çetinkaya-Rundel , the videos are engaging and lively. She makes every concepts interesting and makes me want to learn more.

The second week is ok. I know David Banks is trying his best to make the videos good, but he often goes through the concepts too quick and not explain them deep enough. The tiny questions he sets up during the videos help me a lot on understanding the concepts.

Things become terrible after the first two weeks. Merlise A Clyde is in charge for the videos for week 3 and 4. I know she is great researcher but she is not a great instructor. She goes over the material crazily fast. We are only expected to read one chapter of the pdf per week before week 3. But we have to read 3 chapters per week since week 3. Yes, we have to. Because the videos are not helpful at all. She tries to squeeze all the contains in some videos by sacrifcing lots of detail, and the videos are full of conclusions from no where. If you want to understand what is actually going on, you have to read the pdfs. But sadly, for some key proofs, you need to have some upper level stat knowledges to understand them. And the pdfs are not well written, you can catch typoes in nearly every chapter.

The worst is the last week, the final project. The project requires peer-reviews. You need to be reviewed and review others, but there is literally nobody here to review this. I submitted my project 10+days before the course end, and I got my grade back at the last day. I post and reply threads on the discussion form to try to find somebody to review mine, but there is NOBODY. I guess they all quit at week 3. While I review projects from other fellow, I have seen people submitted maps of somewhere, photo of a dancer, and some R instructions from other course. There is someone submitted the week 4 lab file as the final project. Clearly nobody runs this course anymore.

I would suggest that you split this course in three components, mirroring the frequentist courses of the same specialization: introduction, inference and regression. That would make the material much more digestible, because today, it feels quite compressed and many things are left unexplained (specially the last two weeks of the course, I spent as much time there as with the rest of the specialization altogether!).

Kudos to the overall Specialization; I am not going to complete the last course due logistics, but I truly enjoyed. Specially the first three courses given by Profr. Mine Çetinkaya Rundel, she is a true educator; mastering not only the subject (Statistics) but also the art of teaching it. Her efforts towards quality statistical education are highly appreciated, at least by myself. Please keep on your great work!

I think it was way too much subject matter to fit into the 5 weeks. In previous courses, I could watch lectures and take notes in a reasonable amount of time. This course, 7-minute lectures would take me 45 minutes to listen and take notes. They just went way too fast through the material, even talking much faster. Splitting this into 2 sections would have been much better.

Bottom line: if you're thinking about taking this course, don't. Go take the UC Santa Cruz Bayesian Statistics specialization instead. This course starts off ok, but the further you get, it becomes clear that this course is riddled with content errors and has no one around to fix them. If you think you'll give it a chance because the other Duke Statistics courses are good, don't - I cannot emphasize enough how much worse this one is than the rest of them.

First off, the (relatively) good things: this is actually a pretty challenging class that will make you think. It's got more of a masters-level feel than the other Duke stats courses, in that you're given a basic overview of material and then will need to go figure out the details on your own. I thought the R labs were actually quite good and helpful at explaining material without being very difficult on the pure programming side. I felt like the professors did a decent job at first of explaining how the Bayesian mindset differs from the frequentist (though that totally falls apart by the end), and how to use that to your advantage. Yes, the lectures mention integrals, but you never have to compute one. Literally all you need to know about calculus is that an integral is the area under a curve (more on that later though).

Unfortunately, there are a lot of issues with this course. The foremost is that it was obviously abandoned years ago. While the other Duke stats courses have a good number of students and seem to be actively supported, it seems like almost no one is taking this class. That's probably because it was removed from the certificate program. That's a problem when you have a peer-graded final project that must be completed for a certificate (unlike the other Duke stats classes, where you can just skip the project). It's also an issue because there are basic mistakes and contradictions in the answer keys - about 1 on each quiz - that haven't been corrected despite being reported over 5 years ago.

The quizzes are not necessarily overly difficult, but there are a lot of poorly written questions. For instance, quiz 2 has questions where you're asked to tell whether a new (posterior) mean is 0.6 or 0.7 based on prior function graphs alone. Another question has answers that directly contradict the method used in a textbook example. I can see why others are frustrated, because the quality control just isn't there for this course.

The last thing is that this course would honestly be better if it dove into the calculus rather than avoided it. You can get a surface level understanding without it, but it won't be anything more than looking at a formula and plugging in values. To actually get an appreciation for why Bayesian statistics work the way they do, you need calculus. Would that put it beyond the reach of some people? Yes, probably. And, to be fair, there are resources that go into the math more available. But a lot of depth and clarity is lost by trying to explain concepts without being able to say why they work.

Great course. Difficult to apprehend sometimes as the Frequentist paradigm is learned first but once you get it, it is really amazing to see the believe update in action with data.

Very good introduction to Bayesian Statistics. Very interactive with Labs in Rmarkdown. Definitely requires thinking and a good math/analytic background is helpful.

This is my first course on bayesian statistics, I really like it, it was step by step, and helps to clarify lots of concepts of frequentist statistic.

This is a great course but challenging. Some modules are not that clear, so it is important to have other resources to study. I recommend this book to assist you over the course: https://xcelab.net/rm/statistical-rethinking/.

Nonetheless, I am now a fan of Bayesian statistics!!

The course has seen a lot of improvement with new study materials and videos. I'd say that this is now much better than what the course was previously.

Weeks 1 and 2 were perfect. Weeks 3 and 4 felt rushed. The course suddenly became very formula heavy.