Learner Reviews & Feedback for Statistical Inference by Johns Hopkins University

The course is not meant for beginners, but seems to be advertised as such. Knowledge of Elementary Statistics is a must. The course is fast-paced and most people would not be able to finish it in 4 weeks or understand all the concepts in the course without outside help. Use of Discussion Forums and Mentors such as Leonard Greski is invaluable for completing the course successfully. There are several minor flaws in the videos and textbook that need to be addressed. This course would be much better off broken into two (Elementary + Inferential Statistics) and buffered with longer videos and step-by-step instruction and help.

There is no doubt that topics covered in the course are fundamental and critical. However, instructors rushed through most topics and explain them in a not very intuitive way. To be fair, the "power" section in the course is actually pretty organized. Swirl exercises can improve your understanding to some degree, make sure to take them.

This is definitely not the 1st course if you are completely new to statistics. I'd suggest taking other similar Coursera courses or reading articles from different resources (e.g. http://www.sthda.com/english/) to better internalize all concepts.

A topic such as statistical inference is not complicated, and could be taught in a much more straight forward and comprehendible fashion. Just look at the tons of material and (good old fashion books) that relate this material in a much more concise manner. Moreover, the material in this class including the R-files are not well synchronized (gives low quality impression). A lot of time is needed to sort out the documentation between R-files, the book (Statistical Inference for Data Science) and the slides. I find many errors and sometimes inconsistent notation.

I have mixed feelings about this class. We are rushed through the concepts, I had to study a lot on my own to deeply understand the mechanisms - lucky for me I studied advanced mathematics in College, I mostly had to revive my memories. The video lectures are of very average quality, but the practical exercises in swirl helped a lot. Just the one final project is not enough imho. I can't judge yet if what I learned is enough to properly apprehend the algorithms at play in ML - we'll see. All in all, not the best course so far from the specialization.

This course covers the very important things about statistics, I totally agree with that. But I find that if Coursera can make the entire course easier to understand for the layman, it will be the best. After I took the course, I need to visit youtube to do some researches to understand the more complex stuffs like power t test. Maybe coursera should look at Khan Acedemy and see if they can get some idea from it.

I usually go to https://www.youtube.com/watch?v=uhxtUt_-GyM&list=PL1328115D3D8A2566 to look for those chapters that I need to revise.

This course goes on a very fast pace and simply does not have the charm of all the other courses in the specialization. I understand that a lot of content is covered within a month, but there should be supplementary course material available. Moreover, TAs should be more active on the forums. I have seen most of the questions just being discussed among the students. A little disappointed. Will probably have to watch all the material again to have confidence with it.

Course topics is good and heavily dive into statistical training.

I may say that there is a lot of theoretical stuff and these need to be supported by real world simple examples.

I have spent twice the time to watch the youtube videos about the classes to settle my mind and see some examples.

Course content need to revised and realistic easy to understand content including R coding should be included.

Thanks for the effort spent so far.

I believe the theory part can be greatly improved to provide an understanding. Practical and all is good enough as someone who likes maths, I would like to see more of it in the theory classes. I did not quite understand some topics intutively for which I had to search for other materials, but swirl excercises are a big help actually. Also a big thank you to the professors/mentors who put their time and effort in this .

It was more difficult than I expected. Besides to imagine inside your head some of the theoretical concepts. Instead of "accept or reject", we have "reject" and "fail to reject".... just as an example :) And now there is this discussion about p-values omg....

https://www.vox.com/science-and-health/2017/7/31/16021654/p-values-statistical-significance-redefine-0005?imm_mid=0f55ac&cmp=em-data-na-na-newsltr_20170809

I think it could be taught a tad better. Maybe more explanations in lessons and a bit better (read: less vaguely) worded course project description would be useful. I do understand the importance of this part of the DS specialization, but I would have loved a bit more careful approach to the subject. It is probably hard for an expert to lower himself to this admittedly low level of knowledge, but please do try.

Great R material, powerpoints, and lesson materials, l however the material is extremely fast-paced. Recommend a page dedicated entirely to R formulas. This was a good refresher, but anyone who has not had at lease Stats 3 or 4 in college/graduate school should think twice before doing this program or at least have many outside hours dedicated to completing this program.

Caffo clearly knows his stuff. But some of the lectures start off going slow but then take a leap forward into a conceptual realm that is beyond most people if they are not at least somewhat familiar with statistical concepts. Take your time with this one and make sure to do the reading. The videos kind of cut off prematurely sometimes.

Since it is a very theoretical subject, trying to explain it without proofs and plenty of background is hard. But i feel like most of the course is just to memorize formulas without much explanation where they come from. A few examples are computing the expectation and mean of the average distribution and computing confidence intervals.

Not my favorite course in this specialization. Very poor at developing "intuition" regarding statistical inference concepts. At many times I felt that the instructor was simply reading formulas, assuming that we already had the background. I had to go to YouTube to get real-world explanations of what different concepts meant.

The course covers very important topics pretty well. The instructors knows the subject, materials are well chosen. However, the lectures could be done much better. There are many typos, the instructor is reading from the slights. Isn't it worth putting a little more effort since this course is taken by the thouthands of students?

The course is good, but not very challenging. Anybody having done any course in statistic would have little to no information from the first two weeks. Only week 4 was interesting to me, dealing with boostrapping.

The teacher is very clear and chooses the subject in a clever way. One always understands what he or she is doing.

Brian Caffo is an interesting lecturer - he dives into the key concepts and ideas that are essential to understanding the statistical concepts necessary to gain a better appreciation of the course. However, presentation and materials need a LOT of work. They can be too overwhelming and most of the times feel irrelevant.

This course should be targeted for Data Scientists, in my opinion it is more for statisticians.

Too much about the insight of statistics and some but not enough about how to use the statistic tools.

Some time the professor seems like he is just reading the slides which I think it doesn't intensive the student.

So in my opinion information that ws given in this course was not exlained well, lucky for me I was just refreshing these things, so I knew where I could lok for additional explanation. But if you don't have any background in this scope, it would be rather hard to understand why given formulas are working.

The course contains a lot I want to learn, but as someone with a limited background in statistics - I found many of the lectures not to provide clear explanations for concepts. I had to use a lot of outside material to try to learn and understand the concepts. The course lectures seem incomplete to me.

The class contents are good I guess. But I don't think the professor knew how to teach and enjoyed the teaching process. Based on my experience, all the concepts are not that hard for everyone if they would be explained in a good way. I finished this course only because I want to do the course capstone.

A good class, but I think there are some missing pieces. For example, there was a lecture on the basics of knitr, but nothing related to creating a pdf from R. In the Regression Models class there is a lecture on basic notation. I think it would have been more helpful to have that lecture in this class.

I really learned a lot in this course, but I find that I got most out of the lectures in week 3/4 when Brian actually stopped reading the slides out loud and explained the concepts i his own words. I believe the course could be improved by taking that approach in the other weeks

The lectures are really hard to understand, while the material itself is really not that hard. The lecturer talks as if he is just reminding us everything we've already learned. Had to go to other MOOC (specifically Khan Academy) to obtain proper understanding of the topic.

The course was very dry compared to the other courses I have taken. Though there was a lot to cover in the four weeks but this was not best way to do it. The course covers a lot of concepts in far too little a time span. It should have been spread into at-least two modules.