Opiniones y comentarios de aprendices correspondientes a Inferential Statistics por parte de Universidad Duke

An overview of inference, light on the math, light on the theory, and with an unfortunate failure to reinforce what may be the most important part of practice: what should be done when conditions for a particular method are not met. When you teach students how to evaluate the conditions required for certain methods, but then walk through those methods even when the conditions aren't met, you reinforce poor practice. If you want to use an example where the conditions aren't met, STOP once you find out the conditions aren't met. STOP and REINFORCE the fact that you cannot use a method without meeting conditions. It is not a valuable exercise to walk through the plug and chug calculations anyway. STOP, discuss why you can't proceed, and then move on to another example if you want to give your students an opportunity to practice taking the method through to its conclusion.

The course did teach statistics but there were some problems with R commands, assignment expectations and grading outcomes. For one, this course really needs to do a better job of emphasizing that the student is expected to use R commands provided by Course 1 (I think it's called "Exploring Data") as well as this one. It would be helpful if students had easy access to all of the labs in that course as well as this one. Secondly, the list of expectations given out for the final project omitted several requirements (apparently we are expected to use R commands learned) though this is a small problem.

The biggest problem is the peer-grading. Reviewing other peers was actually very helpful in that it does teach you about your mistakes. For example, I realized after grading others that I had made two major errors and was ready to redo my project if needed or, if I passed, never make those types of errors again. This feeling that you are learning and that this is a quality course is taken away when your grade doesn't match your work. In my case, I got a perfect score which is nonsense. Like I said I made two big mistakes one of which was not including a confidence interval when I should have. What if I made others that I didn't see?

Feedback is important. It is a big part of learning. So is the ability to actually use the skills being taught (the R commands were taught much less clearly than say in Datacamp which this course suggests we use even though it is NOT free and I signed up for the Duke course specifically because it taught statistics using R). It's a good course overall and you will learn statistics; but when you're charging people a fairly high monthly fee, you should deliver on your promises to give feedback and to effectively teach one of the major course goals.

A very good introduction to the fundamentals of inference and NHST. It's very important that you do all excercises and readings or you will not learn as much. Also, the course won't provide a lot of information on how to use R, but if you spend a good amount of time on your project and make sure that it's good you will learn enough. I had to review a lot of R projects that were not very good, which suggests that some students aren't learning what they should.

If you want to learn statistics or have limited knowledge on the topic, and also want to learn a bit about how to use R, take the course. If you already know statistics and you only want to learn R, then this might not be the course for you, as the emphasis is on statistics per se.

Solid 3 stars. Lots of material is covered quickly and I learned a lot. The lectures are informative and supplement the book (I definitely recommend reading the (free) book). On the negative side, I noticed that the online discussion forum for the course isn't monitored by the instructors and the mentors seem to respond to only some questions. I noticed that almost all of the questions posted by students in the past year that went unanswered. I mean no one even bothered to respond to them at all. That's shameful, esp. if those students who submitted questions are paying for a certificate.

At times it feels lazy how it is put together. The examples are confusing (rather than clarifying) and there is close to no teaching of R, but the assignments are meant to be done on it. In fact in the forums it is endorsed by mentors to learn R somewhere else. Likewise, I saw one comment where the student mentioned how they got confused by a core concept (p-value) and could finally wrap their head around it by watching a Khan Academy video. And sadly, this was also endorsed by a mentor. Overall, I found the effort put into this course insufficient for people who are new to Statistics or R. Therefore, the name of the entire specialization becomes misleading as it suggests that we were going to be taught how to use R in statistics. I had high hopes for this course but sadly I will abandon it and spend my money on an alternative course/specialization.

The teacher lacks the ability of mathematical description, including clearing defining concepts, describing everything in mathematical languages, and showing math formulas of t-tests. She hopes to hide everything behind the canvas and just show how statistics are applied. But without enough mathematics nothing she said makes sense. I have to search on the internet to get to know what she didn't teach.

------from a math phd.

I think this course needs more instruction with the R studio lab

Not suitable for beginners

This course by Professor Çetinkaya-Rundel is awesome because it is taught in a very clear and vivid way. Lab section and forum are so dope that I love them so much! Definitely strong recommendation!!!

This course is an excellent overview of inferential statistic tests / hypothesis tests and confidence intervals. The organization and material is quite good, with exercises and applications using R.

No tutor answering questions in the discussion platform.

The course is very useful and helps me understand the formal testing process of data analysis. I just hope it would cover more of non-parametric testing techniques and dive into a bit more into effect size testing. Anyway, It also provides a lot of insights into important statistical measures of information, which could potentially be extended to the field of predictive modeling and machine learning.

Great course. If you put in a little effort, you will come out with a lot of new knowledge. I recommend using the book after you have seen the movies. It gives a deeper picture of how it works. Great!

G

r

e

a

t

C

o

u

r

s

e

.

I

l

e

a

r

n

e

d

a

l

o

t

.

While I understand and appreciate that the scope of the class is more focused on the application / ideas of the statistical methods without delving too much into the mathematics, I would appreciate if some of it was used to explain why some equations work the way that they do. For example, in talking about F test statistics, it was difficult to understand the reasoning behind F = variance between groups / variance within groups until I had to look up other explanations elsewhere. While I believe that the instructor teaches well in most parts, I often find it difficult to follow along because she goes through a lot of assumptions and I'm unclear as how / why she is allowed to go on her assumptions.

It has skills I need to learn for work using R programming. The instructor for me is hard to follow and squeezes way too much into verbally explaining things in such detail before you comprehend what she has said while at the same time presenting information on the slide incrementally. This leads to me not being able to choose to listen to her or ignore the slide. Maybe you are better at multitasking in real time. With that said you can tell she puts lots of energy into the video and makes things relevant. If you take these courses READ FIRST then watch the videos and be ready to click the pause button so you don't miss a concept that vanishes to the next slide in a few seconds.

The instructor for this course is EXCELLENT. She explains every topic completely, thoroughly, and with logical examples. She is fantastic at this topic!

I rated the course low though because of the incorporation of R. I reviewed every lab in detail in attempting to create the final project, and most of what I needed was never covered. Also, R is constantly changing so the coding in the examples no longer works and it's almost impossible to find a course that is up to date on all codes (for example, the select() function does not work, nor does %>% in the downloaded R or R Studio. It's beyond frustrating. The two-hour final project took 14 hours. Be careful - if you do not know R, this course is extremely difficult to do well in - one can understand the inference portion perfectly based on the excellent instruction, but R is very difficult.

So much disjointed information.... I felt absolutely crushed trying to learn and understand all this. Am waiting for another 8 hours before I can reattempt the quiz.

Personally, I feel that this course assumes the student is automatically an expert in statistics (simply due to completing the first intro to statistics course). The logical progression of how to approach different problems - and the terminology of the statements involved has been thrown out the window...

If you're new to statistics, I suggest you should at least double the time allocation they provided...

This is my second course of the specialization. Took this course as I thought it would be beneficial for my work as a research assistant. I have taken basic statistics for life science before back as an undergraduate and did not have any programming background prior to taking the courses from Duke.

So far, I am able to follow through the weeks. Would say that it's rather intense compared to the first course (Introduction to Probability and Data with R) in terms of content. The course is rather insightful and really enjoyed it overall despite the painful hours spent after work. The lecturer is really clear in her explanation. Would really recommend everyone who takes this course with little prior knowledge on stats to do the readings and exercises in the resource book on top of the video lectures. This helps to clear doubts on our side and enables the concepts to be reinforced and sink in our head better.

As for the lab sessions. I would say a lot of googling and figuring things out on my side was required, especially when doing the projects. Sometimes the forum does help. However, I would say the involvement of the mentors for this course is slightly lower than that in the previous course. That said, most of the issues encountered when doing the 'hands on' in R can be found in google.

I would definitely recommend this course to anyone who is interested to learn data analysis with R.

An excellent and rigorous that covers theoretical and simulation approaches to inference. The teaching is first-rate! The textbook and lecture examples are superb. The final project gives you practice in finding a research question that interests you, translating that question into hypotheses, and then challenges you to find the right method to test the hypothesis. The only improvement I could suggest is more examples and exercises on simulation approaches. I spent a great deal of time on my own learning about that topic.

Awesome. I loved the way this course is done. I know what Test Statistic to use for what type of data and under which conditions. I am preparing a cheat-sheet that will be shared with all later on.

Well-taught, but they need to provide more resources to help people learn R. R is not a user-friendly app and I needed to google how to do a lot of the things they're asking us to do. Needless to say, I can google how to work in R on my own without paying Coursera a fee.

The course factually do not teach r. It's much longer than it was declared. The core statistical course is quite good, but information about computer simulation methods looks like was entered retroactively

some topics require more explanation!

It was good. But I feel like I've spent half of the time untangling sly phrasing of questions.