Best free way to teach myself statistics?
![[Post New]](/jforum/templates/default/images/icon_minipost_new.gif){kind=icon}
Anonymous
|
I never took stats at any level but I am in a new position where I would really benefit from learning more about statistics (currently reading a report and having trouble understanding all of the models and such)
How should I approach this? Library books? Khan academy? Free online courses (there seem to be some?) Would appreciate any suggestions |
Anonymous
|
I would suggest you buy a used statistics book that has solved answers to problems within it or in an accompanying study guide. Try online resources but also read the book in chapter order in case the online material doesn't do enough to help you understand.
If you find out an expensive current book "9th edition" you can probably find the just as good "8th edition" for only a few dollars. Statistics isn't a discipline that changes the basics so books only update for fashion and business reasons. Statistics 101 is the same everywhere and doesn't get you very far. You mentioned reading about "models". Can you explain what research field and what is being modeled? Also what kind of statistical tests they are citing? You should expect to drill yourself on problem sets if you want to internalize the material. It's like high school algebra. If you don't work problems, you won't internalize the formulas. Actually now that I think about it, the high school AP Statistics curriculum might be easy for you to investigate and get matching printed resources for. I'm sure you can find AP Statistics online resources because many kids like to use online resources to supplement their class. Believe it or not, statistics.com is a reputable provider of college-level live online courses. I have considered getting a certificate through them in the past. They've been around for years. If your work covers training benefits that might work out for you. Or a community college class. Again, Statistics 101 is extremely uniform everywhere. What will vary most is teaching quality. |
Anonymous
| Khan Academy and a textbook from the library |
Anonymous
Thank you for this. This article is the type I am referencing - I can understand it in general but if I am being honest, I am lost at really understanding the details. https://journals.sagepub.com/doi/10.1177/0014402919868486#:~:text=Unadjusted%20analyses%20of%20large%20samples,minority%20overrepresentation%20in%20special%20education. |
Anonymous
| Sorry, realized that link may not be all that helpful. But for example currently looking at a table labeled “logistics regression (odds ratio coefficient) models of disability status….. and I have very little understanding of how to interpret the data. |
Anonymous
| A community college course would be almost free. |
Anonymous
|
Your article seems to be behind a paywall, but I think I can imagine the type of table/analysis it contains. In my MPH program there was a stats course for people who were going to do statistical analysis, and a stats course for people who didn’t need to know how to run the analysis but needed to be able to interpret what someone else’s analysis means. I am the second kind of public health practitioner!
I’d suggest starting with a Khan Academy or Coursera course. See how far that gets you, and whether you need more. It should enable you to at least read the tables and know what you are looking for. My coursework was now almost 20 years ago. I find that I need to re-teach myself things every few years. It can be complicated! |
Anonymous
| Stats were a blur to me until I took a logic course.Knowing logic clarifies a lot of formerly confusing things. |
Anonymous
|
I learned how to read those tables in graduate school, in regression, causal inference (that's cause-al, not casual!) and econometrics. Those courses came AFTER I had already taken 1) undergrad stats, 2) grad stats, 3) economic modeling with calc. So that's a lot to study on your own. And I won't lie, it's difficult stuff to grasp.
Rather than go to a textbook (they will have more theory and detail than you need) I recommend you start with a study guide like "Statistics for Dummies" or "Head First Statistics" and the like. That will get you the foundation you need to move on to multivariate regression and modeling. (Those subjects will not be covered in the books named above, but you can't jump straight to them without understanding the basics, trust me.) regression and modeling also involve calculus, but if you took algebra and pre-calc in you can probably pick up what you need. |
Anonymous
| I like Mathispower4u better than Khan Academy. I haven’t watched the statistics section but have watched the others a lot to brush up on topics myself. https://mathispower4u.com/statistics.php |
Anonymous
|
MIT has some free offerings: https://ocw.mit.edu/search/?t=Probability%20and%20Statistics |
Anonymous
| I second the suggestion to take a class at a community college. They all offer online classes these days and you’d get a teacher with a masters’ or PhD willing to help you during office hours if you need extra help. |
Anonymous
Hi again. I'm PP with the long answer above. It helps to know that you're looking at education and social research. Let me change up my advice a little now that I know you're not looking at hard science or medical research. I've had Stats 101 3 times in different flavors at good schools (undergrad and grad). I've never had Stats 102 because people kept telling me their 101 was different/better/deeper. I got As in all 3 classes. As it turns out, they were all similar classes. I have a feeling, given the field you're looking at, you could be fine with only Stats 101 and 102 level knowledge but you'd still see (not practical to fill) gaps in understanding when a researcher busts out particularly fancy techniques. And self-studying 3-6 credits of material is a lot if you just want to be a more confident reader of social science/education research. I would recommend you shop for a book that really works for the level you are at right now. Start with the library and get something on the "Dummies" level. If that's too easy, then shop on Amazon for a second used textbook with reviews that speak to you. It definitely could take more than one college class to start to understand the specifics of the statistical methods. So, I would say to focus more on critically analyzing the study itself. Are the sample sizes large, what factors are being analyzed/accounted for and are there any factors that are left out, what are the names of the statistical tests, etc. Read the abstract, intro, and conclusion several times before going through the main paper. Then...google for simple explanations of the terms you are curious about. Start with "regression analysis". Watch a video about what that means. Then google "odds ratio" statistics regression. There are some good examples right away. (I didn't remember anything about this from my coursework.) Controversially, you don't need to understand the data table per se, you need to understand what groups are being compared and why did they use the statistical technique more than you need to confirm the statistical strength of the effect they uncovered. In part that's because in your field, you should most likely be looking for corroborating evidence across multiple studies. I think you may actually do better backing into an understanding by googling and watching highly-rated videos. Having a comprehensive printed text with index could still help you because digital search does not give you a sense of progression like a coursebook does (101 material is usually the first half of the book, more complicated 102 stuff is in the 2nd half). E.g., texts start with hypothesis testing of normal distributions and then move into other topics (chi-square, etc.). Also some people are better at googling than others. I haven't been through AP Statistics myself but my kid just completed it with reasonable success. That's why I think you want to teach yourself, that might be better than true college level material because of the target customer. You want to prioritize comprehension vs. developing the ability to "do" statistics. https://www.khanacademy.org/math/ap-statistics https://www.reddit.com/r/APStudents/comments/15l98ob/recommended_textbook_for_selfstudying_ap/ Good luck with furthering your knowledge! |
Anonymous
PP x 3 . This person above posted while I was writing mine. I think the comments are very helpful corroboration that you are looking at research using grad school level stats well beyond the 6 credit intro sequence. That to me further supports a backing into understanding as you read the articles you want to read vs. getting frustrated that weeks of course time haven't addressed topics you encounter (that's how I feel about being convinced/required to take three flavors of Stats 101 and never getting to 102). I do think everyone who reads published research should have Stats 101. It is a reasonable bare minimum for statistical literacy. It's like algebra. You may not use it a lot, but you definitely should have been exposed to it. So I commend you for looking into this and developing a plan for yourself. |
Anonymous
| There are some popular stats books out there like there are popular Econ books- you could start with something like that to get an idea for how stats can be applied and then move on to more academic materials. Tim Harford has written a couple. I studied Econ and stats because I find them so applicable to real life, not just theoretical. |