Tuesday
Jul202010
Jeremy Taylor | Tuesday, July 20, 2010 at 9:19PM What Makes Stats Scary or Confusing? I Want To Hear From You!
I want to write about and make video tutorials about the things that you want to hear about. Please respond to this post with a comment about what you think is scary or confusing about statistics!
Reader Comments (7)
excellent site dear sir! More videos please.
Thanks Dave! I'll do my best to keep the videos coming!
-Jeremy
So many exercises to complete, but no time. In the UK I sat an exam that required four analyses (including a factor analysis and an rmANOVA) to be completed in SPSS in less than two hours(!). In a Scandinavian country I have heard that students get to sit for 5 hours in a basic statistics course - although it was a handwritten exam. It feels unfair... but how do you get up to speed with statistics (without getting too stuck on checking whether the assumptions for a test hold). I've got roughly 400 exercises to complete for a stats course I'm taking. Although the course covers the basics, the work piling up in front of me seems overwhelming.
How can I get faster at stats?
I certainly understand your frustration. Unfortunately, there is no way around checking assumptions and the learning curve involved. In terms of speed, if you are using SPSS learning syntax is probably the best way to work faster and more efficiently.
SO many things make me fear stats. The biggest thing that I struggle with though, is what test to choose. I understand the need to test for basic assumptions, and I think I'll probably have to use non-parametric stats on my skewed and outlier-rich data set (transformations aren't working). I *think* I understand what my "treatment" is (I need to compare 1 measurement among 4 years), and I think that if the assumptions were met, I'd be using a 1-way ANOVA. However, I've come to this very basic decision after 4 days of pouring over textbooks and course notes, trying to get my brain back in 'stats mode' after being away from it. This is typical, and every time I stop thinking about stats for a few months or even weeks, I'm back to square one.
I have much more complicated decisions to make, and it'd be so great to have some simple way to help me decide what tests to use. I've seen charts in some references, but they tend to only cover 'basic' stats (up to a 2-way ANOVA)...I think I'll be heading into the realm multivariate non-parametric stats, which seem like afterthoughts in most stats treatments, assuming that the person running those tests are stats whizzes!
Thanks!
Kelly:
You said you are comparing one measure over four years... by that, do you mean that you have four years of repeated measures from the same participants?
OR
Do you mean that you have one measure on four DIFFERENT groups of people from four different years?
The answer to that question will result in two very different kinds of analysis.
Best.
Jeremy
I don't really know if this is a helpful answer to your question, but what I think is that what makes Stats scary is thinking that what you got is lower than the average, I mean the numbers that you have are not even close to the standards. Is that answer good enough? :-)
Cape Town<