Would the following stats topics be in the scope of the math community?
Hello,
I've read the thread Scope of site (comparison with CrossValidated), and started thinking about launching a proposal for a "statistics" community on codidact (but maybe others have already thought about it?).
I identified the following topics that might not be in the scope of the math community, but I wonder what more experienced users than me think about it:
- correct use of statistical methods and their interpretation
- meta-analysis
- sampling methods
- prior elicitation
- data mining
- natural language processing
- study design
- simulation
- data visualization
- ethics in statistics
- careers in statistics
- ...
There may be other topics I did not think of.
Here are more specific examples of questions taken from Cross Validated:
- When to use Fisher versus Neyman-Pearson framework?
- Principled way of collapsing categorical variables with many levels?
- How to interpret a QQ plot?
- Is there a good reason for a lab to repeat experiments instead of conducting a single larger blocked experiment?
- Using post-stratification weights in R survey package
- Recommend references on survey sample weighting
- Why are p-values misleading after performing a stepwise selection?
- Why does finding small effects in large studies indicate publication bias?
- Eliciting priors from experts
- Avoiding social discrimination in model building
- What are some good references on the history of ethics in statistics?
- How to simulate artificial data for logistic regression?
- Simulation of logistic regression power analysis - designed experiments
- Why is skip-gram better for infrequent words than CBOW?
- How does Keras 'Embedding' layer work?
- How is the .similarity method in SpaCy computed?
- How to produce a pretty plot of the results of k-means cluster analysis?
- When are Log scales appropriate?
- Having a job in data-mining without a PhD
My underlying question is probably more something along the lines of "In the light of the above examples, do you think there is some room for a stats community on codidact? Or is it already in the scope of the math community?"
2 answers
You are accessing this answer with a direct link, so it's being shown above all other answers regardless of its score. You can return to the normal view.
There is some overlap and it's not always easy to draw the line between mathematics and science (particularly theoretical physics), but most of the examples given seem to me to be more on the "using mathematics in other fields" side of the line than the "mathematics" side of the line.
I think that really the important question is whether the people who would be able to answer statistics questions will participate on this site, which may come down to whether they see themselves as mathematicians. If you don't get any answers from statisticians in the next week, maybe you should make a proposal for a new community and see whether there's any response there.
0 comment threads
Being somewhat active at CrossValidated (and actually a mathematician by training), I would say that:
-
Statistics, as per the scope you outline and per about 98% of the activity at CV, is not a subset of math, so no, I would say that this content is already in scope here. (This is of course not to say that statistics in the sense of measure theory, probability etc. are off-topic here.)
Plus: people who have this kind of question will not believe they will find an answer at a math site.
-
I absolutely think there is room for a stats site at Codidact, with the scope you outline. (Maybe add topics like forecasting, classification etc.)
FWIW, I will gladly commit to contributing at any such site here.
0 comment threads
0 comment threads