5,800 reputation
927
bio website integrativestatistics.com
location Boston, MA
age
visits member for 3 years, 3 months
seen 13 hours ago

1d
comment Clustering using density fields
Elaine...I'm your desnity.
1d
comment How to test for mediation with a continuous mediator and DV, but a categorical predictor?
Your diagram is helpful, although I'd like to know for sure what "aggregated 5 item scale" means.
2d
answered How to do multivariate machine learning? (predicting multiple dependent variables)
Apr
19
revised How to fit a smooth curve to data in order to mimic an elbow?
edited title
Apr
19
comment How to determine probability of score when selecting from pool of values?
2nd paragraph: Shouldn't you divide by 3! as well, yielding 220 combinations (not permutations)?
Apr
19
comment Differences in edition of advanced statistics books
An interesting related question is how many times older editions have been shown to contain demonstrably incorrect material.
Apr
14
comment Why would I want to use my data this way, and what am I supposed to do with this regression?
Sounds as if you should be asking the person who gave the assignment.
Apr
10
comment How to find the probability that a random sample of 10 points out of 52 points of data will have a mean close to the total mean?
I believe you mean the probability that the 10-week sample mean will be close to the 52-week sample mean. Also, are data collected once per week? And, to answer the question requires that one know the standard deviation within the entire 52-week sample.
Mar
17
awarded  Popular Question
Mar
13
awarded  regression
Mar
8
comment Statistical analysis on several data sources - possible?
Odd that you couldn't find good information about this sort of problem. You've tagged the question with "missing-data": just clicking on that tag brings up hundreds of threads on this site, and a generic web search identifies such a great variety of books, articles, tutorials, and trainings that would be relevant.
Feb
28
awarded  Nice Question
Feb
18
awarded  Notable Question
Feb
9
comment What is your favorite “data analysis” cartoon?
"Because medical research findings can be difficult to reconcile, are not always pre-digested, and can seem overwhelming to us casual observers, let us make fun of those who dedicate their lives to obtaining them."
Feb
8
comment Does it make sense to plot train and test results on a common ROC graph?
@gung - +1, yet when you say the shape "shouldn't change in any important way" - that I think is the main point of the question, to ask whether superimposition is a valid way of checking that. imo superimposition would allow only an informal/impressionistic check. It will be hard to test specific differences, e.g., if at a sensitivity of .1, specificity is __ in the training set but only ___ in the test set.
Feb
1
comment A question on sampling distributions
There's at least one article by Thomas Kane and Douglas Staiger, c. 2000, that's relevant to this topic.
Feb
1
comment How should I transform non-negative data including zeros?
@HongOoi - can you suggest any readings on when this approach is and isn't applicable?
Feb
1
answered How should I transform non-negative data including zeros?
Feb
1
comment Accounting for post-intervention bias following propensity matching
Have you considered any type of path analysis or structural equation modeling?
Feb
1
comment Accounting for post-intervention bias following propensity matching
Thanks for contributing this well-done and entertaining post. I hope you get great answers.